HP Hewlett Packard Landscape Lighting HP UX 11i v3 User Manual
Ver ita s Volu m e Ma n a ger 5.0  
Ad m in istr a tor ’s Gu id e  
HP -UX 11i v3  
F ir st Ed ition  
Ma n u fa ct u r in g Pa r t Nu m b er : 5992-3942  
Ma y 2008  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Contents  
Chapter 1  
Understanding Veritas Volume Manager  
VxVM and the operating system .......................................................................19  
How data is stored ........................................................................................19  
How VxVM handles storage management .......................................................20  
Physical objects—physical disks ................................................................20  
Virtual objects ..............................................................................................26  
Volume layouts in VxVM ....................................................................................34  
Implementation of non-layered volumes .................................................34  
Implementation of layered volumes .........................................................34  
Layout methods ............................................................................................35  
Concatenation and spanning .....................................................................35  
Striping (RAID-0) .........................................................................................38  
Mirroring (RAID-1) ......................................................................................42  
Striping plus mirroring (mirrored-stripe or RAID-0+1) ........................42  
Mirroring plus striping (striped-mirror, RAID-1+0 or RAID-10) .........43  
RAID-5 (striping with parity) .....................................................................45  
Layered volumes ..........................................................................................51  
Online relayout .....................................................................................................54  
How online relayout works .........................................................................54  
Limitations of online relayout ...................................................................57  
Transformation characteristics .................................................................58  
Transformations and volume length ........................................................58  
Volume resynchronization .................................................................................59  
Dirty flags ......................................................................................................59  
Resynchronization process ........................................................................59  
Dirty region logging ............................................................................................60  
Dirty region logs ...........................................................................................60  
Log subdisks and plexes ..............................................................................61  
Sequential DRL .............................................................................................61  
SmartSync recovery accelerator ...............................................................62  
Volume snapshots ................................................................................................63  
Comparison of snapshot features ..............................................................65  
FastResync ............................................................................................................66  
FastResync enhancements .........................................................................67  
Non-persistent FastResync ........................................................................67  
Persistent FastResync .................................................................................68  
Download from Www.Somanuals.com. All Manuals Search And Download.  
6
Contents  
DCO volume versioning .............................................................................. 68  
FastResync limitations ............................................................................... 74  
Hot-relocation ...................................................................................................... 75  
Volume sets .......................................................................................................... 75  
Chapter 2  
Administering disks  
Disk devices .......................................................................................................... 77  
Disk device naming in VxVM ..................................................................... 78  
Private and public disk regions ................................................................. 80  
Discovering and configuring newly added disk devices ................................ 82  
Partial device discovery .............................................................................. 82  
Discovering disks and dynamically adding disk arrays ......................... 83  
Third-party driver coexistence .................................................................. 84  
Administering the Device Discovery Layer ............................................. 85  
Placing disks under VxVM control ................................................................... 90  
Changing the disk-naming scheme ................................................................... 91  
Regenerating persistent device names .................................................... 93  
Changing device naming for TPD-controlled enclosures ...................... 94  
Discovering the association between enclosure and OS based disk names 94  
Issues regarding persistent simple or nopriv disks with enclosure-based  
naming ................................................................................................... 94  
Installing and formatting disks ......................................................................... 96  
Displaying and changing default disk layout attributes ............................... 97  
Adding a disk to VxVM ....................................................................................... 97  
Reinitializing a disk ...................................................................................101  
Using vxdiskadd to place a disk under control of VxVM .....................101  
Rootability .........................................................................................................102  
VxVM root disk volume restrictions .......................................................103  
Root disk mirrors .......................................................................................103  
Booting root volumes ................................................................................104  
Setting up a VxVM root disk and mirror ................................................104  
Creating an LVM root disk from a VxVM root disk ..............................106  
Adding swap volumes to a VxVM rootable system ...............................107  
Adding persistent dump volumes to a VxVM rootable system ..........107  
Removing a persistent dump volume .....................................................108  
Dynamic LUN expansion ..................................................................................108  
Removing disks ..................................................................................................110  
Removing a disk with subdisks ................................................................111  
Removing a disk with no subdisks ..........................................................112  
Removing a disk from VxVM control .............................................................112  
Removing and replacing disks .........................................................................112  
Replacing a failed or removed disk .........................................................115  
Enabling a disk ...................................................................................................117  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Contents  
7
Taking a disk offline ..........................................................................................118  
Renaming a disk .................................................................................................119  
Reserving disks ..................................................................................................119  
Displaying disk information ............................................................................120  
Displaying disk information with vxdiskadm .......................................121  
Controlling Powerfail Timeout ........................................................................122  
Setting the PFTO values ............................................................................122  
Displaying the PFTO values .....................................................................122  
Enabling or disabling PFTO ......................................................................123  
Chapter 3  
Administering dynamic multipathing (DMP)  
How DMP works .................................................................................................125  
How DMP monitors I/O on paths ............................................................128  
Load balancing ............................................................................................129  
DMP coexistence with HP-UX native multipathing .............................130  
DMP in a clustered environment .............................................................132  
Disabling and enabling multipathing for specific devices ..........................133  
Disabling multipathing and making devices invisible to VxVM ........133  
Enabling multipathing and making devices visible to VxVM .............134  
Enabling and disabling I/O for controllers and storage processors ..........136  
Displaying DMP database information ..........................................................137  
Displaying the paths to a disk ..........................................................................137  
Administering DMP using vxdmpadm ...........................................................139  
Retrieving information about a DMP node ............................................139  
Displaying the members of a LUN group ...............................................140  
Displaying paths controlled by a DMP node, controller or array port 140  
Displaying information about controllers .............................................141  
Displaying information about enclosures ..............................................142  
Displaying information about array ports .............................................142  
Displaying information about TPD-controlled devices .......................143  
Gathering and displaying I/O statistics ..................................................144  
Setting the attributes of the paths to an enclosure ..............................146  
Displaying the I/O policy ..........................................................................147  
Specifying the I/O policy ..........................................................................147  
Disabling I/O for paths, controllers or array ports ...............................153  
Enabling I/O for paths, controllers or array ports ................................154  
Upgrading disk controller firmware .......................................................154  
Renaming an enclosure .............................................................................155  
Configuring the response to I/O failures ................................................156  
Configuring the I/O throttling mechanism ............................................157  
Displaying recoveryoption values ...........................................................159  
Configuring DMP path restoration policies ...........................................160  
Stopping the DMP path restoration thread ...........................................161  
Download from Www.Somanuals.com. All Manuals Search And Download.  
8
Contents  
Displaying the status of the DMP path restoration thread .................161  
Displaying information about the DMP error-handling thread .........162  
Configuring array policy modules ..........................................................162  
Chapter 4  
Creating and administering disk groups  
Specifying a disk group to commands ............................................................167  
System-wide reserved disk groups .........................................................167  
Rules for determining the default disk group .......................................168  
Displaying disk group information .................................................................169  
Displaying free space in a disk group .....................................................170  
Creating a disk group ........................................................................................170  
Adding a disk to a disk group ...........................................................................171  
Removing a disk from a disk group ................................................................172  
Deporting a disk group .....................................................................................173  
Importing a disk group .....................................................................................174  
Handling disks with duplicated identifiers ...................................................175  
Writing a new UDID to a disk ..................................................................176  
Importing a disk group containing cloned disks ..................................176  
Sample cases of operations on cloned disks ..........................................178  
Renaming a disk group .....................................................................................183  
Moving disks between disk groups .................................................................184  
Moving disk groups between systems ............................................................185  
Handling errors when importing disks ..................................................186  
Reserving minor numbers for disk groups ............................................187  
Compatibility of disk groups between platforms ..................................189  
Handling conflicting configuration copies ....................................................190  
Example of a serial split brain condition in a cluster ..........................190  
Correcting conflicting configuration information ...............................194  
Reorganizing the contents of disk groups .....................................................195  
Limitations of disk group split and join .................................................199  
Listing objects potentially affected by a move ......................................200  
Moving objects between disk groups ......................................................203  
Splitting disk groups .................................................................................205  
Joining disk groups ....................................................................................206  
Disabling a disk group .......................................................................................207  
Destroying a disk group ....................................................................................208  
Recovering a destroyed disk group .........................................................208  
Upgrading a disk group ....................................................................................208  
Managing the configuration daemon in VxVM ............................................212  
Backing up and restoring disk group configuration data ............................213  
Using vxnotify to monitor configuration changes .......................................213  
Chapter 5  
Creating and administering subdisks  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Contents  
9
Creating subdisks ...............................................................................................215  
Displaying subdisk information ......................................................................216  
Moving subdisks .................................................................................................217  
Splitting subdisks ..............................................................................................217  
Joining subdisks .................................................................................................218  
Associating subdisks with plexes ....................................................................218  
Associating log subdisks ...................................................................................220  
Dissociating subdisks from plexes ..................................................................221  
Removing subdisks ............................................................................................221  
Changing subdisk attributes ............................................................................221  
Chapter 6  
Creating and administering plexes  
Creating plexes ...................................................................................................223  
Creating a striped plex ......................................................................................224  
Displaying plex information ............................................................................224  
Plex states ...................................................................................................224  
Plex condition flags ...................................................................................228  
Plex kernel states .......................................................................................229  
Attaching and associating plexes ....................................................................229  
Taking plexes offline .........................................................................................230  
Detaching plexes ................................................................................................231  
Reattaching plexes .............................................................................................231  
Moving plexes .....................................................................................................232  
Copying volumes to plexes ...............................................................................233  
Dissociating and removing plexes ...................................................................233  
Changing plex attributes ..................................................................................234  
Chapter 7  
Creating volumes  
Types of volume layouts ...................................................................................236  
Supported volume logs and maps ...........................................................237  
Creating a volume ..............................................................................................238  
Advanced approach ...................................................................................238  
Assisted approach ......................................................................................239  
Using vxassist .....................................................................................................239  
Setting default values for vxassist ..........................................................241  
Discovering the maximum size of a volume ..................................................242  
Disk group alignment constraints on volumes .............................................242  
Creating a volume on any disk .........................................................................243  
Creating a volume on specific disks ................................................................244  
Specifying ordered allocation of storage to volumes ...........................245  
Creating a mirrored volume .............................................................................249  
Creating a mirrored-concatenated volume ............................................249  
Download from Www.Somanuals.com. All Manuals Search And Download.  
10 Contents  
Creating a concatenated-mirror volume ................................................249  
Creating a volume with a version 0 DCO volume .........................................250  
Creating a volume with a version 20 DCO volume .......................................252  
Creating a volume with dirty region logging enabled ..................................252  
Creating a striped volume ................................................................................253  
Creating a mirrored-stripe volume .........................................................254  
Creating a striped-mirror volume ...........................................................254  
Mirroring across targets, controllers or enclosures ....................................255  
Creating a RAID-5 volume ................................................................................256  
Creating tagged volumes ..................................................................................257  
Creating a volume using vxmake ....................................................................258  
Creating a volume using a vxmake description file .............................259  
Initializing and starting a volume ...................................................................260  
Initializing and starting a volume created using vxmake ...................261  
Accessing a volume ...........................................................................................262  
Chapter 8  
Administering volumes  
Displaying volume information ......................................................................264  
Volume states .............................................................................................265  
Volume kernel states .................................................................................266  
Monitoring and controlling tasks ...................................................................267  
Specifying task tags ...................................................................................267  
Managing tasks with vxtask .....................................................................268  
Stopping a volume .............................................................................................270  
Putting a volume in maintenance mode ................................................270  
Starting a volume ..............................................................................................271  
Adding a mirror to a volume ............................................................................271  
Mirroring all volumes ...............................................................................272  
Mirroring volumes on a VM disk .............................................................272  
Removing a mirror ............................................................................................273  
Adding logs and maps to volumes ...................................................................274  
Preparing a volume for DRL and instant snapshots ....................................275  
Specifying storage for version 20 DCO plexes ......................................276  
Using a DCO and DCO volume with a RAID-5 volume .........................277  
Determining the DCO version number ...................................................277  
Determining if DRL is enabled on a volume ..........................................278  
Determining if DRL logging is active on a volume ...............................278  
Disabling and re-enabling DRL ................................................................278  
Removing support for DRL and instant snapshots from a volume ...279  
Upgrading existing volumes to use version 20 DCOs ..................................279  
Adding traditional DRL logging to a mirrored volume ................................281  
Removing a traditional DRL log ..............................................................282  
Adding a RAID-5 log ..........................................................................................283  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Contents 11  
Adding a RAID-5 log using vxplex ...........................................................283  
Removing a RAID-5 log .............................................................................284  
Resizing a volume ..............................................................................................284  
Resizing volumes using vxresize .............................................................285  
Resizing volumes using vxassist .............................................................286  
Resizing volumes using vxvol ..................................................................287  
Setting tags on volumes ....................................................................................288  
Changing the read policy for mirrored volumes ...........................................289  
Removing a volume ...........................................................................................290  
Moving volumes from a VM disk .....................................................................290  
Enabling FastResync on a volume ...................................................................292  
Checking whether FastResync is enabled on a volume ........................293  
Disabling FastResync ................................................................................293  
Performing online relayout ..............................................................................294  
Permitted relayout transformations .......................................................295  
Specifying a non-default layout ...............................................................298  
Specifying a plex for relayout ..................................................................298  
Tagging a relayout operation ...................................................................298  
Viewing the status of a relayout ..............................................................299  
Controlling the progress of a relayout ....................................................299  
Converting between layered and non-layered volumes ...............................300  
Chapter 9  
Administering volume snapshots  
Traditional third-mirror break-off snapshots ...............................................305  
Full-sized instant snapshots ............................................................................307  
Space-optimized instant snapshots ................................................................309  
Emulation of third-mirror break-off snapshots ............................................310  
Linked break-off snapshot volumes ................................................................311  
Cascaded snapshots ...........................................................................................312  
Creating a snapshot of a snapshot ..........................................................313  
Creating multiple snapshots ............................................................................317  
Restoring the original volume from a snapshot ...........................................317  
Creating instant snapshots ..............................................................................319  
Preparing to create instant and break-off snapshots ..........................321  
Creating and managing space-optimized instant snapshots ..............324  
Creating and managing full-sized instant snapshots ..........................327  
Creating and managing third-mirror break-off snapshots .................329  
Creating and managing linked break-off snapshot volumes ..............331  
Creating multiple instant snapshots ......................................................333  
Creating instant snapshots of volume sets ............................................334  
Adding snapshot mirrors to a volume ....................................................336  
Removing a snapshot mirror ...................................................................336  
Removing a linked break-off snapshot volume .....................................337  
Download from Www.Somanuals.com. All Manuals Search And Download.  
12 Contents  
Adding a snapshot to a cascaded snapshot hierarchy .........................337  
Refreshing an instant snapshot ..............................................................337  
Reattaching an instant snapshot ............................................................338  
Reattaching a linked break-off snapshot volume .................................339  
Restoring a volume from an instant snapshot ......................................340  
Dissociating an instant snapshot ............................................................340  
Removing an instant snapshot ................................................................341  
Splitting an instant snapshot hierarchy ................................................341  
Displaying instant snapshot information ..............................................342  
Controlling instant snapshot synchronization .....................................344  
Listing the snapshots created on a cache ..............................................345  
Tuning the autogrow attributes of a cache ............................................346  
Growing and shrinking a cache ...............................................................347  
Removing a cache ......................................................................................347  
Creating traditional third-mirror break-off snapshots ...............................348  
Converting a plex into a snapshot plex ..................................................351  
Creating multiple snapshots ....................................................................352  
Reattaching a snapshot volume ..............................................................352  
Adding plexes to a snapshot volume ......................................................353  
Dissociating a snapshot volume ..............................................................354  
Displaying snapshot information ...........................................................355  
Adding a version 0 DCO and DCO volume .....................................................356  
Specifying storage for version 0 DCO plexes ........................................357  
Removing a version 0 DCO and DCO volume ........................................358  
Reattaching a version 0 DCO and DCO volume .....................................359  
Chapter 10  
Creating and administering volume sets  
Creating a volume set ........................................................................................362  
Adding a volume to a volume set ....................................................................362  
Listing details of volume sets ..........................................................................363  
Stopping and starting volume sets .................................................................363  
Removing a volume from a volume set ..........................................................364  
Raw device node access to component volumes ...........................................364  
Enabling raw device access when creating a volume set .....................365  
Displaying the raw device access settings for a volume set ................366  
Controlling raw device access for an existing volume set ...................366  
Chapter 11  
Configuring off-host processing  
Implementing off-host processing solutions ................................................370  
Implementing off-host online backup ....................................................371  
Implementing decision support ..............................................................374  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Contents 13  
Chapter 12  
Administering hot-relocation  
How hot-relocation works ................................................................................380  
Partial disk failure mail messages ...........................................................383  
Complete disk failure mail messages ......................................................384  
How space is chosen for relocation .........................................................384  
Configuring a system for hot-relocation ........................................................385  
Displaying spare disk information ..................................................................386  
Marking a disk as a hot-relocation spare .......................................................387  
Removing a disk from use as a hot-relocation spare ....................................388  
Excluding a disk from hot-relocation use ......................................................388  
Making a disk available for hot-relocation use .............................................389  
Configuring hot-relocation to use only spare disks .....................................390  
Moving and unrelocating subdisks .................................................................390  
Moving and unrelocating subdisks using vxdiskadm ..........................391  
Moving and unrelocating subdisks using vxassist ...............................392  
Moving and unrelocating subdisks using vxunreloc ............................392  
Restarting vxunreloc after errors ...........................................................394  
Modifying the behavior of hot-relocation ......................................................395  
Chapter 13  
Administering cluster functionality  
Overview of cluster volume management .....................................................398  
Private and shared disk groups ...............................................................401  
Activation modes of shared disk groups ................................................402  
Connectivity policy of shared disk groups .............................................404  
Effect of disk connectivity on cluster reconfiguration ........................409  
Limitations of shared disk groups ...........................................................409  
Cluster initialization and configuration .........................................................410  
Cluster reconfiguration .............................................................................410  
Volume reconfiguration ............................................................................413  
Node shutdown ...........................................................................................416  
Node abort ...................................................................................................417  
Cluster shutdown .......................................................................................417  
Multiple host failover configurations .............................................................417  
Import lock ..................................................................................................418  
Failover ........................................................................................................418  
Corruption of disk group configuration .................................................419  
Administering VxVM in cluster environments .............................................420  
Requesting node status and discovering the master node ..................420  
Determining if a disk is shareable ...........................................................421  
Listing shared disk groups .......................................................................421  
Creating a shared disk group ...................................................................422  
Importing disk groups as shared .............................................................423  
Download from Www.Somanuals.com. All Manuals Search And Download.  
14 Contents  
Converting a disk group from shared to private ...................................424  
Moving objects between disk groups ......................................................424  
Splitting disk groups .................................................................................424  
Joining disk groups ....................................................................................424  
Changing the activation mode on a shared disk group ........................425  
Setting the disk detach policy on a shared disk group ........................425  
Setting the disk group failure policy on a shared disk group .............426  
Creating volumes with exclusive open access by a node .....................426  
Setting exclusive open access to a volume by a node ..........................426  
Displaying the cluster protocol version .................................................427  
Displaying the supported cluster protocol version range ...................427  
Upgrading the cluster protocol version .................................................428  
Recovering volumes in shared disk groups ...........................................428  
Obtaining cluster performance statistics ..............................................428  
Chapter 14  
Administering  
sites and remote mirrors  
Configuring sites for hosts and disks .............................................................434  
Configuring site-based allocation on a disk group .......................................434  
Configuring site consistency on a disk group ...............................................435  
Configuring site consistency on a volume .....................................................435  
Setting the siteread policy on a volume .........................................................436  
Site-based allocation of storage to volumes ..................................................436  
Examples of storage allocation using sites ............................................438  
Making an existing disk group site consistent ..............................................439  
Fire drill — testing the configuration .............................................................440  
Simulating site failure ..............................................................................440  
Recovery from simulated site failure .....................................................440  
Automatic site reattachment ...........................................................................440  
Failure scenarios and recovery procedures ...................................................441  
Recovery from a loss of site connectivity ..............................................442  
Recovery from host failure ......................................................................442  
Recovery from storage failure .................................................................442  
Recovery from site failure ........................................................................443  
Chapter 15  
Using Storage Expert  
About Storage Expert ........................................................................................445  
How Storage Expert works ...............................................................................446  
Before using Storage Expert ............................................................................446  
Running Storage Expert ...................................................................................446  
Discovering what a rule does ...................................................................447  
Displaying rule attributes and their default values .............................447  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Contents 15  
Running a rule ............................................................................................447  
Identifying configuration problems using Storage Expert .........................449  
Recovery time .............................................................................................449  
Disk groups .................................................................................................451  
Disk striping ...............................................................................................453  
Disk sparing and relocation management .............................................454  
Hardware failures ......................................................................................454  
Rootability ...................................................................................................454  
System name ...............................................................................................454  
Rule definitions and attributes ........................................................................456  
Chapter 16  
Performance monitoring and tuning  
Performance guidelines ....................................................................................463  
Data assignment .........................................................................................463  
Striping ........................................................................................................464  
Mirroring .....................................................................................................464  
Combining mirroring and striping ..........................................................465  
RAID-5 .........................................................................................................465  
Volume read policies .................................................................................466  
Performance monitoring ..................................................................................467  
Setting performance priorities ................................................................467  
Obtaining performance data ....................................................................467  
Using performance data ...........................................................................469  
Tuning VxVM .....................................................................................................472  
General tuning guidelines ........................................................................472  
Tuning guidelines for large systems .......................................................473  
Changing the values of tunables ..............................................................474  
Tunable parameters ...................................................................................475  
Appendix A  
Appendix B  
Commands summary  
Online manual pages .........................................................................................507  
Section 1M — administrative commands ...............................................507  
Section 4 — file formats ............................................................................510  
Section 7 — device driver interfaces .......................................................510  
Configuring Veritas Volume Manager  
Setup tasks after installation ...........................................................................511  
Adding unsupported disk arrays as JBODs ....................................................512  
Adding foreign devices ......................................................................................512  
Adding disks to disk groups .............................................................................512  
Guidelines for configuring storage .................................................................513  
Mirroring guidelines ..................................................................................514  
Download from Www.Somanuals.com. All Manuals Search And Download.  
16 Contents  
Dirty region logging guidelines ...............................................................515  
Striping guidelines ....................................................................................515  
RAID-5 guidelines ......................................................................................516  
Hot-relocation guidelines .........................................................................516  
Accessing volume devices ........................................................................518  
Controlling VxVM’s view of multipathed devices ........................................518  
Configuring cluster support .............................................................................518  
Configuring shared disk groups ..............................................................519  
Converting existing VxVM disk groups to shared disk groups ..........519  
Reconfiguration tasks .......................................................................................520  
Changing the name of the default disk group .......................................520  
Enabling or disabling enclosure-based naming ....................................520  
Glossary  
Index  
521  
531  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter  
1
Understanding Veritas  
Volume Manager  
TM  
Veritas Volume Manager (VxVM) by Symantec is a storage management  
subsystem that allows you to manage physical disks as logical devices called  
volumes. A VxVM volume appears to applications and the operating system as a  
physical disk on which file systems, databases and other managed data objects  
can be configured.  
VxVM provides easy-to-use online disk storage management for computing  
environments and Storage Area Network (SAN) environments. By supporting  
the Redundant Array of Independent Disks (RAID) model, VxVM can be  
configured to protect against disk and hardware failure, and to increase I/O  
throughput. Additionally, VxVM provides features that enhance fault tolerance  
and fast recovery from disk failure.  
VxVM overcomes physical restrictions imposed by hardware disk devices by  
providing a logical volume management layer. This allows volumes to span  
multiple disks.  
VxVM provides the tools to improve performance and ensure data availability  
and integrity. You can also use VxVM to dynamically configure disk storage  
while the system is active.  
The following sections of this chapter explain fundamental concepts of VxVM:  
■
■
■
VxVM and the operating system  
How VxVM handles storage management  
Volume layouts in VxVM  
The following sections introduce you to advanced features of VxVM:  
■
■
■
Online relayout  
Volume resynchronization  
Dirty region logging  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
18 Understanding Veritas Volume Manager  
■
■
■
■
Volume snapshots  
FastResync  
Hot-relocation  
Volume sets  
Further information on administering Veritas Volume Manager may be found in  
the following documents:  
■
Veritas Storage Foundation Cross-Platform Data Sharing Administrator’s  
Guide  
Provides more information on using the Cross-platform Data Sharing (CDS)  
feature of Veritas Volume Manager, which allows you to move VxVM disks  
and objects between machines that are running under different operating  
systems.  
Note: The CDS feature requires a Veritas Storage Foundation license.  
■
Veritas Storage Foundation Intelligent Storage Provisioning Administrator’s  
Guide  
Describes the command-line interface to the Veritas Intelligent Storage  
Provisioning (ISP) feature, which uses a rule-based engine to create VxVM  
objects and make optimal usage of the available storage.  
■
Veritas FlashSnap Point-In-Time Copy Solutions Administrator’s Guide  
Provides guidelines on using the features of the FlashSnap software to  
implement various point-in-time copy solutions for backup, and database  
replication.  
Note: The FlashSnap feature requires a separate license.  
■
■
Veritas Volume Manager Troubleshooting Guide  
Describes recovery from hardware failure, disk group configuration and  
restoration, command and transaction logging, and common error  
messages together with suggested solutions.  
Veritas Enterprise Administrator User’s Guide  
Describes how to use the Veritas Enterprise Administrator — the graphical  
user interface to Veritas Volume Manager. More detailed information is  
available in the VEA online help.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Understanding Veritas Volume Manager 19  
VxVM and the operating system  
VxVM and the operating system  
VxVM operates as a subsystem between your operating system and your data  
management systems, such as file systems and database management systems.  
VxVM is tightly coupled with the operating system. Before a disk can be brought  
under VxVM control, the disk must be accessible through the operating system  
device interface. VxVM is layered on top of the operating system interface  
services, and is dependent upon how the operating system accesses physical  
disks.  
VxVM is dependent upon the operating system for the following functionality:  
■
■
■
operating system (disk) devices  
device handles  
VxVM dynamic multipathing (DMP) metadevice  
This guide introduces you to the VxVM commands which are used to carry out  
the tasks associated with VxVM objects. These commands are described on the  
relevant manual pages and in the chapters of this guide where VxVM tasks are  
described.  
VxVM relies on the following constantly-running daemons and kernel threads  
for its operation:  
■
■
■
vxconfigd—The VxVM configuration daemon maintains disk and group  
configurations and communicates configuration changes to the kernel, and  
modifies configuration information stored on disks.  
vxiod—VxVM I/O kernel threads provide extended I/O operations without  
blocking calling processes. By default, 16 I/O threads are started at boot  
time, and at least one I/O thread must continue to run at all times.  
vxrelocd—The hot-relocation daemon monitors VxVM for events that affect  
redundancy, and performs hot-relocation to restore redundancy.  
How data is stored  
There are several methods used to store data on physical disks. These methods  
organize data on the disk so the data can be stored and retrieved efficiently. The  
basic method of disk organization is called formatting. Formatting prepares the  
hard disk so that files can be written to and retrieved from the disk by using a  
prearranged storage pattern.  
Hard disks are formatted, and information stored, using two methods: physical-  
storage layout and logical-storage layout. VxVM uses the logical-storage layout  
method. The types of storage layout supported by VxVM are introduced in this  
chapter.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
20 Understanding Veritas Volume Manager  
How VxVM handles storage management  
How VxVM handles storage management  
VxVM uses two types of objects to handle storage management: physical objects  
and virtual objects.  
■
■
Physical objects—physical disks or other hardware with block and raw  
operating system device interfaces that are used to store data.  
Virtual objects—When one or more physical disks are brought under the  
control of VxVM, it creates virtual objects called volumes on those physical  
disks. Each volume records and retrieves data from one or more physical  
disks. Volumes are accessed by file systems, databases, or other applications  
in the same way that physical disks are accessed. Volumes are also  
composed of other virtual objects (plexes and subdisks) that are used in  
changing the volume configuration. Volumes and their virtual components  
are called virtual objects or VxVM objects.  
Physical objects—physical disks  
A physical disk is the basic storage device (media) where the data is ultimately  
stored. You can access the data on a physical disk by using a device name to  
locate the disk.  
Figure 1-1 shows how a physical disk and device name (devname) are illustrated  
in this document.  
Figure 1-1  
Physical disk example  
devname  
In HP-UX 11i v3, disks may be identified either by their legacy device name,  
which takes the form c#t#d#, or by their persistent (or agile) device name,  
which takes the form disk##.  
In a legacy device name, c#specifies the controller, t#specifies the target ID,  
and d#specifies the disk. For example, the device name c0t0d0is the entire  
hard disk that is connected to controller number 0in the system, with a target  
ID of 0, and physical disk number of 0. The equivalent persistent device name  
might be disk33.  
In this document, legacy device names are generally shown as this format is the  
same as the default format that is used by the Dynamic Discovery Layer (DDL)  
and Dynamic Multipathing (DMP) features of VxVM.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Understanding Veritas Volume Manager 21  
How VxVM handles storage management  
VxVM writes identification information on physical disks under VxVM control  
(VM disks). VxVM disks can be identified even after physical disk disconnection  
or system outages. VxVM can then re-form disk groups and logical objects to  
provide failure detection and to speed system recovery.  
VxVM accesses all disks as entire physical disks without partitions.  
Disk arrays  
Performing I/O to disks is a relatively slow process because disks are physical  
devices that require time to move the heads to the correct position on the disk  
before reading or writing. If all of the read or write operations are done to  
individual disks, one at a time, the read-write time can become unmanageable.  
Performing these operations on multiple disks can help to reduce this problem.  
A disk array is a collection of physical disks that VxVM can represent to the  
operating system as one or more virtual disks or volumes. The volumes created  
by VxVM look and act to the operating system like physical disks. Applications  
that interact with volumes should work in the same way as with physical disks.  
Figure 1-2 illustrates how VxVM represents the disks in a disk array as several  
volumes to the operating system.  
Data can be spread across several disks within an array to distribute or balance  
I/O operations across the disks. Using parallel I/O across multiple disks in this  
way improves I/O performance by increasing data transfer speed and overall  
throughput for the array.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
22 Understanding Veritas Volume Manager  
How VxVM handles storage management  
Figure 1-2  
How VxVM presents the disks in a disk array as volumes to the  
operating system  
Operating system  
Veritas Volume Manager  
Volumes  
Physical disks  
Disk 1  
Disk 2  
Disk 3  
Disk 4  
Multipathed disk arrays  
Some disk arrays provide multiple ports to access their disk devices. These  
ports, coupled with the host bus adaptor (HBA) controller and any data bus or  
I/O processor local to the array, make up multiple hardware paths to access the  
disk devices. Such disk arrays are called multipathed disk arrays. This type of  
disk array can be connected to host systems in many different configurations,  
(such as multiple ports connected to different controllers on a single host,  
chaining of the ports through a single controller on a host, or ports connected to  
different hosts simultaneously).  
HP-UX 11i v3 provides its own native multipathing solution in addition to the  
Dynamic Multipathing (DMP) feature that is used by VxVM. These two  
multipathing solutions can coexist on the same system. For more information,  
see “DMP coexistence with HP-UX native multipathing” on page 130.  
Device discovery  
Device discovery is the term used to describe the process of discovering the  
disks that are attached to a host. This feature is important for DMP because it  
needs to support a growing number of disk arrays from a number of vendors. In  
conjunction with the ability to discover the devices attached to a host, the  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Understanding Veritas Volume Manager 23  
How VxVM handles storage management  
Device Discovery service enables you to add support dynamically for new disk  
arrays. This operation, which uses a facility called the Device Discovery Layer  
(DDL), is achieved without the need for a reboot.  
This means that you can dynamically add a new disk array to a host, and run a  
command which scans the operating system’s device tree for all the attached  
disk devices, and reconfigures DMP with the new device database. For more  
information, see “Administering the Device Discovery Layer” on page 85.  
Enclosure-based naming  
Enclosure-based naming provides an alternative to the disk device naming  
described in “Physical objects—physical disks” on page 20. This allows disk  
devices to be named for enclosures rather than for the controllers through  
which they are accessed. In a Storage Area Network (SAN) that uses Fibre  
Channel hubs or fabric switches, information about disk location provided by  
the operating system may not correctly indicate the physical location of the  
disks. For example, c#t#d# naming assigns controller-based device names to  
disks in separate enclosures that are connected to the same host controller.  
Enclosure-based naming allows VxVM to access enclosures as separate physical  
entities. By configuring redundant copies of your data on separate enclosures,  
you can safeguard against failure of one or more enclosures.  
In a typical SAN environment, host controllers are connected to multiple  
enclosures in a daisy chain or through a Fibre Channel hub or fabric switch as  
illustrated in Figure 1-3.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
24 Understanding Veritas Volume Manager  
How VxVM handles storage management  
Figure 1-3  
Example configuration for disk enclosures connected via a fibre  
channel hub or switch  
c1  
Host  
Fibre Channel hub  
or switch  
Disk enclosures  
enc0  
enc1  
enc2  
In such a configuration, enclosure-based naming can be used to refer to each  
disk within an enclosure. For example, the device names for the disks in  
enclosure enc0are named enc0_0, enc0_1, and so on. The main benefit of this  
scheme is that it allows you to quickly determine where a disk is physically  
located in a large SAN configuration.  
Note: In many advanced disk arrays, you can use hardware-based storage  
management to represent several physical disks as one logical disk device to the  
operating system. In such cases, VxVM also sees a single logical disk device  
rather than its component disks. For this reason, when reference is made to a  
disk within an enclosure, this disk may be either a physical or a logical device.  
Another important benefit of enclosure-based naming is that it enables VxVM to  
avoid placing redundant copies of data in the same enclosure. This is a good  
thing to avoid as each enclosure can be considered to be a separate fault domain.  
For example, if a mirrored volume were configured only on the disks in  
enclosure enc1, the failure of the cable between the hub and the enclosure  
would make the entire volume unavailable.  
If required, you can replace the default name that VxVM assigns to an enclosure  
with one that is more meaningful to your configuration. See “Renaming an  
enclosure” on page 155 for details.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Understanding Veritas Volume Manager 25  
How VxVM handles storage management  
In High Availability (HA) configurations, redundant-loop access to storage can  
be implemented by connecting independent controllers on the host to separate  
hubs with independent paths to the enclosures as shown in Figure 1-4.  
Figure 1-4  
Example HA configuration using multiple hubs or switches to  
provide redundant loop access  
c1  
c2  
Host  
Fibre Channel  
hubs or switches  
Disk enclosures  
enc0  
enc1  
enc2  
Such a configuration protects against the failure of one of the host controllers  
(c1and c2), or of the cable between the host and one of the hubs. In this  
example, each disk is known by the same name to VxVM for all of the paths over  
which it can be accessed. For example, the disk device enc0_0represents a  
single disk for which two different paths are known, such as c1t99d0and  
c2t99d0.  
Note: The native multipathing feature of HP-UX 11i v3 similarly maps the  
various physical paths to a disk, and presents these as a single persistent device  
with a name of the form disk##. However, this mechanism is independent of that  
used by VxVM. For instructions on administering native multipathing with  
Base-VxVMand VxVM-Full, please consult the Release Notes for VxVM 5.0 on  
11i v3.  
To take account of fault domains when configuring data redundancy, you can  
control how mirrored volumes are laid out across enclosures as described in  
“Mirroring across targets, controllers or enclosures” on page 255.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
26 Understanding Veritas Volume Manager  
How VxVM handles storage management  
See “Disk device naming in VxVM” on page 78 and “Changing the disk-naming  
scheme” on page 91 for details of the standard and the enclosure-based naming  
schemes, and how to switch between them.  
Virtual objects  
Virtual objects in VxVM include the following:  
■
■
■
■
■
Disk groups  
VM disks  
Subdisks  
Plexes  
Volumes  
The connection between physical objects and VxVM objects is made when you  
place a physical disk under VxVM control.  
After installing VxVM on a host system, you must bring the contents of physical  
disks under VxVM control by collecting the VM disks into disk groups and  
allocating the disk group space to create logical volumes.  
Note: To bring the physical disk under VxVM control, the disk must not be under  
LVM control. For more information on how LVM and VM disks co-exist or how  
to convert LVM disks to VM disks, see the Veritas Volume Manager Migration  
Guide  
Bringing the contents of physical disks under VxVM control is accomplished  
only if VxVM takes control of the physical disks and the disk is not under control  
of another storage manager.  
VxVM creates virtual objects and makes logical connections between the objects.  
The virtual objects are then used by VxVM to do storage management tasks.  
Note: The vxprintcommand displays detailed information on existing VxVM  
objects. For additional information on the vxprintcommand, see “Displaying  
volume information” on page 264 and the vxprint(1M) manual page.  
Combining virtual objects in VxVM  
VxVM virtual objects are combined to build volumes. The virtual objects  
contained in volumes are VM disks, disk groups, subdisks, and plexes. Veritas  
Volume Manager objects are organized as follows:  
■
VM disks are grouped into disk groups  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Understanding Veritas Volume Manager 27  
How VxVM handles storage management  
■
■
Subdisks (each representing a specific region of a disk) are combined to form  
plexes  
Volumes are composed of one or more plexes  
Figure 1-5 shows the connections between Veritas Volume Manager virtual  
objects and how they relate to physical disks. The disk group contains three VM  
disks which are used to create two volumes. Volume vol01is simple and has a  
single plex. Volume vol02is a mirrored volume with two plexes.  
Figure 1-5  
Connection between objects in VxVM  
vol01  
vol02  
Volumes  
vol01-01  
vol02-01  
vol02-02  
vol01-01  
vol02-01  
vol02-02  
Plexes  
disk01-01  
disk02-01  
disk03-01  
Subdisks  
disk01-01  
disk02-01  
disk3-01  
disk01-01  
disk01  
disk02-01  
disk02  
disk03-01  
disk03  
VM disks  
Disk group  
devname1  
Physical disks  
devname2  
devname3  
The various types of virtual objects (disk groups, VM disks, subdisks, plexes and  
volumes) are described in the following sections. Other types of objects exist in  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
28 Understanding Veritas Volume Manager  
How VxVM handles storage management  
Veritas Volume Manager, such as data change objects (DCOs), and cache objects,  
to provide extended functionality. These objects are discussed later in this  
chapter.  
Disk groups  
A disk group is a collection of disks that share a common configuration, and  
which are managed by VxVM (see “VM disks” on page 28). A disk group  
configuration is a set of records with detailed information about related VxVM  
objects, their attributes, and their connections. A disk group name can be up to  
31 characters long.  
In releases prior to VxVM 4.0, the default disk group was rootdg(the root disk  
group). For VxVM to function, the rootdgdisk group had to exist and it had to  
contain at least one disk. This requirement no longer exists, and VxVM can work  
without any disk groups configured (although you must set up at least one disk  
group before you can create any volumes of otherVxVM objects). For more  
information about changes to disk group configuration, see “Creating and  
administering disk groups” on page 165.  
You can create additional disk groups when you need them. Disk groups allow  
you to group disks into logical collections. A disk group and its components can  
be moved as a unit from one host machine to another. The ability to move whole  
volumes and disks between disk groups, to split whole volumes and disks  
between disk groups, and to join disk groups is described in “Reorganizing the  
contents of disk groups” on page 195.  
Volumes are created within a disk group. A given volume and its plexes and  
subdisks must be configured from disks in the same disk group.  
VM disks  
When you place a physical disk under VxVM control, a VM disk is assigned to  
the physical disk. A VM disk is under VxVM control and is usually in a disk  
group. Each VM disk corresponds to one physical disk. VxVM allocates storage  
from a contiguous area of VxVM disk space.  
A VM disk typically includes a public region (allocated storage) and a small  
private region where VxVM internal configuration information is stored.  
Each VM disk has a unique disk media name (a virtual disk name). You can either  
define a disk name of up to 31 characters, or allow VxVM to assign a default  
name that takes the form diskgroup##, where diskgroup is the name of the  
disk group to which the disk belongs (see “Disk groups” on page 28).  
Figure 1-6 shows a VM disk with a media name of disk01that is assigned to the  
physical disk devname.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Understanding Veritas Volume Manager 29  
How VxVM handles storage management  
Figure 1-6  
VM disk example  
disk01  
VM disk  
devname  
Physical disk  
Subdisks  
A subdisk is a set of contiguous disk blocks. A block is a unit of space on the disk.  
VxVM allocates disk space using subdisks. A VM disk can be divided into one or  
more subdisks. Each subdisk represents a specific portion of a VM disk, which is  
mapped to a specific region of a physical disk.  
The default name for a VM disk is diskgroup##and the default name for a  
subdisk is diskgroup##-##, where diskgroup is the name of the disk group to  
which the disk belongs (see “Disk groups” on page 28).  
In Figure 1-7, disk01-01is the name of the first subdisk on the VM disk named  
disk01.  
Figure 1-7  
Subdisk example  
disk01-01  
Subdisk  
disk01-01  
disk01  
VM disk with one subdisk  
A VM disk can contain multiple subdisks, but subdisks cannot overlap or share  
the same portions of a VM disk. Figure 1-8 shows a VM disk with three subdisks.  
(The VM disk is assigned to one physical disk.)  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
30 Understanding Veritas Volume Manager  
How VxVM handles storage management  
Figure 1-8  
Example of three subdisks assigned to one VM Disk  
disk01-01  
disk01-02  
disk01-03  
Subdisks  
disk01-01  
disk01-02  
disk01-03  
VM disk with three subdisks  
disk01  
Any VM disk space that is not part of a subdisk is free space. You can use free  
space to create new subdisks.  
VxVM release 3.0 or higher supports the concept of layered volumes in which  
subdisks can contain volumes. For more information, see “Layered volumes” on  
page 51.  
Plexes  
VxVM uses subdisks to build virtual objects called plexes. A plex consists of one  
or more subdisks located on one or more physical disks. For example, see the  
plex vol01-01shown in Figure 1-9.  
Figure 1-9  
Example of a plex with two subdisks  
Plex with two subdisks  
disk01-01  
disk01-02  
vol01-01  
disk01-01  
disk01-02  
Subdisks  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Understanding Veritas Volume Manager 31  
How VxVM handles storage management  
You can organize data on subdisks to form a plex by using the following  
methods:  
■
■
■
■
concatenation  
striping (RAID-0)  
mirroring (RAID-1)  
striping with parity (RAID-5)  
Concatenation, striping (RAID-0), mirroring (RAID-1) and RAID-5 are described  
in “Volume layouts in VxVM” on page 34.  
Volumes  
A volume is a virtual disk device that appears to applications, databases, and file  
systems like a physical disk device, but does not have the physical limitations of  
a physical disk device. A volume consists of one or more plexes, each holding a  
copy of the selected data in the volume. Due to its virtual nature, a volume is not  
restricted to a particular disk or a specific area of a disk. The configuration of a  
volume can be changed by using VxVM user interfaces. Configuration changes  
can be accomplished without causing disruption to applications or file systems  
that are using the volume. For example, a volume can be mirrored on separate  
disks or moved to use different disk storage.  
Note: VxVM uses the default naming conventions of vol## for volumes and  
vol##-## for plexes in a volume. For ease of administration, you can choose to  
select more meaningful names for the volumes that you create.  
A volume may be created under the following constraints:  
■
■
Its name can contain up to 31 characters.  
It can consist of up to 32 plexes, each of which contains one or more  
subdisks.  
■
■
It must have at least one associated plex that has a complete copy of the  
data in the volume with at least one associated subdisk.  
All subdisks within a volume must belong to the same disk group.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
32 Understanding Veritas Volume Manager  
How VxVM handles storage management  
Note: You can use the Veritas Intelligent Storage Provisioning (ISP) feature to  
create and administer application volumes. These volumes are very similar to  
the traditional VxVM volumes that are described in this chapter. However, there  
are significant differences between the functionality of the two types of volume  
that prevent them from being used interchangeably. Refer to the Veritas Storage  
Foundation Intelligent Storage Provisioning Administrator’s Guide for more  
information about creating and administering ISP application volumes.  
In Figure 1-10, volume vol01has the following characteristics:  
■
■
■
It contains one plex named vol01-01.  
The plex contains one subdisk named disk01-01.  
The subdisk disk01-01is allocated from VM disk disk01.  
Figure 1-10  
Example of a volume with one plex  
vol01  
Volume with one plex  
vol01-01  
disk01-01  
Plex with one subdisk  
vol01-01  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Understanding Veritas Volume Manager 33  
How VxVM handles storage management  
In Figure 1-11 a volume, vol06, with two data plexes is mirrored. Each plex of  
the mirror contains a complete copy of the volume data.  
Figure 1-11  
Example of a volume with two plexes  
vol06  
Volume with two plexes  
vol06-01  
vol06-02  
disk01-01  
vol06-01  
disk02-01  
vol06-02  
Plexes  
Volume vol06has the following characteristics:  
■
■
■
It contains two plexes named vol06-01and vol06-02.  
Each plex contains one subdisk.  
Each subdisk is allocated from a different VM disk (disk01and disk02).  
For more information, see “Mirroring (RAID-1)” on page 42.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
34 Understanding Veritas Volume Manager  
Volume layouts in VxVM  
Volume layouts in VxVM  
A VxVM virtual device is defined by a volume. A volume has a layout defined by  
the association of a volume to one or more plexes, each of which map to  
subdisks. The volume presents a virtual device interface that is exposed to other  
applications for data access. These logical building blocks re-map the volume  
address space through which I/O is re-directed at run-time.  
Different volume layouts each provide different levels of storage service. A  
volume layout can be configured and reconfigured to match particular levels of  
desired storage service.  
Implementation of non-layered volumes  
In a non-layered volume, a subdisk is restricted to mapping directly to a VM disk.  
This allows the subdisk to define a contiguous extent of storage space backed by  
the public region of a VM disk. When active, the VM disk is directly associated  
with an underlying physical disk. The combination of a volume layout and the  
physical disks therefore determines the storage service available from a given  
virtual device.  
Implementation of layered volumes  
A layered volume is constructed by mapping its subdisks to underlying volumes.  
The subdisks in the underlying volumes must map to VM disks, and hence to  
attached physical storage.  
Layered volumes allow for more combinations of logical compositions, some of  
which may be desirable for configuring a virtual device. Because permitting free  
use of layered volumes throughout the command level would have resulted in  
unwieldy administration, some ready-made layered volume configurations are  
designed into VxVM.  
See “Layered volumes” on page 51.  
These ready-made configurations operate with built-in rules to automatically  
match desired levels of service within specified constraints. The automatic  
configuration is done on a “best-effort” basis for the current command  
invocation working against the current configuration.  
To achieve the desired storage service from a set of virtual devices, it may be  
necessary to include an appropriate set of VM disks into a disk group, and to  
execute multiple configuration commands.  
To the extent that it can, VxVM handles initial configuration and on-line re-  
configuration with its set of layouts and administration interface to make this  
job easier and more deterministic.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Understanding Veritas Volume Manager 35  
Volume layouts in VxVM  
Layout methods  
Data in virtual objects is organized to create volumes by using the following  
layout methods:  
■
■
■
■
■
■
Concatenation and spanning  
Striping (RAID-0)  
Mirroring (RAID-1)  
Striping plus mirroring (mirrored-stripe or RAID-0+1)  
Mirroring plus striping (striped-mirror, RAID-1+0 or RAID-10)  
RAID-5 (striping with parity)  
The following sections describe each layout method.  
Concatenation and spanning  
Concatenation maps data in a linear manner onto one or more subdisks in a plex.  
To access all of the data in a concatenated plex sequentially, data is first  
accessed in the first subdisk from beginning to end. Data is then accessed in the  
remaining subdisks sequentially from beginning to end, until the end of the last  
subdisk.  
The subdisks in a concatenated plex do not have to be physically contiguous and  
can belong to more than one VM disk. Concatenation using subdisks that reside  
on more than one VM disk is called spanning.  
Figure 1-12 shows the concatenation of two subdisks from the same VM disk.  
The blocks n, n+1, n+2and n+3(numbered relative to the start of the plex) are  
contiguous on the plex, but actually come from two distinct subdisks on the  
same physical disk.  
The remaining free space in the subdisk, disk01-02, on VM disk, disk01, can  
be put to other uses.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
36 Understanding Veritas Volume Manager  
Volume layouts in VxVM  
Figure 1-12  
Example of concatenation  
Data in  
disk01-01  
Data in  
disk01-03  
n
n+1 n+2 n+3  
Data blocks  
Plex with  
concatenated subdisks  
disk01-01 disk01-03  
disk01-01  
disk01-03  
Subdisks  
disk01-02  
VM disk  
disk01-01  
disk01-03  
disk01  
devname  
Physical disk  
n n+1 n+2  
n+3  
You can use concatenation with multiple subdisks when there is insufficient  
contiguous space for the plex on any one disk. This form of concatenation can be  
used for load balancing between disks, and for head movement optimization on  
a particular disk.  
Figure 1-13 shows data spread over two subdisks in a spanned plex. The blocks  
n, n+1, n+2and n+3(numbered relative to the start of the plex) are contiguous  
on the plex, but actually come from two distinct subdisks from two distinct  
physical disks.  
The remaining free space in the subdisk disk02-02on VM disk disk02can be  
put to other uses.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Understanding Veritas Volume Manager 37  
Volume layouts in VxVM  
Figure 1-13  
Example of spanning  
Data in  
disk01-01  
Data in  
disk02-01  
n
n+1 n+2 n+3  
Data blocks  
Plex with  
concatenated subdisks  
disk01-01 disk02-01  
disk01-01  
disk02-01  
Subdisks  
VM disks  
disk02-02  
disk01-01  
disk01  
disk02-01  
disk02  
devname1  
devname2  
n+3  
Physical disks  
n n+1 n+2  
Caution: Spanning a plex across multiple disks increases the chance that a disk  
failure results in failure of the assigned volume. Use mirroring or RAID-5 (both  
described later) to reduce the risk that a single disk failure results in a volume  
failure.  
See “Creating a volume on any disk” on page 243 for information on how to  
create a concatenated volume that may span several disks.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
38 Understanding Veritas Volume Manager  
Volume layouts in VxVM  
Striping (RAID-0)  
Note: You need a full license to use this feature.  
Striping (RAID-0) is useful if you need large amounts of data written to or read  
from physical disks, and performance is important. Striping is also helpful in  
balancing the I/O load from multi-user applications across multiple disks. By  
using parallel data transfer to and from multiple disks, striping significantly  
improves data-access performance.  
Striping maps data so that the data is interleaved among two or more physical  
disks. A striped plex contains two or more subdisks, spread out over two or more  
physical disks. Data is allocated alternately and evenly to the subdisks of a  
striped plex.  
The subdisks are grouped into “columns,” with each physical disk limited to one  
column. Each column contains one or more subdisks and can be derived from  
one or more physical disks. The number and sizes of subdisks per column can  
vary. Additional subdisks can be added to columns, as necessary.  
Caution: Striping a volume, or splitting a volume across multiple disks, increases  
the chance that a disk failure will result in failure of that volume.  
If five volumes are striped across the same five disks, then failure of any one of  
the five disks will require that all five volumes be restored from a backup. If each  
volume is on a separate disk, only one volume has to be restored. (As an  
alternative to striping, use mirroring or RAID-5 to substantially reduce the  
chance that a single disk failure results in failure of a large number of volumes.)  
Data is allocated in equal-sized units (stripe units) that are interleaved between  
the columns. Each stripe unit is a set of contiguous blocks on a disk. The default  
stripe unit size (or width) is 64 kilobytes.  
For example, if there are three columns in a striped plex and six stripe units,  
data is striped over the three columns, as illustrated in Figure 1-14.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Understanding Veritas Volume Manager 39  
Volume layouts in VxVM  
Figure 1-14  
Striping across three columns  
Column Column Column  
0
1
2
Stripe 1  
su1  
su4  
su2  
su5  
su3  
su6  
Stripe 2  
Subdisk Subdisk Subdisk  
1
2
3
Plex  
SU = stripe unit  
A stripe consists of the set of stripe units at the same positions across all  
columns. In the figure, stripe units 1, 2, and 3 constitute a single stripe.  
Viewed in sequence, the first stripe consists of:  
■
■
■
stripe unit 1 in column 0  
stripe unit 2 in column 1  
stripe unit 3 in column 2  
The second stripe consists of:  
■
■
■
stripe unit 4 in column 0  
stripe unit 5 in column 1  
stripe unit 6 in column 2  
Striping continues for the length of the columns (if all columns are the same  
length), or until the end of the shortest column is reached. Any space remaining  
at the end of subdisks in longer columns becomes unused space.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
40 Understanding Veritas Volume Manager  
Volume layouts in VxVM  
Figure 1-15 shows a striped plex with three equal sized, single-subdisk columns.  
There is one column per physical disk. This example shows three subdisks that  
occupy all of the space on the VM disks. It is also possible for each subdisk in a  
striped plex to occupy only a portion of the VM disk, which leaves free space for  
other disk management tasks.  
Figure 1-15  
Example of a striped plex with one subdisk per column  
su1  
su2  
su3  
su4  
su5  
su6  
Stripe units  
Striped plex  
Column 0  
Column 1  
Column 2  
disk03-01  
disk01-01  
disk02-01  
disk01-01  
disk03-01  
Subdisks  
VM disks  
disk02-01  
disk01-01  
disk01  
disk03-01  
disk03  
disk02-01  
disk02  
devname1  
devname3  
devname2  
Physical disks  
su1 su4  
su2 su5  
su3 su6  
Figure 1-16 illustrates a striped plex with three columns containing subdisks of  
different sizes. Each column contains a different number of subdisks. There is  
one column per physical disk. Striped plexes can be created by using a single  
subdisk from each of the VM disks being striped across. It is also possible to  
allocate space from different regions of the same disk or from another disk (for  
example, if the size of the plex is increased). Columns can also contain subdisks  
from different VM disks.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Understanding Veritas Volume Manager 41  
Volume layouts in VxVM  
Figure 1-16  
Example of a striped plex with concatenated subdisks per column  
su1  
su2  
su3  
su4  
su5  
su6  
Stripe units  
Column 0  
Column 1  
Column 2  
disk03-01  
disk03-02  
disk03-03  
disk02-01  
disk02-02  
Striped plex  
disk01-01  
disk03-01  
disk03-02  
disk03-03  
disk02-01  
disk02-02  
Subdisks  
VM disks  
disk01-01  
disk03-01  
disk03-02  
disk02-01  
disk01-01  
disk01  
disk02-02  
disk02  
disk03-03  
disk03  
devname1  
devname3  
devname2  
Physical disks  
su1 su4  
su2 su5  
su3 su6  
See “Creating a striped volume” on page 253 for information on how to create a  
striped volume.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
42 Understanding Veritas Volume Manager  
Volume layouts in VxVM  
Mirroring (RAID-1)  
Note: You need a full license to use this feature with disks other than the root  
disk.  
Mirroring uses multiple mirrors (plexes) to duplicate the information contained  
in a volume. In the event of a physical disk failure, the plex on the failed disk  
becomes unavailable, but the system continues to operate using the unaffected  
mirrors.  
Note: Although a volume can have a single plex, at least two plexes are required  
to provide redundancy of data. Each of these plexes must contain disk space  
from different disks to achieve redundancy.  
When striping or spanning across a large number of disks, failure of any one of  
those disks can make the entire plex unusable. Because the likelihood of one out  
of several disks failing is reasonably high, you should consider mirroring to  
improve the reliability (and availability) of a striped or spanned volume.  
See “Creating a mirrored volume” on page 249 for information on how to create  
a mirrored volume.  
Disk duplexing, in which each mirror exists on a separate controller, is also  
supported. See “Mirroring across targets, controllers or enclosures” on page 255  
for details.  
Striping plus mirroring (mirrored-stripe or RAID-0+1)  
Note: You need a full license to use this feature.  
VxVM supports the combination of mirroring above striping. The combined  
layout is called a mirrored-stripe layout. A mirrored-stripe layout offers the dual  
benefits of striping to spread data across multiple disks, while mirroring  
provides redundancy of data.  
For mirroring above striping to be effective, the striped plex and its mirrors  
must be allocated from separate disks.  
Figure 1-17 shows an example where two plexes, each striped across three disks,  
are attached as mirrors to the same volume to create a mirrored-stripe volume.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
Understanding Veritas Volume Manager 43  
Volume layouts in VxVM  
Figure 1-17  
Mirrored-stripe volume laid out on six disks  
Column 0  
Column 1  
Column 1  
Column 2  
Column 2  
Striped plex  
Mirror  
Column 0  
Striped plex  
Mirrored-stripe  
volume  
See “Creating a mirrored-stripe volume” on page 254 for information on how to  
create a mirrored-stripe volume.  
The layout type of the data plexes in a mirror can be concatenated or striped.  
Even if only one is striped, the volume is still termed a mirrored-stripe volume.  
If they are all concatenated, the volume is termed a mirrored-concatenated  
volume.  
Mirroring plus striping (striped-mirror, RAID-1+0 or RAID-10)  
Note: You need a full license to use this feature.  
VxVM supports the combination of striping above mirroring. This combined  
layout is called a striped-mirror layout. Putting mirroring below striping mirrors  
each column of the stripe. If there are multiple subdisks per column, each  
subdisk can be mirrored individually instead of each column.  
Note: A striped-mirror volume is an example of a layered volume. See “Layered  
volumes” on page 51 for more information.  
As for a mirrored-stripe volume, a striped-mirror volume offers the dual  
benefits of striping to spread data across multiple disks, while mirroring  
provides redundancy of data. In addition, it enhances redundancy, and reduces  
recovery time after disk failure.  
Figure 1-18 shows an example where a striped-mirror volume is created by  
using each of three existing 2-disk mirrored volumes to form a separate column  
within a striped plex.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
44 Understanding Veritas Volume Manager  
Volume layouts in VxVM  
Figure 1-18  
Striped-mirror volume laid out on six disks  
Underlying mirrored volumes  
Column 0  
Column 0  
Column 1  
Column 2  
Column 2  
Mirror  
Column 1  
Striped plex  
Striped-mirror  
volume  
See “Creating a striped-mirror volume” on page 254 for information on how to  
create a striped-mirrored volume.  
As shown in Figure 1-19, the failure of a disk in a mirrored- stripe layout  
detaches an entire data plex, thereby losing redundancy on the entire volume.  
When the disk is replaced, the entire plex must be brought up to date.  
Recovering the entire plex can take a substantial amount of time. If a disk fails  
in a striped-mirror layout, only the failing subdisk must be detached, and only  
that portion of the volume loses redundancy. When the disk is replaced, only a  
portion of the volume needs to be recovered. Additionally, a mirrored-stripe  
volume is more vulnerable to being put out of use altogether should a second  
disk fail before the first failed disk has been replaced, either manually or by hot-  
relocation.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Understanding Veritas Volume Manager 45  
Volume layouts in VxVM  
Figure 1-19  
How the failure of a single disk affects mirrored-stripe and striped-  
mirror volumes  
Striped plex  
Detached  
striped plex  
X
Mirrored-stripe  
volume with  
no redundancy  
Failure of disk  
detaches plex  
Striped plex  
X
Striped-mirror  
volume with  
Failure of disk removes  
redundancy from a mirror  
partial  
redundancy  
Compared to mirrored-stripe volumes, striped-mirror volumes are more  
tolerant of disk failure, and recovery time is shorter.  
If the layered volume concatenates instead of striping the underlying mirrored  
volumes, the volume is termed a concatenated-mirror volume.  
RAID-5 (striping with parity)  
Note: VxVM supports RAID-5 for private disk groups, but not for shareable disk  
groups in a cluster environment. In addition, VxVM does not support the  
mirroring of RAID-5 volumes that are configured using Veritas Volume  
Manager software. Disk devices that support RAID-5 in hardware may be  
mirrored.  
You need a full license to use this feature.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
46 Understanding Veritas Volume Manager  
Volume layouts in VxVM  
Although both mirroring (RAID-1) and RAID-5 provide redundancy of data, they  
use different methods. Mirroring provides data redundancy by maintaining  
multiple complete copies of the data in a volume. Data being written to a  
mirrored volume is reflected in all copies. If a portion of a mirrored volume fails,  
the system continues to use the other copies of the data.  
RAID-5 provides data redundancy by using parity. Parity is a calculated value  
used to reconstruct data after a failure. While data is being written to a RAID-5  
volume, parity is calculated by doing an exclusive OR (XOR) procedure on the  
data. The resulting parity is then written to the volume. The data and calculated  
parity are contained in a plex that is “striped” across multiple disks. If a portion  
of a RAID-5 volume fails, the data that was on that portion of the failed volume  
can be recreated from the remaining data and parity information. It is also  
possible to mix concatenation and striping in the layout.  
Figure 1-20 shows parity locations in a RAID-5 array configuration. Every stripe  
has a column containing a parity stripe unit and columns containing data. The  
parity is spread over all of the disks in the array, reducing the write time for  
large independent writes because the writes do not have to wait until a single  
parity disk can accept the data.  
Figure 1-20  
Parity locations in a RAID-5 mode  
l
Stripe 1  
Stripe 2  
Stripe 3  
Stripe 4  
Data  
Data  
Data  
Parity  
Data  
Parity  
Data  
Parity  
Data  
Data  
Data  
Parity  
RAID-5 volumes can additionally perform logging to minimize recovery time.  
RAID-5 volumes use RAID-5 logs to keep a copy of the data and parity currently  
being written. RAID-5 logging is optional and can be created along with RAID-5  
volumes or added later.  
The implementation of RAID-5 in VxVM is described in “Veritas Volume  
Manager RAID-5 arrays” on page 47.  
Traditional RAID-5 arrays  
A traditional RAID-5 array is several disks organized in rows and columns. A  
column is a number of disks located in the same ordinal position in the array. A  
row is the minimal number of disks necessary to support the full width of a  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Understanding Veritas Volume Manager 47  
Volume layouts in VxVM  
parity stripe. Figure 1-21 shows the row and column arrangement of a  
traditional RAID-5 array.  
Figure 1-21  
Traditional RAID-5 array  
Stripe 1  
Stripe 3  
Row 0  
Row 1  
Stripe 2  
Column 0  
Column 1  
Column 2  
Column 3  
This traditional array structure supports growth by adding more rows per  
column. Striping is accomplished by applying the first stripe across the disks in  
Row 0, then the second stripe across the disks in Row 1, then the third stripe  
across the Row 0 disks, and so on. This type of array requires all disks columns,  
and rows to be of equal size.  
Veritas Volume Manager RAID-5 arrays  
The RAID-5 array structure in Veritas Volume Manager differs from the  
traditional structure. Due to the virtual nature of its disks and other objects,  
VxVM does not use rows. Instead, VxVM uses columns consisting of variable  
length subdisks as shown in Figure 1-22. Each subdisk represents a specific area  
of a disk.  
VxVM allows each column of a RAID-5 plex to consist of a different number of  
subdisks. The subdisks in a given column can be derived from different physical  
disks. Additional subdisks can be added to the columns as necessary. Striping is  
implemented by applying the first stripe across each subdisk at the top of each  
column, then applying another stripe below that, and so on for the length of the  
columns. Equal-sized stripe units are used for each column. For RAID-5, the  
default stripe unit size is 16 kilobytes. See “Striping (RAID-0)” on page 38 for  
further information about stripe units.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
48 Understanding Veritas Volume Manager  
Volume layouts in VxVM  
Figure 1-22  
Veritas Volume Manager RAID-5 array  
Stripe 1  
Stripe 2  
SD  
SD  
SD  
SD  
SD  
SD  
SD  
SD  
Column 0  
Column 1  
Column 2  
Column 3  
SD = subdisk  
Note: Mirroring of RAID-5 volumes is not supported.  
See “Creating a RAID-5 volume” on page 256 for information on how to create a  
RAID-5 volume.  
Left-symmetric layout  
There are several layouts for data and parity that can be used in the setup of a  
RAID-5 array. The implementation of RAID-5 in VxVM uses a left-symmetric  
layout. This provides optimal performance for both random I/O operations and  
large sequential I/O operations. However, the layout selection is not as critical  
for performance as are the number of columns and the stripe unit size.  
Left-symmetric layout stripes both data and parity across columns, placing the  
parity in a different column for every stripe of data. The first parity stripe unit is  
located in the rightmost column of the first stripe. Each successive parity stripe  
unit is located in the next stripe, shifted left one column from the previous  
parity stripe unit location. If there are more stripes than columns, the parity  
stripe unit placement begins in the rightmost column again.  
Figure 1-23 shows a left-symmetric parity layout with five disks (one per  
column).  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Understanding Veritas Volume Manager 49  
Volume layouts in VxVM  
Figure 1-23  
Left-symmetric layout  
Column  
Parity stripe unit  
0
1
2
7
3
P1  
8
P0  
4
Stripe  
5
6
10  
15  
P4  
11  
P3  
16  
P2  
12  
17  
9
13  
18  
14  
19  
(Data)  
stripe unit  
For each stripe, data is organized starting to the right of the parity stripe unit. In  
the figure, data organization for the first stripe begins at P0 and continues to  
stripe units 0-3. Data organization for the second stripe begins at P1, then  
continues to stripe unit 4, and on to stripe units 5-7. Data organization proceeds  
in this manner for the remaining stripes.  
Each parity stripe unit contains the result of an exclusive OR (XOR) operation  
performed on the data in the data stripe units within the same stripe. If one  
column’s data is inaccessible due to hardware or software failure, the data for  
each stripe can be restored by XORing the contents of the remaining columns  
data stripe units against their respective parity stripe units.  
For example, if a disk corresponding to the whole or part of the far left column  
fails, the volume is placed in a degraded mode. While in degraded mode, the data  
from the failed column can be recreated by XORing stripe units 1-3 against  
parity stripe unit P0 to recreate stripe unit 0, then XORing stripe units 4, 6, and  
7 against parity stripe unit P1 to recreate stripe unit 5, and so on.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
50 Understanding Veritas Volume Manager  
Volume layouts in VxVM  
Note: Failure of more than one column in a RAID-5 plex detaches the volume.  
The volume is no longer allowed to satisfy read or write requests. Once the failed  
columns have been recovered, it may be necessary to recover user data from  
backups.  
RAID-5 logging  
Logging is used to prevent corruption of data during recovery by immediately  
recording changes to data and parity to a log area on a persistent device such as  
a volume on disk or in non-volatile RAM. The new data and parity are then  
written to the disks.  
Without logging, it is possible for data not involved in any active writes to be  
lost or silently corrupted if both a disk in a RAID-5 volume and the system fail. If  
this double-failure occurs, there is no way of knowing if the data being written to  
the data portions of the disks or the parity being written to the parity portions  
have actually been written. Therefore, the recovery of the corrupted disk may be  
corrupted itself.  
Figure 1-24 illustrates a RAID-5 volume configured across three disks (A, B and  
C). In this volume, recovery of disk B’s corrupted data depends on disk A’s data  
and disk C’s parity both being complete. However, only the data write to disk A is  
complete. The parity write to disk C is incomplete, which would cause the data  
on disk B to be reconstructed incorrectly.  
Figure 1-24  
Incomplete write to a RAID-5 volume  
Completed  
data write  
Incomplete  
parity write  
Corrupted data  
Disk B  
Disk A  
Disk C  
This failure can be avoided by logging all data and parity writes before  
committing them to the array. In this way, the log can be replayed, causing the  
data and parity updates to be completed before the reconstruction of the failed  
drive takes place.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Understanding Veritas Volume Manager 51  
Volume layouts in VxVM  
Logs are associated with a RAID-5 volume by being attached as log plexes. More  
than one log plex can exist for each RAID-5 volume, in which case the log areas  
are mirrored.  
See “Adding a RAID-5 log” on page 283 for information on how to add a RAID-5  
log to a RAID-5 volume.  
Layered volumes  
A layered volume is a virtual Veritas Volume Manager object that is built on top  
of other volumes. The layered volume structure tolerates failure better and has  
greater redundancy than the standard volume structure. For example, in a  
striped-mirror layered volume, each mirror (plex) covers a smaller area of  
storage space, so recovery is quicker than with a standard mirrored volume.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
52 Understanding Veritas Volume Manager  
Volume layouts in VxVM  
Figure 1-25  
Example of a striped-mirror layered volume  
vol01  
Striped-mirror  
volume  
vol01-01  
vol01-01  
Column 0 Column 1  
Striped plex  
Managed  
by User  
Managed  
by VxVM  
vop01  
vop01  
vop02  
vop02  
Subdisks  
Underlying  
mirrored  
volumes  
Concatenated  
plexes  
disk05-01  
disk05-01  
disk06-01  
disk06-01  
disk04-01  
disk04-01  
disk07-01  
disk07-01  
Subdisks on  
VM disks  
Figure 1-25 illustrates the structure of a typical layered volume. It shows  
subdisks with two columns, built on underlying volumes with each volume  
internally mirrored. The volume and striped plex in the “Managed by User” area  
allow you to perform normal tasks in VxVM. User tasks can be performed only  
on the top-level volume of a layered volume.  
Underlying volumes in the “Managed by VxVM” area are used exclusively by  
VxVM and are not designed for user manipulation. You cannot detach a layered  
volume or perform any other operation on the underlying volumes by  
manipulating the internal structure. You can perform all necessary operations  
in the “Managed by User” area that includes the top-level volume and striped  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Understanding Veritas Volume Manager 53  
Volume layouts in VxVM  
plex (for example, resizing the volume, changing the column width, or adding a  
column).  
System administrators can manipulate the layered volume structure for  
troubleshooting or other operations (for example, to place data on specific  
disks). Layered volumes are used by VxVM to perform the following tasks and  
operations:  
■
■
■
Creating striped-mirrors. (See “Creating a striped-mirror volume” on  
page 254, and the vxassist(1M) manual page.)  
Creating concatenated-mirrors. (See “Creating a concatenated-mirror  
volume” on page 249, and the vxassist(1M) manual page.)  
Online Relayout. (See “Online relayout” on page 54, and the  
vxrelayout(1M) and vxassist(1M) manual pages.)  
■
■
RAID-5 subdisk moves. (See the vxsd(1M) manual page.)  
Snapshots. (See “Administering volume snapshots” on page 303, and the  
vxsnap(1M) and vxassist(1M) manual pages.)  
Download from Www.Somanuals.com. All Manuals Search And Download.  
54 Understanding Veritas Volume Manager  
Online relayout  
Online relayout  
Note: You need a full license to use this feature.  
Online relayout allows you to convert between storage layouts in VxVM, with  
uninterrupted data access. Typically, you would do this to change the  
redundancy or performance characteristics of a volume. VxVM adds  
redundancy to storage either by duplicating the data (mirroring) or by adding  
parity (RAID-5). Performance characteristics of storage in VxVM can be changed  
by changing the striping parameters, which are the number of columns and the  
stripe width.  
See “Performing online relayout” on page 294 for details of how to perform  
online relayout of volumes in VxVM. Also see “Converting between layered and  
non-layered volumes” on page 300 for information about the additional volume  
conversion operations that are possible.  
How online relayout works  
Online relayout allows you to change the storage layouts that you have already  
created in place without disturbing data access. You can change the  
performance characteristics of a particular layout to suit your changed  
requirements. You can transform one layout to another by invoking a single  
command.  
For example, if a striped layout with a 128KB stripe unit size is not providing  
optimal performance, you can use relayout to change the stripe unit size.  
File systems mounted on the volumes do not need to be unmounted to achieve  
this transformation provided that the file system (such as Veritas File System)  
supports online shrink and grow operations.  
Online relayout reuses the existing storage space and has space allocation  
policies to address the needs of the new layout. The layout transformation  
process converts a given volume to the destination layout by using minimal  
temporary space that is available in the disk group.  
The transformation is done by moving one portion of data at a time in the source  
layout to the destination layout. Data is copied from the source volume to the  
temporary area, and data is removed from the source volume storage area in  
portions. The source volume storage area is then transformed to the new layout,  
and the data saved in the temporary area is written back to the new layout. This  
operation is repeated until all the storage and data in the source volume has  
been transformed to the new layout.  
The default size of the temporary area used during the relayout depends on the  
size of the volume and the type of relayout. For volumes larger than 50MB, the  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Understanding Veritas Volume Manager 55  
Online relayout  
amount of temporary space that is required is usually 10% of the size of the  
volume, from a minimum of 50MB up to a maximum of 1GB. For volumes  
smaller than 50MB, the temporary space required is the same as the size of the  
volume.  
The following error message displays the number of blocks required if there is  
insufficient free space available in the disk group for the temporary area:  
tmpsize too small to perform this relayout (nblks minimum  
required)  
You can override the default size used for the temporary area by using the  
tmpsizeattribute to vxassist. See the vxassist(1M) manual page for more  
information.  
As well as the temporary area, space is required for a temporary intermediate  
volume when increasing the column length of a striped volume. The amount of  
space required is the difference between the column lengths of the target and  
source volumes. For example, 20GB of temporary additional space is required to  
relayout a 150GB striped volume with 5 columns of length 30GB as 3 columns of  
length 50GB. In some cases, the amount of temporary space that is required is  
relatively large. For example, a relayout of a 150GB striped volume with 5  
columns as a concatenated volume (with effectively one column) requires 120GB  
of space for the intermediate volume.  
Additional permanent disk space may be required for the destination volumes,  
depending on the type of relayout that you are performing. This may happen,  
for example, if you change the number of columns in a striped volume. Figure 1-  
26 shows how decreasing the number of columns can require disks to be added  
to a volume. The size of the volume remains the same but an extra disk is needed  
to extend one of the columns.  
Figure 1-26  
Example of decreasing the number of columns in a volume  
Five columns of length L  
Three columns of length 5L/3  
The following are examples of operations that you can perform using online  
relayout:  
■
Change a RAID-5 volume to a concatenated, striped, or layered volume  
(remove parity). See Figure 1-27 for an example. Note that removing parity  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
56 Understanding Veritas Volume Manager  
Online relayout  
(shown by the shaded area) decreases the overall storage space that the  
volume requires.  
Figure 1-27  
Example of relayout of a RAID-5 volume to a striped volume  
RAID-5 volume  
Striped volume  
■
Change a volume to a RAID-5 volume (add parity). See Figure 1-28 for an  
example. Note that adding parity (shown by the shaded area) increases the  
overall storage space that the volume requires.  
Figure 1-28  
Example of relayout of a concatenated volume to a RAID-5 volume  
Concatenated  
volume  
RAID-5 volume  
■
Change the number of columns in a volume. See Figure 1-29 for an example.  
Note that the length of the columns is reduced to conserve the size of the  
volume.  
Figure 1-29  
Example of increasing the number of columns in a volume  
Two columns  
Three columns  
■
Change the column stripe width in a volume. See Figure 1-30 for an  
example.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Understanding Veritas Volume Manager 57  
Online relayout  
Figure 1-30  
Example of increasing the stripe width for the columns in a volume  
For details of how to perform online relayout operations, see “Performing online  
relayout” on page 294. For information about the relayout transformations that  
are possible, see “Permitted relayout transformations” on page 295.  
Limitations of online relayout  
Note the following limitations of online relayout:  
■
■
Log plexes cannot be transformed.  
Volume snapshots cannot be taken when there is an online relayout  
operation running on the volume.  
■
Online relayout cannot create a non-layered mirrored volume in a single  
step. It always creates a layered mirrored volume even if you specify a non-  
layered mirrored layout, such as mirror-stripeor mirror-concat. Use  
the vxassistconvertcommand to turn the layered mirrored volume that  
results from a relayout into a non-layered volume. See “Converting between  
layered and non-layered volumes” on page 300 for more information.  
■
■
Online relayout can be used only with volumes that have been created using  
the vxassistcommand or the Veritas Enterprise Administrator (VEA).  
The usual restrictions apply for the minimum number of physical disks that  
are required to create the destination layout. For example, mirrored volumes  
require at least as many disks as mirrors, striped and RAID-5 volumes  
require at least as many disks as columns, and striped-mirror volumes  
require at least as many disks as columns multiplied by mirrors.  
■
To be eligible for layout transformation, the plexes in a mirrored volume  
must have identical stripe widths and numbers of columns. Relayout is not  
possible unless you make the layouts of the individual plexes identical.  
■
■
Online relayout involving RAID-5 volumes is not supported for shareable  
disk groups in a cluster environment.  
Online relayout cannot transform sparse plexes, nor can it make any plex  
sparse. (A sparse plex is not the same size as the volume, or has regions that  
are not mapped to any subdisk.)  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
58 Understanding Veritas Volume Manager  
Online relayout  
■
■
The number of mirrors in a mirrored volume cannot be changed using  
relayout.  
Only one relayout may be applied to a volume at a time.  
Transformation characteristics  
Transformation of data from one layout to another involves rearrangement of  
data in the existing layout to the new layout. During the transformation, online  
relayout retains data redundancy by mirroring any temporary space used. Read  
and write access to data is not interrupted during the transformation.  
Data is not corrupted if the system fails during a transformation. The  
transformation continues after the system is restored and both read and write  
access are maintained.  
You can reverse the layout transformation process at any time, but the data may  
not be returned to the exact previous storage location. Any existing  
transformation in the volume must be stopped before doing a reversal.  
You can determine the transformation direction by using the vxrelayout  
statusvolume command.  
These transformations are protected against I/O failures if there is sufficient  
redundancy and space to move the data.  
Transformations and volume length  
Some layout transformations can cause the volume length to increase or  
decrease. If either of these conditions occurs, online relayout uses the  
vxresize(1M) command to shrink or grow a file system as described in  
“Resizing a volume” on page 284.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Understanding Veritas Volume Manager 59  
Volume resynchronization  
Volume resynchronization  
When storing data redundantly and using mirrored or RAID-5 volumes, VxVM  
ensures that all copies of the data match exactly. However, under certain  
conditions (usually due to complete system failures), some redundant data on a  
volume can become inconsistent or unsynchronized. The mirrored data is not  
exactly the same as the original data. Except for normal configuration changes  
(such as detaching and reattaching a plex), this can only occur when a system  
crashes while data is being written to a volume.  
Data is written to the mirrors of a volume in parallel, as is the data and parity in  
a RAID-5 volume. If a system crash occurs before all the individual writes  
complete, it is possible for some writes to complete while others do not. This can  
result in the data becoming unsynchronized. For mirrored volumes, it can cause  
two reads from the same region of the volume to return different results, if  
different mirrors are used to satisfy the read request. In the case of RAID-5  
volumes, it can lead to parity corruption and incorrect data reconstruction.  
VxVM needs to ensure that all mirrors contain exactly the same data and that  
the data and parity in RAID-5 volumes agree. This process is called volume  
resynchronization. For volumes that are part of the disk group that is  
automatically imported at boot time (usually aliased as the reserved system-  
wide disk group, bootdg), the resynchronization process takes place when the  
system reboots.  
Not all volumes require resynchronization after a system failure. Volumes that  
were never written or that were quiescent (that is, had no active I/O) when the  
system failure occurred could not have had outstanding writes and do not  
require resynchronization.  
Dirty flags  
VxVM records when a volume is first written to and marks it as dirty. When a  
volume is closed by all processes or stopped cleanly by the administrator, and all  
writes have been completed, VxVM removes the dirty flag for the volume. Only  
volumes that are marked dirty when the system reboots require  
resynchronization.  
Resynchronization process  
The process of resynchronization depends on the type of volume. RAID-5  
volumes that contain RAID-5 logs can “replay” those logs. If no logs are  
available, the volume is placed in reconstruct-recovery mode and all parity is  
regenerated. For mirrored volumes, resynchronization is done by placing the  
volume in recovery mode (also called read-writeback recovery mode).  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
60 Understanding Veritas Volume Manager  
Dirty region logging  
Resynchronization of data in the volume is done in the background. This allows  
the volume to be available for use while recovery is taking place.  
The process of resynchronization can impact system performance. The recovery  
process reduces some of this impact by spreading the recoveries to avoid  
stressing a specific disk or controller.  
For large volumes or for a large number of volumes, the resynchronization  
process can take time. These effects can be addressed by using dirty region  
logging (DRL) and FastResync (fast mirror resynchronization) for mirrored  
volumes, or by ensuring that RAID-5 volumes have valid RAID-5 logs. See the  
sections “Dirty region logging” on page 60 and “FastResync” on page 66 for  
more information.  
For raw volumes used by database applications, the SmartSync feature can be  
used if this is supported by the database vendor (see “SmartSync recovery  
accelerator” on page 62).  
Dirty region logging  
Note: If a version 20 DCO volume is associated with a volume, a portion of the  
DCO volume can be used to store the DRL log. There is no need to create a  
separate DRL log for a volume which has a version 20 DCO volume. For more  
information, see “DCO volume versioning” on page 68.  
You need a full license to use this feature.  
Dirty region logging (DRL), if enabled, speeds recovery of mirrored volumes  
after a system crash. DRL keeps track of the regions that have changed due to  
I/O writes to a mirrored volume. DRL uses this information to recover only those  
portions of the volume that need to be recovered.  
If DRL is not used and a system failure occurs, all mirrors of the volumes must  
be restored to a consistent state. Restoration is done by copying the full contents  
of the volume between its mirrors. This process can be lengthy and I/O  
intensive. It may also be necessary to recover the areas of volumes that are  
already consistent.  
Dirty region logs  
DRL logically divides a volume into a set of consecutive regions, and maintains a  
log on disk where each region is represented by a status bit. This log records  
regions of a volume for which writes are pending. Before data is written to a  
region, DRL synchronously marks the corresponding status bit in the log as  
dirty. To enhance performance, the log bit remains set to dirty until the region  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                 
Understanding Veritas Volume Manager 61  
Dirty region logging  
becomes the least recently accessed for writes. This allows writes to the same  
region to be written immediately to disk if the region’s log bit is set to dirty.  
On restarting a system after a crash, VxVM recovers only those regions of the  
volume that are marked as dirty in the dirty region log.  
Log subdisks and plexes  
DRL log subdisks store the dirty region log of a mirrored volume that has DRL  
enabled. A volume with DRL has at least one log subdisk; multiple log subdisks  
can be used to mirror the dirty region log. Each log subdisk is associated with  
one plex of the volume. Only one log subdisk can exist per plex. If the plex  
contains only a log subdisk and no data subdisks, that plex is referred to as a log  
plex.  
The log subdisk can also be associated with a regular plex that contains data  
subdisks. In that case, the log subdisk risks becoming unavailable if the plex  
must be detached due to the failure of one of its data subdisks.  
If the vxassistcommand is used to create a dirty region log, it creates a log plex  
containing a single log subdisk by default. A dirty region log can also be set up  
manually by creating a log subdisk and associating it with a plex. The plex then  
contains both a log and data subdisks.  
Sequential DRL  
Some volumes, such as those that are used for database replay logs, are written  
sequentially and do not benefit from delayed cleaning of the DRL bits. For these  
volumes, sequential DRL can be used to limit the number of dirty regions. This  
allows for faster recovery should a crash occur. However, if applied to volumes  
that are written to randomly, sequential DRL can be a performance bottleneck  
as it limits the number of parallel writes that can be carried out.  
The maximum number of dirty regions allowed for sequential DRL is controlled  
by a tunable as detailed in the description of voldrl_max_seq_dirtyin  
“Tunable parameters” on page 475.  
Note: DRL adds a small I/O overhead for most write access patterns.  
For details of how to configure DRL and sequential DRL, see “Adding traditional  
DRL logging to a mirrored volume” on page 281, and “Preparing a volume for  
DRL and instant snapshots” on page 275.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
62 Understanding Veritas Volume Manager  
Dirty region logging  
SmartSync recovery accelerator  
The SmartSync feature of Veritas Volume Manager increases the availability of  
mirrored volumes by only resynchronizing changed data. (The process of  
resynchronizing mirrored databases is also sometimes referred to as  
resilvering.) SmartSync reduces the time required to restore consistency, freeing  
more I/O bandwidth for business-critical applications. If supported by the  
database vendor, the SmartSync feature uses an extended interface between  
VxVM volumes and the database software to avoid unnecessary work during  
®
mirror resynchronization. For example, Oracle automatically takes advantage  
of SmartSync to perform database resynchronization when it is available.  
Note: The SmartSync feature of Veritas Volume Manager is only applicable to  
databases that are configured on raw volumes. You cannot use it with volumes  
that contain file systems. Use an alternative solution such as the Oracle  
Resilvering feature of Veritas File System (VxFS).  
You must configure volumes correctly to use SmartSync. For VxVM, there are  
two types of volumes used by the database, as follows:  
■
Data volumes are all other volumes used by the database (control files and  
tablespace files).  
■
Redo log volumes contain redo logs of the database.  
SmartSync works with these two types of volumes differently, so they must be  
configured as described in the following sections.  
To enable the use of SmartSync with database volumes in shared disk groups,  
set the value of the volcvm_smartsynctunable to 1. For a description of  
volcvm_smartsync, see “Tunable parameters” on page 475.  
Data volume configuration  
The recovery takes place when the database software is started, not at system  
startup. This reduces the overall impact of recovery when the system reboots.  
Because the recovery is controlled by the database, the recovery time for the  
volume is the resilvering time for the database (that is, the time required to  
replay the redo logs).  
Because the database keeps its own logs, it is not necessary for VxVM to do  
logging. Data volumes should be configured as mirrored volumes without dirty  
region logs. In addition to improving recovery time, this avoids any run-time I/O  
overhead due to DRL, and improves normal database write access.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Understanding Veritas Volume Manager 63  
Volume snapshots  
Redo log volume configuration  
A redo log is a log of changes to the database data. Because the database does not  
maintain changes to the redo logs, it cannot provide information about which  
sections require resilvering. Redo logs are also written sequentially, and since  
traditional dirty region logs are most useful with randomly-written data, they  
are of minimal use for reducing recovery time for redo logs. However, VxVM can  
reduce the number of dirty regions by modifying the behavior of its dirty region  
logging feature to take advantage of sequential access patterns. Sequential DRL  
decreases the amount of data needing recovery and reduces recovery time  
impact on the system.  
The enhanced interfaces for redo logs allow the database software to inform  
VxVM when a volume is to be used as a redo log. This allows VxVM to modify the  
DRL behavior of the volume to take advantage of the access patterns. Since the  
improved recovery time depends on dirty region logs, redo log volumes should  
be configured as mirrored volumes with sequential DRL.  
For additional information, see “Sequential DRL” on page 61.  
Volume snapshots  
Veritas Volume Manager provides the capability for taking an image of a volume  
at a given point in time. Such an image is referred to as a volume snapshot. Such  
snapshots should not be confused with file system snapshots, which are point-  
in-time images of a Veritas File System.  
Figure 1-31 illustrates how a snapshot volume represents a copy of an original  
volume at a given point in time. Even though the contents of the original volume  
can change, the snapshot volume can be used to preserve the contents of the  
original volume as they existed at an earlier time.  
The snapshot volume provides a stable and independent base for making  
backups of the contents of the original volume, or for other applications such as  
decision support. In the figure, the contents of the snapshot volume are  
eventually resynchronized with the original volume at a later point in time.  
Another possibility is to use the snapshot volume to restore the contents of the  
original volume. This may be useful if the contents of the original volume have  
become corrupted in some way.  
Note: If you choose to write to the snapshot volume, it may no longer be suitable  
for use in restoring the contents of the original volume.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
64 Understanding Veritas Volume Manager  
Volume snapshots  
Figure 1-31  
Volume snapshot as a point-in-time image of a volume  
Original  
volume  
T1  
T2  
T3  
Snapshot volume is  
created at time T2  
Original  
volume  
Snapshot  
volume  
Snapshot volume retains  
image taken at time T2  
Original  
volume  
Snapshot  
volume  
Snapshot volume is  
updated at time T4  
Original  
volume  
Snapshot  
volume  
T4  
Time  
Resynchronize  
snapshot volume  
from original volume  
The traditional type of volume snapshot in VxVM is of the third-mirror break-off  
type. This name comes from its implementation where a snapshot plex (or third  
mirror) is added to a mirrored volume. The contents of the snapshot plex are  
then synchronized from the original plexes of the volume. When this  
synchronization is complete, the snapshot plex can be detached as a snapshot  
volume for use in backup or decision support applications. At a later time, the  
snapshot plex can be reattached to the original volume, requiring a full  
resynchronization of the snapshot plex’s contents. For more information about  
this type of snapshot, see “Traditional third-mirror break-off snapshots” on  
page 305.  
The FastResync feature was introduced to track writes to the original volume.  
This tracking means that only a partial, and therefore much faster,  
resynchronization is required on reattaching the snapshot plex. In later  
releases, the snapshot model was enhanced to allow snapshot volumes to  
contain more than a single plex, reattachment of a subset of a snapshot  
volume’s plexes, and persistence of FastResync across system reboots or cluster  
restarts.  
For more information about FastResync, see “FastResync” on page 66.  
Release 4.0 of VxVM introduced full-sized instant snapshots and space-  
optimized instant snapshots, which offer advantages over traditional third-  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Understanding Veritas Volume Manager 65  
Volume snapshots  
mirror snapshots such as immediate availability and easier configuration and  
administration. You can also use the third-mirror break-off usage model with  
full-sized snapshots, where this is necessary for write-intensive applications.  
For more information, see the following sections:  
■
■
■
■
“Full-sized instant snapshots” on page 307.  
“Space-optimized instant snapshots” on page 309.  
“Emulation of third-mirror break-off snapshots” on page 310.  
“Linked break-off snapshot volumes” on page 311.  
“Comparison of snapshot features” on page 65 compares the features that are  
supported by the different types of snapshot.  
For more information about taking snapshots of a volume, see “Administering  
volume snapshots” on page 303, and the vxsnap(1M) and vxassist(1M) manual  
pages.  
Comparison of snapshot features  
The table, “Comparison of snapshot features for supported snapshot types” on  
page 65, compares the features of the various types of snapshots that are  
supported in VxVM.  
Full-sized instant snapshots are easier to configure and offer more flexibility of  
use than do traditional third-mirror break-off snapshots. For preference, new  
volumes should be configured to use snapshots that have been created using the  
vxsnapcommand rather than using the vxassistcommand. Legacy volumes  
can also be reconfigured to use vxsnapsnapshots, but this requires rewriting of  
administration scripts that assume the vxassistsnapshot model.  
If storage space is at a premium, space-optimized instant snapshots can be  
configured with some reduction of supported functionality. For example, space-  
optimized snapshots cannot be turned into independent volumes, nor can they  
be moved into a separate disk group for off-host processing.  
Table 1-1  
Comparison of snapshot features for supported snapshot types  
Snapshot feature  
Full-sized  
Space-  
Break-off  
instant (vxsnap) optimized  
(vxassist or  
instant (vxsnap) vxsnap)  
Yes  
No  
Yes  
No  
No  
Immediately available for  
use on creation  
Yes  
Requires less storage  
space than original  
volume  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
66 Understanding Veritas Volume Manager  
FastResync  
Table 1-1  
Comparison of snapshot features for supported snapshot types  
Snapshot feature  
Full-sized  
Space-  
Break-off  
instant (vxsnap) optimized  
(vxassist or  
instant (vxsnap) vxsnap)  
Yes  
Yes  
No  
Yes  
Yes  
Can be reattached to  
original volume  
Yes  
Can be used to restore  
contents of original  
volume  
Yes  
Yes  
Yes  
Yes  
No  
No  
No  
No  
Yes  
Can quickly be refreshed  
without being reattached  
Snapshot hierarchy can  
be split  
Can be moved into  
separate disk group from  
original volume  
Yes  
Yes  
No  
Yes  
Yes  
Can be turned into an  
independent volume  
Yes  
FastResync ability  
persists across system  
reboots or cluster  
restarts  
Yes  
No  
No  
Synchronization can be  
controlled  
FastResync  
Note: You need a Veritas FlashSnap or FastResync license to use this feature.  
The FastResync feature (previously called Fast Mirror Resynchronization or  
FMR) performs quick and efficient resynchronization of stale mirrors (a mirror  
that is not synchronized). This increases the efficiency of the VxVM snapshot  
mechanism, and improves the performance of operations such as backup and  
decision support applications. Typically, these operations require that the  
volume is quiescent, and that they are not impeded by updates to the volume by  
other activities on the system. To achieve these goals, the snapshot mechanism  
in VxVM creates an exact copy of a primary volume at an instant in time. After a  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Understanding Veritas Volume Manager 67  
FastResync  
snapshot is taken, it can be accessed independently of the volume from which it  
was taken. In a clustered VxVM environment with shared access to storage, it is  
possible to eliminate the resource contention and performance overhead of  
using a snapshot simply by accessing it from a different node.  
For details of how to enable FastResync on a per-volume basis, see “Enabling  
FastResync on a volume” on page 292.  
FastResync enhancements  
FastResync provides two fundamental enhancements to VxVM:  
■
FastResync optimizes mirror resynchronization by keeping track of updates  
to stored data that have been missed by a mirror. (A mirror may be  
unavailable because it has been detached from its volume, either  
automatically by VxVM as the result of an error, or directly by an  
administrator using a utility such as vxplexor vxassist. A returning  
mirror is a mirror that was previously detached and is in the process of  
being re-attached to its original volume as the result of the vxrecoveror  
vxplexattoperation.) When a mirror returns to service, only the updates  
that it has missed need to be re-applied to resynchronize it. This requires  
much less effort than the traditional method of copying all the stored data  
to the returning mirror.  
Once FastResync has been enabled on a volume, it does not alter how you  
administer mirrors. The only visible effect is that repair operations  
conclude more quickly.  
■
FastResync allows you to refresh and re-use snapshots rather than discard  
them. You can quickly re-associate (snapback) snapshot plexes with their  
original volumes. This reduces the system overhead required to perform  
cyclical operations such as backups that rely on the snapshot functionality  
of VxVM.  
Non-persistent FastResync  
Non-persistent FastResync allocates its change maps in memory. If non-  
persistent FastResync is enabled, a separate FastResync map is kept for the  
original volume and for each snapshot volume. Unlike a dirty region log (DRL),  
they do not reside on disk nor in persistent store. This has the advantage that  
updates to the FastResync map have little impact on I/O performance, as no disk  
updates needed to be performed. However, if a system is rebooted, the  
information in the map is lost, so a full resynchronization is required on  
snapback. This limitation can be overcome for volumes in cluster-shareable disk  
groups, provided that at least one of the nodes in the cluster remained running  
to preserve the FastResync map in its memory. However, a node crash in a High  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
68 Understanding Veritas Volume Manager  
FastResync  
Availability (HA) environment requires the full resynchronization of a mirror  
when it is reattached to its parent volume.  
How non-persistent FastResync works with snapshots  
The snapshot feature of VxVM takes advantage of FastResync change tracking  
to record updates to the original volume after a snapshot plex is created. After a  
snapshot is taken, the snapbackoption is used to reattach the snapshot plex.  
Provided that FastResync is enabled on a volume before the snapshot is taken,  
and that it is not disabled at any time before the snapshot is reattached, the  
changes that FastResync records are used to resynchronize the volume during  
the snapback. This considerably reduces the time needed to resynchronize the  
volume.  
Non-Persistent FastResync uses a map in memory to implement change  
tracking. Each bit in the map represents a contiguous number of blocks in a  
volume’s address space. The default size of the map is 4 blocks. The kernel  
tunable vol_fmr_logszcan be used to limit the maximum size in blocks of the  
map as described on “Tunable parameters” on page 475.  
Persistent FastResync  
Unlike non-persistent FastResync, persistent FastResync keeps the FastResync  
maps on disk so that they can survive system reboots, system crashes and  
cluster crashes. Persistent FastResync can also track the association between  
volumes and their snapshot volumes after they are moved into different disk  
groups. When the disk groups are rejoined, this allows the snapshot plexes to be  
quickly resynchronized. This ability is not supported by non-persistent  
FastResync. See “Reorganizing the contents of disk groups” on page 195 for  
details.  
If persistent FastResync is enabled on a volume or on a snapshot volume, a data  
change object (DCO) and a DCO volume are associated with the volume.  
DCO volume versioning  
The internal layout of the DCO volume changed in VxVM 4.0 to support new  
features such as full-sized and space-optimized instant snapshots. Because the  
DCO volume layout is versioned, VxVM software continues to support the  
version 0 layout for legacy volumes. However, you must configure a volume to  
have a version 20 DCO volume if you want to take instant snapshots of the  
volume. Future releases of Veritas Volume Manager may introduce new versions  
of the DCO volume layout.  
See “Determining the DCO version number” on page 277 for a description of  
how to find out the version number of a DCO that is associated with a volume.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Understanding Veritas Volume Manager 69  
FastResync  
Version 0 DCO volume layout  
In VxVM releases 3.2 and 3.5, the DCO object only managed information about  
the FastResync maps. These maps track writes to the original volume and to  
each of up to 32 snapshot volumes since the last snapshotoperation. Each plex  
of the DCO volume on disk holds 33 maps, each of which is 4 blocks in size by  
default.  
Persistent FastResync uses the maps in a version 0 DCO volume on disk to  
implement change tracking. As for non-persistent FastResync, each bit in the  
map represents a region (a contiguous number of blocks) in a volume’s address  
space. The size of each map can be changed by specifying the dcolenattribute  
to the vxassistcommand when the volume is created. The default value of  
dcolenis 132 1024-byte blocks (the plex contains 33 maps, each of length 4  
blocks). To use a larger map size, multiply the desired map size by 33 to calculate  
the value of dcolenthat you need to specify. For example, to use an 8-block  
map, you would specify dcolen=264. The maximum possible map size is 64  
blocks, which corresponds to a dcolenvalue of 2112 blocks.  
Note: The size of a DCO plex is rounded up to the nearest integer multiple of the  
disk group alignment value. The alignment value is 8KB for disk groups that  
support the Cross-platform Data Sharing (CDS) feature. Otherwise, the  
alignment value is 1 block.  
Only traditional (third-mirror) volume snapshots that are administered using  
the vxassistcommand are supported for the version 0 DCO volume layout.  
Full-sized and space-optimized instant snapshots are not supported.  
Version 20 DCO volume layout  
In VxVM 4.0 and later releases, the DCO object is used not only to manage the  
FastResync maps, but also to manage DRL recovery maps (see “Dirty region  
logging” on page 60) and special maps called copymaps that allow instant  
snapshot operations to resume correctly following a system crash.  
Each bit in a map represents a region (a contiguous number of blocks) in a  
volume’s address space. A region represents the smallest portion of a volume for  
which changes are recorded in a map. A write to a single byte of storage  
anywhere within a region is treated in the same way as a write to the entire  
region.  
The layout of a version 20 DCO volume includes an accumulator that stores the  
DRL map and a per-region state map for the volume, plus 32 per-volume maps  
(by default) including a DRL recovery map, and a map for tracking detaches that  
are initiated by the kernel due to I/O error. The remaining 30 per-volume maps  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                 
70 Understanding Veritas Volume Manager  
FastResync  
(by default) are used either for tracking writes to snapshots, or as copymaps.  
The size of the DCO volume is determined by the size of the regions that are  
tracked, and by the number of per-volume maps. Both the region size and the  
number of per-volume maps in a DCO volume may be configured when a volume  
is prepared for use with snapshots. The region size must be a power of 2 and be  
greater than or equal to 16KB.  
As the accumulator is approximately 3 times the size of a per-volume map, the  
size of each plex in the DCO volume can be estimated from this formula:  
DCO_plex_size = ( 3 + number_of_per-volume_maps ) * map_size  
where the size of each map in bytes is:  
map_size = 512 + ( volume_size / ( region_size * 8 ))  
rounded up to the nearest multiple of 8KB. Note that each map includes a 512-  
byte header.  
For the default number of 32 per-volume maps and region size of 64KB, a 10GB  
volume requires a map size of 24KB, and so each plex in the DCO volume  
requires 840KB of storage.  
Note: Full-sized and space-optimized instant snapshots, which are administered  
using the vxsnapcommand, are supported for a version 20 DCO volume layout.  
The use of the vxassistcommand to administer traditional (third-mirror break-  
off) snapshots is not supported for a version 20 DCO volume layout.  
How persistent FastResync works with snapshots  
Persistent FastResync uses a map in a DCO volume on disk to implement change  
tracking. As for non-persistent FastResync, each bit in the map represents a  
contiguous number of blocks in a volume’s address space.  
Figure 1-32 shows an example of a mirrored volume with two plexes on which  
Persistent FastResync is enabled. Associated with the volume are a DCO object  
and a DCO volume with two plexes.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Understanding Veritas Volume Manager 71  
FastResync  
Figure 1-32  
Mirrored volume with persistent FastResync enabled  
Mirrored volume  
Data plex  
Data plex  
Data change object  
DCO  
plex  
DCO  
plex  
DCO volume  
To create a traditional third-mirror snapshot or an instant (copy-on-write)  
snapshot, the vxassistsnapstartor vxsnap makeoperation respectively is  
performed on the volume. This sets up a snapshot plex in the volume and  
associates a disabled DCO plex with it, as shown in Figure 1-33.  
Figure 1-33  
Mirrored volume after completion of a snapstart operation  
Mirrored volume  
Snapshot  
Data plex  
Data plex  
Data change object  
plex  
Disabled  
DCO  
plex  
DCO  
plex  
DCO  
plex  
DCO volume  
Multiple snapshot plexes and associated DCO plexes may be created in the  
volume by re-running the vxassistsnapstartcommand for traditional  
snapshots, or the vxsnapmakecommand for space-optimized snapshots. You  
can create up to a total of 32 plexes (data and log) in a volume.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
72 Understanding Veritas Volume Manager  
FastResync  
Note: Space-optimized instant snapshots do not require additional full-sized  
plexes to be created. Instead, they use a storage cache that typically requires  
only 10% of the storage that is required by full-sized snapshots. There is a trade-  
off in functionality in using space-optimized snapshots as described in  
“Comparison of snapshot features” on page 65. The storage cache is formed  
within a cache volume, and this volume is associated with a cache object. For  
convenience of operation, this cache can be shared by all the instant space-  
optimized snapshots within a disk group.  
A traditional snapshot volume is created from a snapshot plex by running the  
vxassistsnapshotoperation on the volume. For instant snapshots, however,  
the vxsnapmakecommand makes an instant snapshot volume immediately  
available for use. There is no need to run an additional command.  
As illustrated in Figure 1-34, creation of the snapshot volume also sets up a DCO  
object and a DCO volume for the snapshot volume. This DCO volume contains  
the single DCO plex that was associated with the snapshot plex. If two snapshot  
plexes were taken to form the snapshot volume, the DCO volume would contain  
two plexes. For instant space-optimized snapshots, the DCO object and DCO  
volume are associated with a snapshot volume that is created on a cache object  
and not on a VM disk.  
Associated with both the original volume and the snapshot volume are snap  
objects. The snap object for the original volume points to the snapshot volume,  
and the snap object for the snapshot volume points to the original volume. This  
allows VxVM to track the relationship between volumes and their snapshots  
even if they are moved into different disk groups.  
The snap objects in the original volume and snapshot volume are automatically  
deleted in the following circumstances:  
■
For traditional snapshots, the vxassistsnapbackoperation is run to return  
all of the plexes of the snapshot volume to the original volume.  
■
For traditional snapshots, the vxassistsnapclearoperation is run on a  
volume to break the association between the original volume and the  
snapshot volume. If the volumes are in different disk groups, the command  
must be run separately on each volume.  
■
■
For full-sized instant snapshots, the vxsnapreattachoperation is run to  
return all of the plexes of the snapshot volume to the original volume.  
For full-sized instant snapshots, the vxsnapdisor vxsnapsplitoperations  
are run on a volume to break the association between the original volume  
and the snapshot volume. If the volumes are in different disk groups, the  
command must be run separately on each volume.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Understanding Veritas Volume Manager 73  
FastResync  
Note: The vxsnapreattach, disand splitoperations are not supported for  
instant space-optimized snapshots.  
See “Administering volume snapshots” on page 303, and the vxsnap(1M) and  
vxassist(1M) manual pages for more information.  
Figure 1-34  
Mirrored volume and snapshot volume after completion of a  
snapshot operation  
Mirrored volume  
Data plex  
Data plex  
Data change object  
Snap object  
DCO  
log plex  
DCO  
log plex  
DCO volume  
Snapshot volume  
Data plex  
Data change object  
Snap object  
DCO  
log plex  
DCO volume  
Effect of growing a volume on the FastResync map  
It is possible to grow the replica volume, or the original volume, and still use  
FastResync. According to the DCO volume layout, growing the volume has  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
74 Understanding Veritas Volume Manager  
FastResync  
different effects on the map that FastResync uses to track changes to the  
original volume:  
■
For a version 20 DCO volume, the size of the map is increased and the size of  
the region that is tracked by each bit in the map stays the same.  
■
For a version 0 DCO volume, the size of the map remains the same and the  
region size is increased.  
In either case, the part of the map that corresponds to the grown area of the  
volume is marked as “dirty” so that this area is resynchronized. The snapback  
operation fails if it attempts to create an incomplete snapshot plex. In such  
cases, you must grow the replica volume, or the original volume, before invoking  
any of the commands vxsnapreattach, vxsnaprestore, or vxassistsnapback.  
Growing the two volumes separately can lead to a snapshot that shares physical  
disks with another mirror in the volume. To prevent this, grow the volume after  
the snapbackcommand is complete.  
FastResync limitations  
The following limitations apply to FastResync:  
■
Persistent FastResync is supported for RAID-5 volumes, but this prevents  
the use of the relayout or resize operations on the volume while a DCO is  
associated with it.  
■
Neither non-persistent nor persistent FastResync can be used to  
resynchronize mirrors after a system crash. Dirty region logging (DRL),  
which can coexist with FastResync, should be used for this purpose. In  
VxVM 4.0 and later releases, DRL logs may be stored in a version 20 DCO  
volume.  
■
■
When a subdisk is relocated, the entire plex is marked “dirty” and a full  
resynchronization becomes necessary.  
If a snapshot volume is split off into another disk group, non-persistent  
FastResync cannot be used to resynchronize the snapshot plexes with the  
original volume when the disk group is rejoined with the original volume’s  
disk group. Persistent FastResync must be used for this purpose.  
■
If you move or split an original volume (on which persistent FastResync is  
enabled) into another disk group, and then move or join it to a snapshot  
volume’s disk group, you cannot use vxassistsnapbackto resynchronize  
traditional snapshot plexes with the original volume. This restriction arises  
because a snapshot volume references the original volume by its record ID at  
the time that the snapshot volume was created. Moving the original volume  
to a different disk group changes the volume’s record ID, and so breaks the  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Understanding Veritas Volume Manager 75  
Hot-relocation  
association. However, in such a case, you can use the vxplexsnapback  
command with the -f(force) option to perform the snapback.  
Note: This restriction only applies to traditional snapshots. It does not apply to  
instant snapshots.  
■
Any operation that changes the layout of a replica volume can mark the  
FastResync change map for that snapshot “dirty” and require a full  
resynchronization during snapback. Operations that cause this include  
subdisk split, subdisk move, and online relayout of the replica. It is safe to  
perform these operations after the snapshot is completed. For more  
information, see the vxvol(1M), vxassist(1M), and vxplex(1M) manual  
pages.  
Hot-relocation  
Note: You need a full license to use this feature.  
Hot-relocation is a feature that allows a system to react automatically to I/O  
failures on redundant objects (mirrored or RAID-5 volumes) in VxVM and  
restore redundancy and access to those objects. VxVM detects I/O failures on  
objects and relocates the affected subdisks. The subdisks are relocated to disks  
designated as spare disks and/or free space within the disk group. VxVM then  
reconstructs the objects that existed before the failure and makes them  
accessible again.  
When a partial disk failure occurs (that is, a failure affecting only some subdisks  
on a disk), redundant data on the failed portion of the disk is relocated. Existing  
volumes on the unaffected portions of the disk remain accessible. For further  
details, see “Administering hot-relocation” on page 379.  
Volume sets  
Note: You need a full license to use this feature.  
Volume sets are an enhancement to VxVM that allow several volumes to be  
represented by a single logical object. All I/O from and to the underlying  
volumes is directed via the I/O interfaces of the volume set. The volume set  
feature supports the multi-volume enhancement to Veritas File System (VxFS).  
This feature allows file systems to make best use of the different performance  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
76 Understanding Veritas Volume Manager  
Volume sets  
and availability characteristics of the underlying volumes. For example, file  
system metadata could be stored on volumes with higher redundancy, and user  
data on volumes with better performance.  
For more information about creating and administering volume sets, see  
“Creating and administering volume sets” on page 361.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter  
2
Administering disks  
This chapter describes the operations for managing disks used by the Veritas  
Volume Manager (VxVM). This includes placing disks under VxVM control,  
initializing disks, mirroring the root disk, and removing and replacing disks.  
Note: Most VxVM commands require superuser or equivalent privileges.  
Disks that are controlled by the LVM subsystem cannot be used directly as  
VxVM disks, but they can be converted so that their volume groups and logical  
volumes become VxVM disk groups and volumes. For more information on  
conversion, see the Veritas Volume Manager Migration Guide.  
For information about configuring and administering the dynamic multipathing  
(DMP) feature of VxVM that is used with multiported disk arrays, see  
“Administering dynamic multipathing (DMP)” on page 125.  
Disk devices  
When performing disk administration, it is important to understand the  
difference between a disk name and a device name.  
When a disk is placed under VxVM control, a VM disk is assigned to it. You can  
define a symbolic disk name (also known as a disk media name) to refer to a VM  
disk for the purposes of administration. A disk name can be up to 31 characters  
long. If you do not assign a disk name, it defaults to diskgroup## where  
diskgroup is the name of the disk group to which the disk is being added, and ##  
is a sequence number. Your system may use device names that differ from those  
given in the examples.  
The device name (sometimes referred to as devname or disk access name) defines  
the name of a disk device as it is known to the operating system. In HP-UX 11i  
v3, the persistent (agile) forms of such devices are located in the /dev/disk  
Download from Www.Somanuals.com. All Manuals Search And Download.  
               
78 Administering disks  
Disk devices  
and /dev/rdiskdirectories. To maintain backward compatibility, HP-UX also  
creates legacy devices in the /dev/dskand /dev/rdskdirectories.  
VxVM recreates disk devices for all paths in the operating system’s hardware  
device tree as metadevices (DMP nodes) in the /dev/vx/dmpand  
/dev/vx/rdmpdirectories. The dynamic multipathing (DMP) feature of VxVM  
uses a DMP node to represent a disk that can be accessed by one or more  
physical paths, perhaps via different controllers. The number of access paths  
that are available depends on whether the disk is a single disk, or is part of a  
multiported disk array that is connected to a system. DMP nodes are not used by  
the native multipathing feature of HP-UX.  
If a legacy device special file does not exist for the path to a LUN, DMP generates  
the DMP subpath name using the c#t#d#format, where the controller number  
in c#is set to 512 plus the instance number of the target path to which the LUN  
path belongs, the target is set to t0, and the device number in d#is set to the  
instance number of the LUN path. As the controller number is greater than 512,  
DMP subpath names that are generated in this way do not conflict with any  
legacy device names provided by the operating system. If a DMP subpath name  
has a controller number that is greater than 512, this implies that the operating  
system does not provide a legacy device special file for the device.  
You can use the vxdiskutility to display the paths that are subsumed by a DMP  
metadevice, and to display the status of each path (for example, whether it is  
enabled or disabled).  
For more information, see “Administering dynamic multipathing (DMP)” on  
page 125.  
Device names may also be remapped as enclosure-based names as described in  
the following section.  
Disk device naming in VxVM  
There are two different methods of naming disk devices:  
■
■
Operating system-based naming  
Enclosure-based naming  
Operating system-based naming  
Under operating system-based naming, all disk devices except fabric mode disks  
are displayed either using the legacy c#t#d#format or the persistent disk##  
format. By default, VxVM commands display the names of these devices in the  
legacy format as these correspond to the names of the metanodes that are  
created by DMP.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                 
Administering disks 79  
Disk devices  
The syntax of a legacy device name is c#t#d#, where c#represents a controller  
on a host bus adapter, t#is the target controller ID, and d#identifies a disk on  
the target controller.  
Fabric mode disk devices are named as follows:  
■
Disks in supported disk arrays are named using the enclosure name_#  
format. For example, disks in the supported disk array name FirstFloor  
are named FirstFloor_0, FirstFloor_1, FirstFloor_2and so on.  
(You can use the vxdmpadmcommand to administer enclosure names.)  
■
■
Disks in the DISKScategory (JBOD disks) are named using the Disk_#  
format.  
Disks in the OTHER_DISKScategory (disks that are not multipathed by  
DMP) are named using the fabric_# format.  
Enclosure-based naming  
Enclosure-based naming operates as follows:  
■
All fabric or non-fabric disks in supported disk arrays are named using the  
enclosure_name_#format. For example, disks in the supported disk array,  
enggdeptare named enggdept_0, enggdept_1, enggdept_2and so on.  
(You can use the vxdmpadmcommand to administer enclosure names. See  
“Administering DMP using vxdmpadm” on page 139 and the vxdmpadm(1M)  
manual page for more information.)  
■
■
Disks in the DISKScategory (JBOD disks) are named using the Disk_#  
format.  
Disks in the OTHER_DISKScategory (disks that are not multipathed by  
DMP) are named as follows:  
■
Non-fabric disks are named using the c#t#d#or disk##format.  
Fabric disks are named using the fabric_#format.  
■
See “Changing the disk-naming scheme” on page 91 for details of how to switch  
between the operating system and enclosure based naming schemes.  
To display the native OS device names of a VM disk (such as mydg01), use the  
following command:  
# vxdisk path | egrep diskname  
For information on how to rename an enclosure, see “Renaming an enclosure”  
on page 155.  
For a description of disk categories, see “Disk categories” on page 83.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
80 Administering disks  
Disk devices  
Private and public disk regions  
Most VM disks have two regions:  
private region  
A small area where configuration information is stored. A disk  
header label, configuration records for VxVM objects (such as  
volumes, plexes and subdisks), and an intent log for the  
configuration database are stored here. The default private  
region size is 32 megabytes (except for VxVM boot disk groups  
where the private region size must be 1 megabyte), which is  
large enough to record the details of several thousand VxVM  
objects in a disk group.  
Under most circumstances, the default private region size  
should be sufficient. For administrative purposes, it is usually  
much simpler to create more disk groups that contain fewer  
volumes, or to split large disk groups into several smaller ones  
(as described in “Splitting disk groups” on page 205). If  
required, the value for the private region size may be  
overridden when you add or replace a disk using the  
vxdiskadmcommand.  
Each disk that has a private region holds an entire copy of the  
configuration database for the disk group. The size of the  
configuration database for a disk group is limited by the size of  
the smallest copy of the configuration database on any of its  
member disks.  
public region  
An area that covers the remainder of the disk, and which is  
used for the allocation of storage space to subdisks.  
A disk’s type identifies how VxVM accesses a disk, and how it manages the disk’s  
private and public regions. The following disk access types are used by VxVM:  
simple  
The public and private regions are on the same disk area (with  
the public area following the private area).  
nopriv  
There is no private region (only a public region for allocating  
subdisks). This is the simplest disk type consisting only of  
space for allocating subdisks. Such disks are most useful for  
defining special devices (such as RAM disks, if supported) on  
which private region data would not persist between reboots.  
They can also be used to encapsulate disks where there is  
insufficient room for a private region. The disks cannot store  
configuration and log copies, and they do not support the use  
of the vxdiskaddregioncommand to define reserved regions.  
VxVM cannot track the movement of noprivdisks on a SCSI  
chain or between controllers.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Administering disks 81  
Disk devices  
auto  
When the vxconfigddaemon is started, VxVM obtains a list of  
known disk device addresses from the operating system and  
configures disk access records for them automatically.  
Auto-configured disks (with disk access type auto) support the following disk  
formats:  
cdsdisk  
The disk is formatted as a Cross-platform Data Sharing (CDS)  
disk that is suitable for moving between different operating  
systems. This is the default format for disks that are not used  
to boot the system.Typically, most disks on a system are  
configured as this disk type. However, it is not a suitable  
format for boot, root or swap disks, for mirrors or  
hot-relocation spares of such disks, or for Extensible Firmware  
Interface (EFI) disks.  
hpdisk  
The disk is formatted as a simple disk. This format can be  
applied to disks that are used to boot the system. The disk can  
be converted to a CDS disk if it was not initialized for use as a  
boot disk.  
See the vxcdsconvert(1M) manual page for information about the utility that  
you can use to convert disks to the cdsdiskformat.  
Caution: The CDS disk format is incompatible with EFI disks. If a disk is  
initialized by VxVM as a CDS disk, the CDS header occupies the portion of the  
disk where the partition table would usually be located. If you subsequently use  
a command such as fdiskto create a partition table on a CDS disk, this erases  
the CDS information and could cause data corruption.  
By default, auto-configured non-EFI disks are formatted as cdsdiskdisks when  
they are initialized for use with VxVM. You can change the default format by  
using the vxdiskadm(1M) command to update the /etc/default/vxdisk  
defaults file as described in “Displaying and changing default disk layout  
attributes” on page 97. See the vxdisk(1M) manual page for details of the usage  
of this file, and for more information about disk types and their configuration.  
Auto-configured EFI disks are formatted as hpdiskdisks by default.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
82 Administering disks  
Discovering and configuring newly added disk devices  
Discovering and configuring newly added disk  
devices  
When you physically connect new disks to a host or when you zone new fibre  
channel devices to a host, you can use the vxdctl enablecommand to rebuild  
the volume device node directories and to update the DMP internal database to  
reflect the new state of the system.  
To reconfigure the DMP database, first run ioscanfollowed by insfto make the  
operating system recognize the new disks, and then invoke the vxdctlenable  
command. See the vxdctl(1M) manual page for more information.  
You can also use the vxdiskscandiskscommand to scan devices in the  
operating system device tree, and to initiate dynamic reconfiguration of  
multipathed disks.  
If you want VxVM to scan only for new devices that have been added to the  
system, and not for devices that have been enabled or disabled, specify the -f  
option to either of the commands, as shown here:  
# vxdctl -f enable  
# vxdisk -f scandisks  
However, a complete scan is initiated if the system configuration has been  
modified by changes to:  
■
■
■
Installed array support libraries.  
The devices that are listed as being excluded from use by VxVM.  
DISKS (JBOD), SCSI3, or foreign device definitions.  
See the vxdctl(1M) and vxdisk(1M) manual pages for more information.  
Partial device discovery  
The Dynamic Multipathing (DMP) feature of VxVM supports partial device  
discovery where you can include or exclude sets of disks or disks attached to  
controllers from the discovery process.  
The vxdiskscandiskscommand rescans the devices in the OS device tree and  
triggers a DMP reconfiguration. You can specify parameters to vxdisk  
scandisksto implement partial device discovery. For example, this command  
makes VxVM discover newly added devices that were unknown to it earlier:  
# vxdisk scandisks new  
The next example discovers fabric devices:  
# vxdisk scandisks fabric  
The following command scans for the devices c1t1d0and c2t2d0:  
# vxdisk scandisks device=c1t1d0,c2t2d0  
Download from Www.Somanuals.com. All Manuals Search And Download.  
               
Administering disks 83  
Discovering and configuring newly added disk devices  
Alternatively, you can specify a ! prefix character to indicate that you want to  
scan for all devices except those that are listed:  
# vxdisk scandisks !device=c1t1d0,c2t2d0  
You can also scan for devices that are connected (or not connected) to a list of  
logical or physical controllers. For example, this command discovers and  
configures all devices except those that are connected to the specified logical  
controllers:  
# vxdisk scandisks !ctlr=c1,c2  
The next command discovers devices that are connected to the specified  
physical controller:  
# vxdisk scandisks pctlr=8/12.8.0.255.0  
Note: The items in a list of physical controllers are separated by +characters.  
You can use the command vxdmpadmgetctlrallto obtain a list of physical  
controllers.  
You can specify only one selection argument to the vxdiskscandisks  
command. Specifying multiple options results in an error.  
For more information, see the vxdisk(1M) manual page.  
Discovering disks and dynamically adding disk arrays  
You can dynamically add support for a new type of disk array which has been  
developed by a third-party vendor. The support comes in the form of  
vendor-supplied libraries, and is added to an HP-UX system by using the  
swinstallcommand.  
Disk categories  
Disk arrays that have been certified for use with Veritas Volume Manager are  
supported by an array support library (ASL), and are categorized by the vendor  
ID string that is returned by the disks (for example, “HITACHI”).  
Disks in JBODs for which DMP (see “Administering dynamic multipathing  
(DMP)” on page 125) can be supported in Active/Active mode, and which are  
capable of being multipathed, are placed in the DISKScategory. Disks in  
unsupported arrays can be placed in this category by following the steps given  
in “Adding unsupported disk arrays to the DISKS category” on page 87.  
Disks in JBODs that do not fall into any supported category, and which are not  
capable of being multipathed by DMP are placed in the OTHER_DISKScategory.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
84 Administering disks  
Discovering and configuring newly added disk devices  
Adding support for a new disk array  
The following example illustrates how to add support for a new disk array  
named vrtsdato an HP-UX system using a vendor-supplied package on a  
mounted CD-ROM:  
# swinstall -s /cdrom vrtsda  
The new disk array does not need to be already connected to the system when  
the package is installed. If any of the disks in the new disk array are  
subsequently connected, first run the ioscancommand, and then run either  
the vxdiskscandisksor the vxdctlenablecommand to include the devices  
in the VxVM device list.  
Enabling discovery of new devices  
To have VxVM discover a new disk array, use the following command:  
# vxdctl enable  
This command scans all of the disk devices and their attributes, updates the  
VxVM device list, and reconfigures DMP with the new device database. There is  
no need to reboot the host.  
Note: This command ensures that dynamic multipathing is set up correctly on  
the array. Otherwise, VxVM treats the independent paths to the disks as  
separate devices, which can result in data corruption.  
Removing support for a disk array  
To remove support for the vrtsdadisk array, use the following command:  
# swremove vrtsda  
If the arrays remain physically connected to the host after support has been  
removed, they are listed in the OTHER_DISKScategory, and the volumes remain  
available.  
Third-party driver coexistence  
The third-party driver (TPD) coexistence feature of VxVM allows I/O that is  
controlled by third-party multipathing drivers to bypass DMP while retaining  
the monitoring capabilities of DMP. Provided that a suitable ASL is available,  
devices that use TPDs can be discovered without requiring you to set up a  
specification file, or to run a special command. In previous releases, VxVM only  
supported TPD coexistence if the code of the third-party driver was intrusively  
modified. The new TPD coexistence feature maintains backward compatibility  
with such methods, but it also permits coexistence without require any change  
in a third-party multipathing driver.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
Administering disks 85  
Discovering and configuring newly added disk devices  
See “Changing device naming for TPD-controlled enclosures” on page 94 for  
information on how to change the form of TPD device names that are displayed  
by VxVM.  
See “Displaying information about TPD-controlled devices” on page 143 for  
details of how to find out the TPD configuration information that is known to  
DMP.  
Administering the Device Discovery Layer  
Dynamic addition of disk arrays is possible because of the existence of the  
Device Discovery Layer (DDL) which is a facility for discovering disks and their  
attributes that are required for VxVM and DMP operations. The DDL is  
administered using the vxddladmutility, which can be used to perform the  
following tasks:  
■
■
■
■
■
■
■
■
List the types of arrays that are supported.  
Add support for an array to DDL.  
Remove support for an array from DDL.  
List information about excluded disk arrays.  
List disks that are supported in the DISKS(JBOD) category.  
Add disks from different vendors to the DISKScategory.  
Remove disks from the DISKScategory.  
Add disks as foreign devices.  
The following sections explain these tasks in more detail. For further  
information, see the vxddladm(1M) manual page.  
Listing details of supported disk arrays  
To list all currently supported disk arrays, use the following command:  
# vxddladm listsupport all  
Note: Use this command to obtain values for the vidand pidattributes that are  
used with other forms of the vxddladmcommand.  
To display more detailed information about a particular array library, use this  
form of the command:  
# vxddladm listsupport libname=library_name.sl  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
86 Administering disks  
Discovering and configuring newly added disk devices  
This command displays the vendor ID (VID), product IDs (PIDs) for the arrays,  
array types (for example, A/A or A/P), and array names. The following is sample  
output.  
# vxddladm listsupport libname=libvxfujitsu.so  
ATTR_NAME  
ATTR_VALUE  
=================================================  
LIBNAME  
VID  
libvxfujitsu.so  
vendor  
PID  
GR710, GR720, GR730  
GR740, GR820, GR840  
A/A, A/P  
FJ_GR710, FJ_GR720, FJ_GR730  
FJ_GR740, FJ_GR820, FJ_GR840  
ARRAY_TYPE  
ARRAY_NAME  
Excluding support for a disk array library  
To exclude all arrays that depend on a particular array library from  
participating in device discovery, use the following command:  
# vxddladm excludearray libname=libvxenc.sl  
This example excludes support for disk arrays that depends on the library  
libvxenc.sl. You can also exclude support for disk arrays from a particular  
vendor, as shown in this example:  
# vxddladm excludearray vid=ACME pid=X1  
For more information about excluding disk array support, see the vxddladm(1M)  
manual page.  
Re-including support for an excluded disk array library  
If you have excluded support for all arrays that depend on a particular disk  
array library, you can use the includearraykeyword to remove the entry  
from the exclude list, as shown in the following example:  
# vxddladm includearray libname=libvxenc.sl  
This command adds the array library to the database so that the library can once  
again be used in device discovery. If vxconfigdis running, you can use the  
vxdiskscandiskscommand to discover the arrays and add their details to the  
database.  
Listing excluded disk arrays  
To list all disk arrays that are currently excluded from use by VxVM, use the  
following command:  
# vxddladm listexclude  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Administering disks 87  
Discovering and configuring newly added disk devices  
Listing supported disks in the DISKS category  
To list disks that are supported in the DISKS(JBOD) category, use the following  
command:  
# vxddladm listjbod  
Adding unsupported disk arrays to the DISKS category  
Disk arrays should be added as JBOD devices if no ASL is available for the array.  
JBODs are assumed to be Active/Active (A/A) unless otherwise specified. If a  
suitable ASL is not available, an A/A-A, A/P or A/PF array must be claimed as an  
Active/Passive (A/P) JBOD to prevent path delays and I/O failures.  
Caution: The procedure in this section ensures that Dynamic Multipathing  
(DMP) is set up correctly on an array that is not supported by Veritas Volume  
Manager. If the native multipathing in HP-UX 11i v3 does not recognize such an  
array as being multipathed, Veritas Volume Manager treats the independent  
legacy paths to the disks as separate devices, which can result in data  
corruption.  
To add an unsupported disk array  
1
Use the following command to identify the vendor ID and product ID of the  
disks in the array:  
# /etc/vx/diag.d/vxdmpinq device_name  
where device_name is the device name of one of the disks in the array (for  
example, /dev/rdsk/c1t20d0). Note the values of the vendor ID (VID)  
and product ID (PID) in the output from this command. For Fujitsu disks,  
also note the number of characters in the serial number that is displayed.  
The following is sample output:  
# /etc/vx/diag.d/vxdmpinq /dev/rdsk/c1t20d0  
Vendor id (VID)  
Product id (PID):  
Revision  
: SEAGATE  
ST318404LSUN18G  
: 8507  
Serial Number  
: 0025T0LA3H  
In this example, the vendor ID is SEAGATEand the product ID is  
ST318404LSUN18G.  
2
Stop all applications, such as databases, from accessing VxVM volumes that  
are configured on the array, and unmount all VxFS file systems and  
checkpoints that are configured on the array.  
3
4
If the array is of type A/A-A, A/P or A/PF, configure it in autotrespass mode.  
Enter the following command to add a new JBOD category:  
# vxddladm addjbod vid=vendorid pid=productid \  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
88 Administering disks  
Discovering and configuring newly added disk devices  
[length=serialno_length] [policy=ap]  
where vendorid and productid are the VID and PID values that you found  
from the previous step. For example, vendorid might be FUJITSU, IBM, or  
SEAGATE. For Fujitsu devices, you must also specify the number of  
characters in the serial number as the argument to the lengthargument  
(for example, 10). If the array is of type A/A-A, A/P or A/PF, you must also  
specify the policy=apattribute.  
Continuing the previous example, the command to define an array of disks  
of this type as a JBOD would be:  
# vxddladm addjbod vid=SEAGATE pid=ST318404LSUN18G  
5
6
Use the vxdctlenablecommand to bring the array under VxVM control.  
# vxdctl enable  
See “Enabling discovery of new devices” on page 84.  
To verify that the array is now supported, enter the following command:  
# vxddladm listjbod  
The following is sample output from this command for the example array:  
VID  
=============================================================  
SEAGATE ALL PIDs 18 -1 36 12  
PID  
Opcode Page Code Page Offset SNO length  
7
To verify that the array is recognized, use the vxdmpadmlistenclosure  
command as shown in the following sample output for the example array:  
# vxdmpadm listenclosure all  
ENCLR_NAME  
ENCLR_TYPE  
ENCLR_SNO  
STATUS  
=============================================================  
OTHER_DISKS  
Disk  
OTHER_DISKS  
Disk  
OTHER_DISKS  
DISKS  
CONNECTED  
CONNECTED  
The enclosure name and type for the array are both shown as being set to  
Disk. You can use the vxdisklistcommand to display the disks in the  
array:  
# vxdisk list  
DEVICE  
Disk_0  
Disk_1  
...  
TYPE  
auto:none  
auto:none  
DISK  
-
-
GROUP  
-
-
STATUS  
online invalid  
online invalid  
8
To verify that the DMP paths are recognized, use the vxdmpadmgetdmpnode  
command as shown in the following sample output for the example array:  
# vxdmpadm getdmpnode enclosure=Disk  
NAME  
STATE  
ENCLR-TYPE PATHS ENBL  
DSBL ENCLR-NAME  
=============================================================  
Disk_0  
Disk_1  
...  
ENABLED Disk  
ENABLED Disk  
2
2
2
2
0 Disk  
0 Disk  
This shows that there are two paths to the disks in the array.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Administering disks 89  
Discovering and configuring newly added disk devices  
For more information, enter the command vxddladmhelpaddjbod, or see  
the vxddladm(1M) and vxdmpadm(1M) manual pages.  
Removing disks from the DISKS category  
To remove disks from the DISKS(JBOD) category, use the vxddladmcommand  
with the rmjbodkeyword. The following example illustrates the command for  
removing disks supplied by the vendor, Seagate:  
# vxddladm rmjbod vid=SEAGATE  
Adding foreign devices  
DDL may not be able to discover some devices that are controlled by third-party  
drivers, such as those that provide multipathing or RAM disk capabilities. For  
these devices it may be preferable to use the multipathing capability that is  
provided by the third-party drivers for some arrays rather than using the  
Dynamic Multipathing (DMP) feature. Such foreign devices can be made  
available as simple disks to VxVM by using the vxddladmaddforeigncommand.  
This also has the effect of bypassing DMP for handling I/O. The following  
example shows how to add entries for block and character devices in the  
specified directories:  
# vxddladm addforeign blockdir=/dev/foo/dsk \  
chardir=/dev/foo/rdsk  
By default, this command suppresses any entries for matching devices in the  
OS-maintained device tree that are found by the autodiscovery mechanism. You  
can override this behavior by using the -fand -noptions as described on the  
vxddladm(1M) manual page.  
After adding entries for the foreign devices, use either the vxdiskscandisksor  
the vxdctlenablecommand to discover the devices as simple disks. These disks  
then behave in the same way as autoconfigured disks.  
The foreign device mechanism was introduced in VxVM 4.0 to support  
non-standard devices such as RAM disks, some solid state disks, and  
pseudo-devices such as EMC PowerPath. This mechanism has a number of  
limitations:  
■
A foreign device is always considered as a disk with a single path. Unlike an  
autodiscovered disk, it does not have a DMP node.  
■
It is not supported for shared disk groups in a clustered environment. Only  
standalone host systems are supported.  
■
■
It is not supported for Persistent Group Reservation (PGR) operations.  
It is not under the control of DMP, so enabling of a failed disk cannot be  
automatic, and DMP administrative commands are not applicable.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
90 Administering disks  
Placing disks under VxVM control  
■
■
Enclosure information is not available to VxVM. This can reduce the  
availability of any disk groups that are created using such devices.  
EFI disks that are under the control of HP-UX native multipathing cannot be  
initialized as foreign disks. You must migrate the system to DMP, initialize  
the disk as an EFI disk, and then migrate the system back to HP-UX native  
multipathing.  
See “Migrating between DMP and HP-UX native multipathing” on page 130.  
■
Foreign devices, such as HP-UX native multipathing metanodes, do not have  
enclosures, controllers or DMP nodes that can be administered using VxVM  
commands. An error message is displayed if you attempt to use the  
vxddladmor vxdmpadmcommands to administer such devices while HP-UX  
native multipathing is configured.  
■
The I/O Fencing and Cluster File System features are not supported for  
foreign devices.  
If a suitable ASL is available for an array, these limitations are removed.  
See “Third-party driver coexistence” on page 84.  
Placing disks under VxVM control  
When you add a disk to a system that is running VxVM, you need to put the disk  
under VxVM control so that VxVM can control the space allocation on the disk.  
Unless you specify a disk group, VxVM places new disks in a default disk group  
according to the rules given in “Rules for determining the default disk group” on  
page 168.  
The method by which you place a disk under VxVM control depends on the  
circumstances:  
■
■
If the disk is new, it must be initialized and placed under VxVM control. You  
can use the menu-based vxdiskadmutility to do this.  
Caution: Initialization destroys existing data on disks.  
If the disk is not needed immediately, it can be initialized (but not added to a  
disk group) and reserved for future use. To do this, enter nonewhen asked  
to name a disk group. Do not confuse this type of “spare disk” with a  
hot-relocation spare disk.  
■
■
If the disk was previously initialized for future use by VxVM, it can be  
reinitialized and placed under VxVM control.  
If the disk was previously used for a file system, VxVM prompts you to  
confirm that you really want to destroy the file system.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Administering disks 91  
Changing the disk-naming scheme  
■
If the disk was previously in use by the LVM subsystem, you can preserve  
existing data while still letting VxVM take control of the disk. This is  
accomplished using conversion. With conversion, the virtual layout of the  
data is fully converted to VxVM control (see the Veritas Volume Manager  
Migration Guide).  
■
■
If the disk was previously in use by the LVM subsystem, but you do not want  
to preserve the data on it, use the LVM command, pvremove, before  
attempting to initialize the disk for VxVM.  
Multiple disks on one or more controllers can be placed under VxVM control  
simultaneously. Depending on the circumstances, all of the disks may not be  
processed the same way.  
It is possible to configure the vxdiskadmutility not to list certain disks or  
controllers as being available. For example, this may be useful in a SAN  
environment where disk enclosures are visible to a number of separate systems.  
To exclude a device from the view of VxVM, select item 17 (Prevent  
multipathing/Suppress devices from VxVM’s view) from the vxdiskadm  
main menu. See “Disabling and enabling multipathing for specific devices” on  
page 133 for details.  
Changing the disk-naming scheme  
You can either use enclosure-based naming for disks or the operating system’s  
naming convention. Select menu item 20 from the vxdiskadmmain menu to  
change the disk-naming scheme that you want VxVM to use. When prompted,  
enter yto change the naming scheme. For operating system based naming, you  
are asked to select between default, legacy or new device names. The vxconfigd  
daemon is then restarted to bring the new disk naming scheme into effect.  
VxVM commands display device names according the current naming scheme. If  
operating system-based naming is selected, all VxVM commands that list DMP  
node devices will display device names according to the mode that is specified.  
Mode  
Format of output from VxVM command  
default  
The same format is used as in the input to the command (if this can  
be determined). Otherwise, legacy names are used. This is the default  
mode.  
legacy  
new  
Only legacy names are displayed.  
Only new (agile) names are displayed.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
92 Administering disks  
Changing the disk-naming scheme  
Alternatively, you can change the naming scheme from the command line. The  
following commands select enclosure-based and operating system-based  
naming respectively:  
# vxddladm set namingscheme=ebn [persistence={yes|no}]  
# vxddladm set namingscheme=osn [mode={default|legacy|new}] \  
[persistence={yes|no}]  
The optional persistenceargument allows you to select whether the  
DDL-assigned names of disk devices that are displayed by VxVM remain  
unchanged after disk hardware has been reconfigured and the system rebooted.  
By default, both enclosure-based naming and operating system-based naming  
are persistent.  
The following examples illustrate the use of vxddladm set namingscheme  
command:  
# vxddladm set namingscheme=osn mode=default  
# vxdmpadm getlungroup dmpnodename=c1t65d0  
NAME  
===============================================================  
c1t65d0 ENABLED Disk Disk  
STATE  
ENCLR-TYPE PATHS  
ENBL  
DSBL ENCLR-NAME  
2
2
0
# vxdmpadm getlungroup dmpnodename=disk25  
NAME STATE ENCLR-TYPE PATHS ENBL  
===============================================================  
disk25 ENABLED Disk Disk  
DSBL ENCLR-NAME  
2
2
0
# vxddladm set namingscheme=osn mode=legacy  
# vxdmpadm getlungroup dmpnodename=c1t65d0  
NAME  
===============================================================  
c1t65d0 ENABLED Disk Disk  
STATE  
ENCLR-TYPE PATHS  
ENBL  
DSBL ENCLR-NAME  
2
2
0
# vxdmpadm getlungroup dmpnodename=disk25  
NAME STATE ENCLR-TYPE PATHS ENBL  
===============================================================  
c1t65d0 ENABLED Disk Disk  
DSBL ENCLR-NAME  
2
2
0
# vxddladm set namingscheme=osn mode=new  
# vxdmpadm getlungroup dmpnodename=c1t65d0  
NAME  
===============================================================  
disk25 ENABLED Disk Disk  
STATE  
ENCLR-TYPE PATHS  
ENBL  
DSBL ENCLR-NAME  
2
2
0
# vxdmpadm getlungroup dmpnodename=disk25  
NAME STATE ENCLR-TYPE PATHS ENBL  
===============================================================  
disk25 ENABLED Disk Disk  
DSBL ENCLR-NAME  
2
2
0
# vxddladm set namingscheme=ebn  
# vxdmpadm getlungroup dmpnodename=c1t65d0  
VxVM vxdmpadm ERROR V-5-1-10910 Invalid da-name  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Administering disks 93  
Changing the disk-naming scheme  
# vxdmpadm getlungroup dmpnodename=disk25  
VxVM vxdmpadm ERROR V-5-1-10910 Invalid da-name  
# vxdmpadm getlungroup dmpnodename=Disk_11  
NAME  
===============================================================  
Disk_11 ENABLED Disk Disk  
STATE  
ENCLR-TYPE PATHS  
ENBL  
DSBL ENCLR-NAME  
2
2
0
To find out which sort of naming is currently enabled, use the vxddladm get  
namingschemecommand, as shown in the following example:  
# vxddladm get namingscheme  
NAMING_SCHEME  
===============================================  
OS Native Yes New  
PERSISTENCE  
MODE  
The effect of enabling persistent device names in conjunction with operating  
system-based naming is discussed in “Regenerating persistent device names” on  
page 93.  
Regenerating persistent device names  
The persistent device naming feature makes the DDL-assigned names of disk  
devices persistent across system reboots. If operating system-based naming is  
selected, each disk name is usually set to the name of one of the paths to the  
disk. After hardware reconfiguration and a subsequent reboot, the operating  
system may generate different names for the paths to the disks. As DDL assigns  
persistent disk names using the persistent device name database that was  
generated during a previous boot session, the disk names may no longer  
correspond to the actual paths. This does not prevent the disks from being used,  
but the association between the disk name and one of its paths is lost.  
To find the relationship between a disk and its paths, run one of the following  
commands:  
# vxdmpadm getsubpaths dmpnodename=disk_access_name  
# vxdisk list disk_access_name  
To update the disk names so that they correspond to the new path names  
1
Remove the file that contains the existing persistent device name database:  
# rm /etc/vx/disk.info  
2
Restart the VxVM configuration demon:  
# vxconfigd -k  
This regenerates the persistent name database.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
94 Administering disks  
Changing the disk-naming scheme  
Changing device naming for TPD-controlled enclosures  
Note: This feature is available only if the default disk-naming scheme is set to  
use operating system-based naming, and the TPD-controlled enclosure does not  
contain fabric disks.  
For disk enclosures that are controlled by third-party drivers (TPD) whose  
coexistence is supported by an appropriate ASL, the default behavior is to assign  
device names that are based on the TPD-assigned node names. You can use the  
vxdmpadmcommand to switch between these names and the device names that  
are known to the operating system:  
# vxdmpadm setattr enclosure enclosure tpdmode=native|pseudo  
The argument to the tpdmodeattribute selects names that are based on those  
used by the operating system (native), or TPD-assigned node names (pseudo).  
If tpdmodeis set to native, the path with the smallest device number is  
displayed.  
Discovering the association between enclosure and OS based disk  
names  
If you enable enclosure-based naming, and use the vxprintcommand to display  
the structure of a volume, it shows enclosure-based disk device names (disk  
access names) rather than operating system-based names. To discover the  
operating system-based names that are associated with a given enclosure-based  
disk name, use either of the following commands:  
# vxdisk list enclosure-based_name  
# vxdmpadm getsubpaths dmpnodename=enclosure-based_name  
For example, to find the physical device that is associated with disk ENC0_21,  
the appropriate commands would be:  
# vxdisk list ENC0_21  
# vxdmpadm getsubpaths dmpnodename=ENC0_21  
To obtain the full pathname for the block and character disk device from these  
commands, append the displayed device name to /dev/vx/dmpor  
/dev/vx/rdmp.  
Issues regarding persistent simple or nopriv disks with  
enclosure-based naming  
If you change from c#t#d# based naming to enclosure-based naming, persistent  
simple or nopriv disks may be put in the “error” state and cause VxVM objects  
on those disks to fail. If this happens, use the following procedures to correct the  
problem:  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                       
Administering disks 95  
Changing the disk-naming scheme  
■
■
Persistent simple or nopriv disks in the boot disk group  
Persistent simple or nopriv disks in non-boot disk groups  
These procedures use the vxdarestoreutility to handle errors in persistent  
simple and nopriv disks that arise from changing to the enclosure-based naming  
scheme. You do not need to perform either procedure if the devices on which  
any simple or nopriv disks are present are not automatically configured by  
VxVM (for example, non-standard disk devices such as ramdisks).  
Note: The disk access records for simple disks are either persistent or  
non-persistent. The disk access record for a persistent simple disk is stored in  
the disk’s private region. The disk access record for a non-persistent simple disk  
is automatically configured in memory at VxVM startup. A simple disk has a  
non-persistent disk access record if autoconfigis included in the flagsfield  
that is displayed by the vxdisklistdisk_access_name command. If the  
autoconfigflag is not present, the disk access record is persistent. Nopriv  
disks are always persistent.  
Note: You cannot run vxdarestoreif c#t#d# naming is in use. Additionally,  
vxdarestoredoes not handle failures on persistent simple/nopriv disks that are  
caused by renaming enclosures, by hardware reconfiguration that changes  
device names. or by removing support from the JBOD category for disks that  
belong to a particular vendor when enclosure-based naming is in use.  
For more information about the vxdarestorecommand, see the  
vxdarestore(1M) manual page.  
Persistent simple or nopriv disks in the boot disk group  
If all persistent simple and nopriv disks in the boot disk group (usually aliased as  
bootdg) go into the error state, the vxconfigddaemon is disabled after the  
naming scheme change.  
To remove the error state for persistent simple or nopriv disks in the boot  
disk group  
1
2
Use vxdiskadm to change back to c#t#d# naming.  
Enter the following command to restart the VxVM configuration daemon:  
# vxconfigd -kr reset  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
96 Administering disks  
Installing and formatting disks  
3
If you want to use enclosure-based naming, use vxdiskadmto add a  
non-persistent simple disk to the bootdgdisk group, change back to the  
enclosure-based naming scheme, and then run the following command:  
# /etc/vx/bin/vxdarestore  
Note: If not all the disks in bootdggo into the error state, you need only run  
vxdarestoreto restore the disks that are in the error state and the objects that  
they contain.  
Persistent simple or nopriv disks in non-boot disk groups  
If an imported disk group, other than bootdg, consists only of persistent simple  
and/or nopriv disks, it is put in the “onlinedgdisabled” state after the  
change to the enclosure-based naming scheme.  
To remove the error state for persistent simple or nopriv disks in non-boot  
disk groups  
1
Deport the disk group using the following command:  
# vxdg deport diskgroup  
2
Use the vxdarestorecommand to restore the failed disks, and to recover the  
objects on those disks:  
# /etc/vx/bin/vxdarestore  
3
Re-import the disk group using the following command:  
# vxdg import diskgroup  
Installing and formatting disks  
Depending on the hardware capabilities of your disks and of your system, you  
may either need to shut down and power off your system before installing the  
disks, or you may be able to hot-insert the disks into the live system. Many  
operating systems can detect the presence of the new disks on being rebooted. If  
the disks are inserted while the system is live, you may need to enter an  
operating system-specific command to notify the system.  
If the disks require low or intermediate-level formatting before use, use the  
operating system-specific formatting command to do this.  
Note: SCSI disks are usually preformatted. Reformatting is needed only if the  
existing formatting has become damaged.  
The following sections provide detailed examples of how to use the vxdiskadm  
utility to place disks under VxVM control in various ways and circumstances.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Administering disks 97  
Displaying and changing default disk layout attributes  
Displaying and changing default disk layout  
attributes  
To display or change the default values for initializing disks, select menu item  
21 (Change/display the default disk layout)from the vxdiskadm  
main menu. For disk initialization, you can change the default format and the  
default length of the private region.  
The attribute settings for initializing disks are stored in the file,  
/etc/default/vxdisk.  
See the vxdisk(1M) manual page for more information.  
Adding a disk to VxVM  
Formatted disks being placed under VxVM control may be new or previously  
used outside VxVM. The set of disks can consist of all disks on the system, all  
disks on a controller, selected disks, or a combination of these.  
Depending on the circumstances, all of the disks may not be processed in the  
same way.  
Caution: Initialization does not preserve data on disks.  
When initializing multiple disks at one time, it is possible to exclude certain  
disks or certain controllers. To exclude disks, list the names of the disks to be  
excluded in the file /etc/vx/disks.excludebefore the initialization. You  
can exclude all disks on specific controllers from initialization by listing those  
controllers in the file /etc/vx/cntrls.exclude.  
To initialize disks for VxVM use  
1
Select menu item 1 (Add or initialize one or more disks)from  
the vxdiskadmmain menu.  
2
At the following prompt, enter the disk device name of the disk to be added  
to VxVM control (or enter listfor a list of disks):  
Add or initialize disks  
Menu: VolumeManager/Disk/AddDisks  
Use this operation to add one or more disks to a disk group.  
You can add the selected disks to an existing disk group or to  
a new disk group that will be created as a part of the  
operation. The selected disks may also be added to a disk group  
as spares. Or they may be added as nohotuses to be excluded  
from hot-relocation use. The selected disks may also be  
initialized without adding them to a disk group leaving the  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
98 Administering disks  
Adding a disk to VxVM  
disks available for use as replacement disks.  
More than one disk or pattern may be entered at the prompt.  
Here are some disk selection examples:  
all:  
c3 c4t2: all disks on both controller 3 and controller 4,  
target 2  
c3t4d2:  
all disks  
a single disk (in the c#t#d# naming scheme)  
a single disk (in the enclosure based naming scheme)  
all disks on the enclosure whose name is xyz  
xyz_0:  
xyz_:  
disk#:  
a single disk (in the new naming scheme)  
Select disk devices to add:  
[<pattern-list>,all,list,q,?]  
<pattern-list> can be a single disk, or a series of disks and/or controllers  
(with optional targets). If <pattern-list> consists of multiple items, separate  
them using white space, for example:  
c3t0d0 c3t1d0 c3t2d0 c3t3d0  
specifies fours disks at separate target IDs on controller 3.  
If you enter listat the prompt, the vxdiskadmprogram displays a list of  
the disks available to the system:  
DEVICE  
c2t4d0  
c2t5d0  
c2t6d0  
c3t0d0  
c3t1d0  
c3t2d0  
c3t3d0  
c3t8d0  
c3t9d0  
c3t10d0  
c4t1d0  
c4t2d0  
c4t13d0  
c4t14d0  
.
DISK  
-
-
-
GROUP  
-
-
-
STATUS  
LVM  
LVM  
LVM  
mydg01  
mydg03  
mydg04  
mydg05  
mydg06  
mydg07  
mydg02  
mydg08  
TCd1-18238  
-
mydg  
mydg  
mydg  
mydg  
mydg  
mydg  
mydg  
mydg  
TCg1-18238  
-
online  
online  
online  
online  
online  
online  
online  
online  
online  
online invalid  
online  
-
-
.
.
Select disk devices to add:  
[<pattern-list>,all,list,q,?]  
The phrase online invalidin the STATUSline indicates that a disk has  
yet to be added or initialized for VxVM control. Disks that are listed as  
onlinewith a disk name and disk group are already under VxVM control.  
Enter the device name or pattern of the disks that you want to initialize at  
the prompt and press Return.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Administering disks 99  
Adding a disk to VxVM  
3
4
To continue with the operation, enter y(or press Return) at the following  
prompt:  
Here are the disks selected. Output format: [Device]  
list of device names  
Continue operation? [y,n,q,?] (default: y) y  
At the following prompt, specify the disk group to which the disk should be  
added, or noneto reserve the disks for future use:  
You can choose to add these disks to an existing disk group, a  
new disk group, or you can leave these disks available for use  
by future add or replacement operations. To create a new disk  
group, select a disk group name that does not yet exist. To  
leave the disks available for future use, specify a disk group  
name of “none”.  
Which disk group [<group>,none,list,q,?]  
5
If you specified the name of a disk group that does not already exist,  
vxdiskadmprompts for confirmation that you really want to create this  
new disk group:  
There is no active disk group named disk group name.  
Create a new group named disk group name? [y,n,q,?]  
(default: y)y  
You are then prompted to confirm whether the disk group should support  
the Cross-platform Data Sharing (CDS) feature:  
Create the disk group as a CDS disk group? [y,n,q,?]  
(default: n)  
If the new disk group may be moved between different operating system  
platforms, enter y. Otherwise, enter n.  
6
7
At the following prompt, either press Return to accept the default disk name  
or enter nto allow you to define your own disk names:  
Use default disk names for the disks? [y,n,q,?] (default: y)  
When prompted whether the disks should become hot-relocation spares,  
enter n(or press Return):  
Add disks as spare disks for disk group name? [y,n,q,?]  
(default: n) n  
8
9
When prompted whether to exclude the disks from hot-relocation use, enter  
n(or press Return).  
Exclude disks from hot-relocation use? [y,n,q,?}  
(default: n) n  
You are next prompted to choose whether you want to add a site tag to the  
disks:  
Add site tag to disks? [y,n,q,?] (default: n)  
Download from Www.Somanuals.com. All Manuals Search And Download.  
100 Administering disks  
Adding a disk to VxVM  
A site tag is usually applied to disk arrays or enclosures, and is not required  
unless you want to use the Remote Mirror feature. If you enter yto choose  
to add a site tag, you are prompted to the site name at step 11.  
10 To continue with the operation, enter y(or press Return) at the following  
prompt:  
The selected disks will be added to the disk group  
disk group name with default disk names.  
list of device names  
Continue with operation? [y,n,q,?] (default: y) y  
11 If you chose to tag the disks with a site in step 9, you are now prompted to  
enter the site name that should be applied to the disks in each enclosure:  
The following disk(s):  
list of device names  
belong to enclosure(s):  
list of enclosure names  
Enter site tag for disks on enclosure enclosure_name  
[<name>,q,?] site_name  
12 If one or more disks already contains a file system, vxdiskadmasks if you  
are sure that want to destroy it. Enter yto confirm this:  
The following disk device appears to contain a currently  
unmounted file system.  
list of device names  
Are you sure you want to destroy these file systems  
[y,n,q,?] (default: n) y  
vxdiskadmasks you to confirm that the devices are to be reinitialized before  
proceeding:  
Reinitialize these devices? [y,n,q,?] (default: n) y  
VxVM INFO V-5-2-205 Initializing device device name.  
13 You can now choose whether the disk is to be formatted as a CDS disk that is  
portable between different operating systems, or as a non-portable  
hpdisk-format disk:  
Enter the desired format [cdsdisk,hpdisk,q,?]  
(default: cdsdisk)  
Enter the format that is appropriate for your needs. In most cases, this is  
the default format, cdsdisk.  
14 At the following prompt, vxdiskadmasks if you want to use the default  
private region size of 32768 blocks (32MB). Press Return to confirm that you  
want to use the default value, or enter a different value. (The maximum  
value that you can specify is 524288 blocks.)  
Enter desired private region length [<privlen>,q,?]  
(default: 32768)  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Administering disks 101  
Adding a disk to VxVM  
vxdiskadmthen proceeds to add the disks.  
Adding disk device device name to disk group disk group name with  
disk name disk name.  
.
.
.
Note: To bring LVM disks under VxVM control, use the Migration Utilities.  
See the Veritas Volume Manager Migration Guide for details.  
15 At the following prompt, indicate whether you want to continue to initialize  
more disks (y) or return to the vxdiskadmmain menu (n):  
Add or initialize other disks? [y,n,q,?] (default: n)  
See “Displaying and changing default disk layout attributes” on page 97 for  
details of how to change the default layout that is used to initialize disks.  
Reinitializing a disk  
You can reinitialize a disk that has previously been initialized for use by VxVM  
by putting it under VxVM control as you would a new disk. See “Adding a disk to  
VxVM” on page 97 for details.  
Caution: Reinitialization does not preserve data on the disk. If you want to  
reinitialize the disk, make sure that it does not contain data that should be  
preserved.  
If the disk you want to add has previously been under LVM control, you can  
preserve the data it contains on a VxVM disk by the process of conversion (see  
the Veritas Volume Manager Migration Guide for more details).  
Using vxdiskadd to place a disk under control of VxVM  
As an alternative to vxdiskadm, you can use the vxdiskaddcommand to put a  
disk under VxVM control. For example, to initialize the second disk on the first  
controller, use the following command:  
# vxdiskadd c0t1d0  
The vxdiskaddcommand examines your disk to determine whether it has been  
initialized and also checks for disks that have been added to VxVM, and for  
other conditions.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
102 Administering disks  
Rootability  
Note: If you are adding an uninitialized disk, warning and error messages are  
displayed on the console during the vxdiskaddcommand. Ignore these  
messages. These messages should not appear after the disk has been fully  
initialized; the vxdiskaddcommand displays a success message when the  
initialization completes.  
The interactive dialog for adding a disk using vxdiskadd is similar to that for  
vxdiskadm, described in “Adding a disk to VxVM” on page 97.  
Rootability  
Rootability indicates that the volumes containing the rootfile system and the  
system swap area are under VxVM control. Without rootability, VxVM is usually  
started after the operating system kernel has passed control to the initial user  
mode process at boot time. However, if the volume containing the rootfile  
system is under VxVM control, the kernel starts portions of VxVM before  
starting the first user mode process.  
Under HP-UX, a bootable root disk contains a Logical Interchange Format (LIF)  
area. The LIF LABEL record in the LIF area contains information about the  
starting block number, and the length of the volumes that contain the stand  
and rootfile systems and the system swap area. When a VxVM root disk is  
made bootable, the LIF LABEL record is initialized with volume extent  
information for the stand, root, swap, and dump(if present) volumes.  
See “Setting up a VxVM root disk and mirror” on page 104 for details of how to  
configure a bootable VxVM root disk from an existing LVM root disk.  
Note: From the AR0902 release of HP-UX 11i onward, you can choose to  
configure either a VxVM root disk or an LVM root disk at install time. See the  
HP-UX Installation and Configuration Guide for more information.  
See the chapter “Recovery from Boot Disk Failure” in the Veritas Volume  
Manager Troubleshooting Guide, for information on how to replace a failed boot  
disk.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Administering disks 103  
Rootability  
VxVM root disk volume restrictions  
Volumes on a bootable VxVM root disk have the following configuration  
restrictions:  
■
■
All volumes on the root disk must be in the disk group that you choose to be  
the bootdgdisk group.  
The names of the volumes with entries in the LIF LABEL record must be  
standvol, rootvol, swapvol, and dumpvol(if present). The names of the  
volumes for other file systems on the root disk are generated by appending  
volto the name of their mount point under /.  
■
■
■
Any volume with an entry in the LIF LABEL record must be contiguous. It  
can have only one subdisk, and it cannot span to another disk.  
The rootvoland swapvolvolumes must have the special volume usage  
types rootand swaprespectively.  
Only the disk access types autowith format hpdisk, and simpleare  
suitable for use as VxVM root disks, root disk mirrors, or as hot-relocation  
spares for such disks. An auto-configured cdsdiskformat disk, which  
supports the Cross-platform Data Sharing (CDS) feature, cannot be used.  
The vxcp_lvmrootand vxrootmircommands automatically configure a  
suitable disk type on the physical disks that you specify are to be used as  
VxVM root disks and mirrors.  
■
The volumes on the root disk cannot use dirty region logging (DRL).  
In addition, the size of the private region for disks in a VxVM boot disk group is  
limited to 1MB, rather than the usual default value of 32MB. This restriction is  
necessary to allow the boot loader to find the /standfile system during  
Maintenance Mode Boot.  
Root disk mirrors  
All the volumes on a VxVM root disk may be mirrored. The simplest way to  
achieve this is to mirror the VxVM root disk onto an identically sized or larger  
physical disk. If a mirror of the root disk must also be bootable, the restrictions  
listed in “Booting root volumes” on page 104 also apply to the mirror disk.  
Note: If you mirror only selected volumes on the root disk and use spanning or  
striping to enhance performance, these mirrors are not bootable.  
See “Setting up a VxVM root disk and mirror” on page 104 for details of how to  
create a mirror of a VxVM root disk.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
104 Administering disks  
Rootability  
Booting root volumes  
Note: At boot time, the system firmware provides you with a short time period  
during which you can manually override the automatic boot process and select  
an alternate boot device. For information on how to boot your system from a  
device other than the primary or alternate boot devices, and how to change the  
primary and alternate boot devices, see the HP-UX documentation and the  
boot(1M), pdc(1M) and isl(1M) manual pages.  
Before the kernel mounts the rootfile system, it determines if the boot disk is a  
rootable VxVM disk. If it is such a disk, the kernel passes control to its VxVM  
rootability code. This code extracts the starting block number and length of the  
rootand swapvolumes from the LIF LABEL record, builds temporary volume  
and disk configuration objects for these volumes, and then loads this  
configuration into the VxVM kernel driver. At this point, I/O can take place for  
these temporary rootand swapvolumes by referencing the device number set  
up by the rootability code.  
When the kernel has passed control to the initial user procedure, the VxVM  
configuration daemon (vxconfigd) is started. vxconfigdreads the configuration  
of the volumes in the bootdgdisk group and loads them into the kernel. The  
temporary rootand swapvolumes are then discarded. Further I/O for these  
volumes is performed using the VxVM configuration objects that were loaded  
into the kernel.  
Setting up a VxVM root disk and mirror  
Note: These procedures should be carried out at initlevel 1.  
To set up a VxVM root disk and a bootable mirror of this disk, use the  
vxcp_lvmrootutility. This command initializes a specified physical disk as a  
VxVM root disk named rootdisk##(where ##is the first number starting at  
01that creates a unique disk name), copies the contents of the volumes on the  
LVM root disk to the new VxVM root disk, optionally creates a mirror of the  
VxVM root disk on another specified physical disk, and make the VxVM root  
disk and its mirror (if any) bootable by HP-UX.  
The following example shows how to set up a VxVM root disk on the physical  
disk c0t4d0:  
# /etc/vx/bin/vxcp_lvmroot -b c0t4d0  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Administering disks 105  
Rootability  
Note: The -boption to vxcp_lvmrootuses the setbootcommand to define  
c0t4d0as the primary boot device. If this option is not specified, the primary  
boot device is not changed.  
If the destination VxVM root disk is not big enough to accommodate the  
contents of the LVM root disk, you can use the -Roption to specify a percentage  
by which to reduce the size of the file systems on the target disk. (This takes  
advantage of the fact that most of these file systems are usually nowhere near  
100% full.) For example, to specify a size reduction of 30%, the following  
command would be used:  
# /etc/vx/bin/vxcp_lvmroot -R 30 -v -b c0t4d0  
The verbose option, -v, is specified to give an indication of the progress of the  
operation.  
Caution: Only create a VxVM root disk if you also intend to mirror it. There is no  
benefit in having a non-mirrored VxVM root disk for its own sake.  
The next example uses the same command and additionally specifies the -m  
option to set up a root mirror on disk c1t1d0:  
# /etc/vx/bin/vxcp_lvmroot -m c1t1d0 -R 30 -v -b c0t4d0  
In this example, the -boption to vxcp_lvmrootsets c0t4d0as the primary boot  
device and c1t1d0as the alternate boot device.  
This command is equivalent to using vxcp_lvmrootto create the VxVM-rootable  
disk, and then using the vxrootmircommand to create the mirror:  
# /etc/vx/bin/vxcp_lvmroot -R 30 -v -b c0t4d0  
# /etc/vx/bin/vxrootmir -v -b c1t1d0  
The disk name assigned to the VxVM root disk mirror also uses the format  
rootdisk##with ##set to the next available number.  
Note: The target disk for a mirror that is added using the vxrootmircommand  
must be large enough to accommodate the volumes from the VxVM root disk.  
Once you have successfully rebooted the system from a VxVM root disk to init  
level 1, you can use the vxdestroy_lvmrootcommand to completely remove the  
original LVM root disk (and its associated LVM volume group), and re-use this  
disk as a mirror of the VxVM root disk, as shown in this example:  
# /etc/vx/bin/vxdestroy_lvmroot -v c0t0d0  
# /etc/vx/bin/vxrootmir -v -b c0t0d0  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
106 Administering disks  
Rootability  
Note: You may want to keep the LVM root disk in case you ever need a boot disk  
that does not depend on VxVM being present on the system. However, this may  
require that you update the contents of the LVM root disk in parallel with  
changes that you make to the VxVM root disk. See “Creating an LVM root disk  
from a VxVM root disk” on page 106 for a description of how to create a bootable  
LVM root disk from the VxVM root disk.  
For more information, see the vxcp_lvmroot(1M), vxrootmir(1M),  
vxdestroy_lvmroot(1M) and vxres_lvmroot(1M) manual pages.  
Creating an LVM root disk from a VxVM root disk  
Note: These procedures should be carried out at initlevel 1.  
In some circumstances, it may be necessary to boot the system from an LVM  
root disk. If an LVM root disk is no longer available or an existing LVM root disk  
is out-of-date, you can use the vxres_lvmrootcommand to create an LVM root  
disk on a spare physical disk that is not currently under LVM or VxVM control.  
The contents of the volumes on the existing VxVM root disk are copied to the  
new LVM root disk, and the LVM disk is then made bootable. This operation does  
not remove the VxVM root disk or any mirrors of this disk, nor does it affect  
their bootability.  
Note: The target disk must be large enough to accommodate the volumes from  
the VxVM root disk.  
This example shows how to create an LVM root disk on physical disk c0t1d0  
after removing the existing LVM root disk configuration from that disk.  
# /etc/vx/bin/vxdestroy_lvmroot -v c0t1d0  
# /etc/vx/bin/vxres_lvmroot -v -b c0t1d0  
The -boption to vxres_lvmrootsets c0t1d0as the primary boot device.  
As these operations can take some time, the verbose option, -v, is specified to  
indicate how far the operation has progressed.  
For more information, see the vxres_lvmroot(1M) manual page.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Administering disks 107  
Rootability  
Adding swap volumes to a VxVM rootable system  
To add a swap volume to an HP-UX system with a VxVM root disk  
1
Initialize the disk that is to be used to hold the swap volume (for example,  
c2t5d0), and add it to the boot disk group with the disk media name  
“swapdisk”:  
# /etc/vx/bin/vxdisksetup -i c2t5d0  
# vxdg -g bootdg adddisk swapdisk=c2t5d0  
2
3
4
Create a VxVM volume on swapdisk(with a size of 4 gigabytes in this  
example):  
# vxassist -g bootdg -U swap make swapvol1 4g dm:swapdisk  
In this example, the size of the volume is 4 gigabytes.  
Add the volume to the /etc/fstabfile, and enable the volume as a swap  
device.  
# echo "/dev/vx/dsk/bootdg/swapvol1 - swap defaults 0 0" \  
>> /etc/fstab  
# swapon -a  
View the changed swap configuration:  
# swapinfo  
Adding persistent dump volumes to a VxVM rootable system  
A persistent dump volume is used when creating crash dumps, which are  
eventually saved in the /var/adm/crashdirectory. A maximum of ten VxVM  
volumes can be configured as persistent dump volumes.  
To add a persistent dump volume to an HP-UX system with a VxVM root disk  
1
Create the VxVM volume that is to be used as the dump volume in the boot  
disk group:  
# vxassist -g bootdg make dumpvol 5g  
In this example, the size of the volume is 5 gigabytes.  
2
3
4
Display the configuration of the volume:  
# vxprint -g bootdg  
Display the initial crash dump configuration:  
# crashconf -v  
Add the volume as a persistent dump device to the crash dump  
configuration:  
# crashconf -s /dev/vx/dsk/bootdg/dumpvol  
5
Display the new crash dump configuration:  
# crashconf -v  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
108 Administering disks  
Dynamic LUN expansion  
Removing a persistent dump volume  
Caution: The system will not boot correctly if you delete a dump volume without  
first removing it from the crash dump configuration.  
Use this procedure to remove a dump volume from the crash dump  
configuration.  
To remove a persistent dump volume  
1
Run the following command to remove a VxVM volume that is being used as  
a dump volume from the crash dump configuration:  
# crashconf -ds /dev/vx/dsk/bootdg/dumpvol  
In this example, the dump volume is named dumpvolin the boot disk  
group.  
2
Display the new crash dump configuration:  
# crashconf -v  
You can now remove the volume if required.  
Dynamic LUN expansion  
Note: A license is required to use the dynamic LUN expansion feature.  
The following form of the vxdiskcommand can be used to make VxVM aware of  
the new size of a virtual disk device that has been resized:  
# vxdisk [-f] [-g diskgroup] resize {accessname|medianame} \  
[length=value]  
The device must have a SCSI interface that is presented by a smart switch, smart  
array or RAID controller. Following a resize operation to increase the length  
that is defined for a device, additional disk space on the device is available for  
allocation. You can optionally specify the new size by using the lengthattribute.  
If a disk media name rather than a disk access name is specified, the disk group  
must either be specified using the -goption or the default disk group will be  
used. If the default disk group has not been set up, an error message will be  
generated.  
This facility is provided to support dynamic LUN expansion by updating disk  
headers and other VxVM structures to match a new LUN size. It does not resize  
the LUN itself.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Administering disks 109  
Dynamic LUN expansion  
Any volumes on the device should only be grown after the device itself has first  
been grown. Otherwise, storage other than the device may be used to grow the  
volumes, or the volume resize may fail if no free storage is available.  
Resizing should only be performed on devices that preserve data. Consult the  
array documentation to verify that data preservation is supported and has been  
qualified. The operation also requires that only storage at the end of the LUN is  
affected. Data at the beginning of the LUN must not be altered. No attempt is  
made to verify the validity of pre-existing data on the LUN. The operation  
should be performed on the host where the disk group is imported (or on the  
master node for a cluster-shared disk group).  
Resizing of LUNs that are not part of a disk group is not supported.It is not  
possible to resize LUNs that are in the boot disk group (aliased as bootdg), in a  
deported disk group, or that are offline, uninitialized, being reinitialized, or in  
an error state.  
Caution: Do not perform this operation when replacing a physical disk with a  
disk of a different size as data is not preserved.  
Before reducing the size of a device, any volumes on the device should first be  
reduced in size or moved off the device. By default, the resize fails if any  
subdisks would be disabled as a result of their being removed in whole or in part  
during a shrink operation.  
If the device that is being resized has the only valid configuration copy for a disk  
group, the -foption may be specified to forcibly resize the device.  
Resizing a device that contains the only valid configuration copy for a disk  
group can result in data loss if a system crash occurs during the resize.  
Resizing a virtual disk device is a non-transactional operation outside the  
control of VxVM. This means that the resize command may have to be re-issued  
following a system crash. In addition, a system crash may leave the private  
region on the device in an unusable state. If this occurs, the disk must be  
reinitialized, reattached to the disk group, and its data resynchronized or  
recovered from a backup.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
110 Administering disks  
Removing disks  
Removing disks  
Note: You must disable a disk group as described in “Disabling a disk group” on  
page 207 before you can remove the last disk in that group. Alternatively, you  
can destroy the disk group as described in “Destroying a disk group” on  
page 208.  
You can remove a disk from a system and move it to another system if the disk is  
failing or has failed.  
To prepare your system for the removal of the disk  
1
Stop all activity by applications to volumes that are configured on the disk  
that is to be removed. Unmount file systems and shut down databases that  
are configured on the volumes.  
2
3
Use the following command to stop the volumes:  
# vxvol [-g diskgroup] stop volume1 volume2 ...  
Move the volumes to other disks or back up the volumes. To move a volume,  
use vxdiskadmto mirror the volume on one or more disks, then remove the  
original copy of the volume. If the volumes are no longer needed, they can  
be removed instead of moved.  
4
Check that any data on the disk has either been moved to other disks or is no  
longer needed.  
To remove the disk from its disk group  
1
2
Select menu item 2(Remove a disk)from the vxdiskadmmain menu.  
At the following prompt, enter the disk name of the disk to be removed:  
Remove a disk  
Menu: VolumeManager/Disk/RemoveDisk  
Use this operation to remove a disk from a disk group. This  
operation takes a disk name as input. This is the same name  
that you gave to the disk when you added the disk to the disk  
group.  
Enter disk name [<disk>,list,q,?] mydg01  
3
4
If there are any volumes on the disk, VxVM asks you whether they should be  
evacuated from the disk. If you wish to keep the volumes, answer y.  
Otherwise, answer n.  
At the following verification prompt, press Return to continue:  
VxVM NOTICE V-5-2-284 Requested operation is to remove disk  
mydg01 from group mydg.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Administering disks 111  
Removing disks  
Continue with operation? [y,n,q,?] (default: y)  
The vxdiskadmutility removes the disk from the disk group and displays  
the following success message:  
VxVM INFO V-5-2-268 Removal of disk mydg01 is complete.  
You can now remove the disk or leave it on your system as a replacement.  
5
At the following prompt, indicate whether you want to remove other disks  
(y) or return to the vxdiskadmmain menu (n):  
Remove another disk? [y,n,q,?] (default: n)  
Removing a disk with subdisks  
You can remove a disk on which some subdisks are defined. For example, you  
can consolidate all the volumes onto one disk. If you use the vxdiskadm  
program to remove a disk, you can choose to move volumes off that disk. To do  
this, run the vxdiskadmprogram and select item 2 (Remove a disk)from  
the main menu.  
If the disk is used by some subdisks, the following message is displayed:  
VxVM ERROR V-5-2-369 The following volumes currently use part of  
disk mydg02:  
home usrvol  
Volumes must be moved from mydg02 before it can be removed.  
Move volumes to other disks? [y,n,q,?] (default: n)  
If you choose y, then all subdisks are moved off the disk, if possible. Some  
subdisks are not movable. A subdisk may not be movable for one of the following  
reasons:  
■
There is not enough space on the remaining disks in the subdisk’s disk  
group.  
■
Plexes or striped subdisks cannot be allocated on different disks from  
existing plexes or striped subdisks in the volume.  
If the vxdiskadmprogram cannot move some subdisks, remove some plexes  
from some disks to free more space before proceeding with the disk removal  
operation. See “Removing a volume” on page 290 and “Taking plexes offline” on  
page 230 for information on how to remove volumes and plexes.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
112 Administering disks  
Removing a disk from VxVM control  
Removing a disk with no subdisks  
To remove a disk that contains no subdisks from its disk group, run the  
vxdiskadmprogram and select item 2 (Remove a disk)from the main  
menu, and respond to the prompts as shown in this example to remove mydg02:  
Enter disk name [<disk>,list,q,?] mydg02  
VxVM NOTICE V-5-2-284 Requested operation is to remove disk  
mydg02 from group mydg.  
Continue with operation? [y,n,q,?] (default: y) y  
VxVM INFO V-5-2-268 Removal of disk mydg02 is complete.  
Clobber disk headers? [y,n,q,?] (default: n) y  
Enter yto remove the disk completely from VxVM control. If you do not want to  
remove the disk completely from VxVM control, enter n.  
Removing a disk from VxVM control  
After removing a disk from a disk group, you can permanently remove it from  
Veritas Volume Manager control by running the vxdiskunsetupcommand:  
# /usr/lib/vxvm/bin/vxdiskunsetup c#t#d#  
Caution: The vxdiskunsetupcommand removes a disk from Veritas Volume  
Manager control by erasing the VxVM metadata on the disk. To prevent data  
loss, any data on the disk should first be evacuated from the disk. The  
vxdiskunsetupcommand should only be used by a system administrator who is  
trained and knowledgeable about Veritas Volume Manager.  
Removing and replacing disks  
Note: A replacement disk should have the same disk geometry as the disk that  
failed. That is, the replacement disk should have the same bytes per sector,  
sectors per track, tracks per cylinder and sectors per cylinder, same number of  
cylinders, and the same number of accessible cylinders.  
If failures are starting to occur on a disk, but the disk has not yet failed  
completely, you can replace the disk. This involves detaching the failed or  
failing disk from its disk group, followed by replacing the failed or failing disk  
with a new one. Replacing the disk can be postponed until a later date if  
necessary.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                 
Administering disks 113  
Removing and replacing disks  
To replace a disk  
1
Select menu item 3(Remove a disk for replacement)from the  
vxdiskadmmain menu.  
2
At the following prompt, enter the name of the disk to be replaced (or enter  
listfor a list of disks):  
Remove a disk for replacement  
Menu: VolumeManager/Disk/RemoveForReplace  
Use this menu operation to remove a physical disk from a disk  
group, while retaining the disk name. This changes the state  
for the disk name to a removed disk. If there are any  
initialized disks that are not part of a disk group, you will  
be given the option of using one of these disks as a  
replacement.  
Enter disk name [<disk>,list,q,?] mydg02  
3
When you select a disk to remove for replacement, all volumes that are  
affected by the operation are displayed, for example:  
VxVM NOTICE V-5-2-371 The following volumes will lose mirrors  
as a result of this operation:  
home src  
No data on these volumes will be lost.  
The following volumes are in use, and will be disabled as a  
result of this operation:  
mkting  
Any applications using these volumes will fail future  
accesses. These volumes will require restoration from backup.  
Are you sure you want do this? [y,n,q,?] (default: n)  
To remove the disk, causing the named volumes to be disabled and data to  
be lost when the disk is replaced, enter yor press Return.  
To abandon removal of the disk, and back up or move the data associated  
with the volumes that would otherwise be disabled, enter nor qand press  
Return.  
For example, to move the volume mktingto a disk other than mydg02, use  
this command:  
# vxassist move mkting !mydg02  
After backing up or moving the data in the volumes, start again from step 1  
above.  
4
At the following prompt, either select the device name of the replacement  
disk (from the list provided), press Return to choose the default disk, or  
enter noneif you are going to replace the physical disk:  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
114 Administering disks  
Removing and replacing disks  
The following devices are available as replacements:  
c0t1d0  
You can choose one of these disks now, to replace mydg02.  
Select “none” if you do not wish to select a replacement disk.  
Choose a device, or select “none”  
[<device>,none,q,?] (default: c0t1d0)  
Note: Do not choose the old disk drive as a replacement even though it  
appears in the selection list. If necessary, you can choose to initialize a new  
disk.  
If you enter nonebecause you intend to replace the physical disk, see the  
section “Replacing a failed or removed disk” on page 115.  
5
If you chose to replace the disk in step 4, press Return at the following  
prompt to confirm this:  
VxVM NOTICE V-5-2-285 Requested operation is to remove mydg02  
from group mydg. The removed disk will be replaced with disk  
device c0t1d0.  
Continue with operation? [y,n,q,?] (default: y)  
vxdiskadmdisplays the following messages to indicate that the original  
disk is being removed:  
VxVM NOTICE V-5-2-265 Removal of disk mydg02 completed  
successfully.  
VxVM NOTICE V-5-2-260 Proceeding to replace mydg02 with device  
c0t1d0.  
6
You can now choose whether the disk is to be formatted as a CDS disk that is  
portable between different operating systems, or as a non-portable  
hpdisk-format disk:  
Enter the desired format [cdsdisk,hpdisk,q,?]  
(default: cdsdisk)  
Enter the format that is appropriate for your needs. In most cases, this is  
the default format, cdsdisk.  
7
8
At the following prompt, vxdiskadmasks if you want to use the default  
private region size of 32768 blocks (32 MB). Press Return to confirm that you  
want to use the default value, or enter a different value. (The maximum  
value that you can specify is 524288 blocks.)  
Enter desired private region length [<privlen>,q,?]  
(default: 32768)  
If one of more mirror plexes were moved from the disk, you are now  
prompted whether FastResync should be used to resynchronize the plexes:  
Use FMR for plex resync? [y,n,q,?] (default: n) y  
vxdiskadm displays the following success message:  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Administering disks 115  
Removing and replacing disks  
VxVM NOTICE V-5-2-158 Disk replacement completed successfully.  
9
At the following prompt, indicate whether you want to remove another disk  
(y) or return to the vxdiskadmmain menu (n):  
Remove another disk? [y,n,q,?] (default: n)  
Note: If removing a disk causes one or more volumes to be disabled, see the  
section, “Restarting a Disabled Volume” in the chapter “Recovery from  
Hardware Failure” in the Veritas Volume Manager Troubleshooting Guide, for  
information on how to restart a disabled volume so that you can restore its data  
from a backup.  
If you wish to move hot-relocate subdisks back to a replacement disk, see  
“Configuring hot-relocation to use only spare disks” on page 390.  
Replacing a failed or removed disk  
Note: You may need to run commands that are specific to the operating system  
or disk array when replacing a physical disk.  
To specify a disk that has replaced a failed or removed disk  
1
Select menu item 4(Replace a failed or removed disk)from the  
vxdiskadmmain menu.  
2
At the following prompt, enter the name of the disk to be replaced (or enter  
listfor a list of disks):  
Replace a failed or removed disk  
Menu: VolumeManager/Disk/ReplaceDisk  
VxVM INFO V-5-2-479 Use this menu operation to specify a  
replacement disk for a disk that you removed with the “Remove  
a disk for replacement” menu operation, or that failed during  
use. You will be prompted for a disk name to replace and a disk  
device to use as a replacement.  
You can choose an uninitialized disk, in which case the disk  
will be initialized, or you can choose a disk that you have  
already initialized using the Add or initialize a disk menu  
operation.  
Select a removed or failed disk [<disk>,list,q,?] mydg02  
3
The vxdiskadmprogram displays the device names of the disk devices  
available for use as replacement disks. Your system may use a device name  
that differs from the examples. Enter the device name of the disk or press  
Return to select the default device:  
The following devices are available as replacements:  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
116 Administering disks  
Removing and replacing disks  
c0t1d0 c1t1d0  
You can choose one of these disks to replace mydg02.  
Choose "none" to initialize another disk to replace mydg02.  
Choose a device, or select "none"  
[<device>,none,q,?] (default: c0t1d0)  
4
Depending on whether the replacement disk was previously initialized,  
perform the appropriate step from the following:  
◆
If the disk has not previously been initialized, press Return at the following  
prompt to replace the disk:  
VxVM INFO V-5-2-378 The requested operation is to initialize  
disk device c0t1d0 and to then use that device to  
replace the removed or failed disk mydg02 in disk group mydg.  
Continue with operation? [y,n,q,?] (default: y)  
◆
If the disk has already been initialized, press Return at the following prompt  
to replace the disk:  
VxVM INFO V-5-2-382 The requested operation is to use the  
initialized device c0t1d0 to replace the removed or  
failed disk mydg02 in disk group mydg.  
Continue with operation? [y,n,q,?] (default: y)  
5
You can now choose whether the disk is to be formatted as a CDS disk that is  
portable between different operating systems, or as a non-portable  
hpdisk-format disk:  
Enter the desired format [cdsdisk,hpdisk,q,?]  
(default: cdsdisk)  
Enter the format that is appropriate for your needs. In most cases, this is  
the default format, cdsdisk.  
6
7
At the following prompt, vxdiskadmasks if you want to use the default  
private region size of 32768 blocks (32 MB). Press Return to confirm that you  
want to use the default value, or enter a different value. (The maximum  
value that you can specify is 524288 blocks.)  
Enter desired private region length [<privlen>,q,?]  
(default: 32768)  
The vxdiskadmprogram then proceeds to replace the disk, and returns the  
following message on success:  
VxVM NOTICE V-5-2-158 Disk replacement completed successfully.  
At the following prompt, indicate whether you want to replace another disk  
(y) or return to the vxdiskadmmain menu (n):  
Replace another disk? [y,n,q,?] (default: n)  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Administering disks 117  
Enabling a disk  
8
After using the vxdiskadmcommand to replace one or more failed disks in a  
VxVM cluster, run the following command on all the cluster nodes:  
# vxdctl enable  
Then run the following command on the master node:  
# vxreattach -r accesname  
where accessname is the disk access name (such as c0t1d0). This initiates  
the recovery of all the volumes on the disks.  
Alternatively, halt the cluster, reboot all the cluster nodes, and restart the  
cluster.  
If you do not perform this step for a cluster, the nodes may not be able to see  
the disks, and the error Device path not validwill be displayed.  
Enabling a disk  
If you move a disk from one system to another during normal system operation,  
VxVM does not recognize the disk automatically. The enable disk task enables  
VxVM to identify the disk and to determine if this disk is part of a disk group.  
Also, this task re-enables access to a disk that was disabled by either the disk  
group deport task or the disk device disable (offline) task.  
To enable a disk  
1
Select menu item 9 (Enable (online) a disk device)from the  
vxdiskadmmain menu.  
2
At the following prompt, enter the device name of the disk to be enabled (or  
enter listfor a list of devices):  
Enable (online) a disk device  
Menu: VolumeManager/Disk/OnlineDisk  
VxVM INFO V-5-2-998 Use this operation to enable access to a  
disk that was disabled with the “Disable (offline) a disk  
device” operation.  
You can also use this operation to re-scan a disk that may have  
been changed outside of the Volume Manager. For example, if a  
disk is shared between two systems, the Volume Manager running  
on the other system may have changed the disk. If so, you can  
use this operation to re-scan the disk.  
NOTE: Many vxdiskadm operations re-scan disks without user  
intervention. This will eliminate most needs to online a  
disk directly, except when the disk is directly offlined.  
Select a disk device to enable [<address>,list,q,?]  
c0t2d0  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
118 Administering disks  
Taking a disk offline  
vxdiskadmenables the specified device.  
3
At the following prompt, indicate whether you want to enable another  
device (y) or return to the vxdiskadmmain menu (n):  
Enable another device? [y,n,q,?] (default: n)  
Taking a disk offline  
There are instances when you must take a disk offline. If a disk is corrupted, you  
must disable the disk before removing it. You must also disable a disk before  
moving the physical disk device to another location to be connected to another  
system.  
Note: Taking a disk offline is only useful on systems that support hot-swap  
removal and insertion of disks without needing to shut down and reboot the  
system.  
To take a disk offline  
1
Select menu item 10 (Disable (offline) a disk device)from the  
vxdiskadmmain menu.  
2
At the following prompt, enter the address of the disk you want to disable:  
Disable (offline) a disk device  
Menu: VolumeManager/Disk/OfflineDisk  
VxVM INFO V-5-2-474 Use this menu operation to disable all  
access to a disk device by the Volume Manager. This operation  
can be applied only to disks that are not currently in a disk  
group. Use this operation if you intend to remove a disk from  
a system without rebooting.  
NOTE: Many systems do not support disks that can be removed  
from a system during normal operation. On such systems, the  
offline operation is seldom useful.  
Select a disk device to disable [<address>,list,q,?]  
c0t2d0  
The vxdiskadmprogram disables the specified disk.  
3
At the following prompt, indicate whether you want to disable another  
device (y) or return to the vxdiskadmmain menu (n):  
Disable another device? [y,n,q,?] (default: n)  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Administering disks 119  
Renaming a disk  
Renaming a disk  
If you do not specify a VM disk name, VxVM gives the disk a default name when  
you add the disk to VxVM control. The VM disk name is used by VxVM to  
identify the location of the disk or the disk type. To change the disk name to  
reflect a change of use or ownership, use the following command:  
# vxedit [-g diskgroup] rename old_diskname new_diskname  
For example, you might want to rename disk mydg03, as shown in the following  
output from vxdisklist, to mydg02: #  
# vxdisk list  
DEVICE  
c0t0d0  
c1t0d0  
c1t1d0  
TYPE  
DISK  
mydg01  
mydg03  
-
GROUP  
mydg  
mydg  
-
STATUS  
online  
online  
online  
auto:hpdisk  
auto:hpdisk  
auto:hpdisk  
You would use the following command to rename the disk.  
# vxedit -g mydg rename mydg03 mydg02  
To confirm that the name change took place, use the vxdisklistcommand  
again:  
# vxdisk list  
DEVICE  
c0t0d0  
c1t0d0  
c1t1d0  
TYPE  
DISK  
mydg01  
mydg02  
-
GROUP  
mydg  
mydg  
-
STATUS  
online  
online  
online  
auto:hpdisk  
auto:hpdisk  
auto:hpdisk  
Note: By default, VxVM names subdisk objects after the VM disk on which they  
are located. Renaming a VM disk does not automatically rename the subdisks on  
that disk.  
Reserving disks  
By default, the vxassistcommand allocates space from any disk that has free  
space. You can reserve a set of disks for special purposes, such as to avoid  
general use of a particularly slow or a particularly fast disk.  
To reserve a disk for special purposes, use the following command:  
# vxedit [-g diskgroup] set reserve=on diskname  
After you enter this command, the vxassistprogram does not allocate space  
from the selected disk unless that disk is specifically mentioned on the vxassist  
command line. For example, if mydg03is reserved, use the following command:  
# vxassist [-g diskgroup] make vol03 20m mydg03  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                 
120 Administering disks  
Displaying disk information  
The vxassistcommand overrides the reservation and creates a 20 megabyte  
volume on mydg03. However, the command:  
# vxassist -g mydg make vol04 20m  
does not use mydg03, even if there is no free space on any other disk.  
To turn off reservation of a disk, use the following command:  
# vxedit [-g diskgroup] set reserve=off diskname  
See the vxedit(1M) manual page for more information.  
Displaying disk information  
Before you use a disk, you need to know if it has been initialized and placed  
under VxVM control. You also need to know if the disk is part of a disk group,  
because you cannot create volumes on a disk that is not part of a disk group. The  
vxdisklistcommand displays device names for all recognized disks, the disk  
names, the disk group names associated with each disk, and the status of each  
disk.  
To display information on all disks that are known to VxVM, use the following  
command:  
# vxdisk list  
VxVM returns a display similar to the following:  
DEVICE  
c0t0d0  
c1t0d0  
c1t1d0  
enc0_2  
enc0_3  
enc0_0  
enc0_1  
TYPE  
DISK  
mydg04  
mydg03  
-
mydg02  
mydg05  
-
GROUP  
mydg  
mydg  
-
mydg  
mydg  
-
STATUS  
online  
online  
online invalid  
online  
online  
online  
online  
auto:hpdisk  
auto:hpdisk  
auto:hpdisk  
auto:hpdisk  
auto:hpdisk  
auto:hpdisk  
auto:hpdisk  
-
-
The phrase online invalidin the STATUSline indicates that a disk has not  
yet been added to VxVM control. These disks may or may not have been  
initialized by VxVM previously. Disks that are listed as onlineare already  
under VxVM control.  
VxVM cannot access stale device entries in the /dev/diskand /dev/rdisk  
directories. I/O cannot be performed to such devices, which are shown as being  
in the errorstate.  
To display details on a particular disk that is defined to VxVM, use the following  
command:  
# vxdisk [-v] list diskname  
The -voption causes the command to additionally list all tags and tag values  
that are defined for the disk. Without this option, no tags are displayed.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
Administering disks 121  
Displaying disk information  
Displaying disk information with vxdiskadm  
Displaying disk information shows you which disks are initialized, to which disk  
groups they belong, and the disk status. The listcommand displays device  
names for all recognized disks, the disk names, the disk group names associated  
with each disk, and the status of each disk.  
To display disk information  
1
Start the vxdiskadmprogram, and select list (List disk information)  
from the main menu.  
2
At the following display, enter the address of the disk you want to see, or  
enter allfor a list of all disks:  
List disk information  
Menu: VolumeManager/Disk/ListDisk  
VxVM INFO V-5-2-475 Use this menu operation to display a list of  
disks. You can also choose to list detailed information about  
the disk at a specific disk device address.  
Enter disk device or "all" [<address>,all,q,?] (default: all)  
■
■
If you enter all, VxVM displays the device name, disk name, group,  
and status.  
If you enter the address of the device for which you want information,  
complete disk information (including the device name, the type of disk,  
and information about the public and private areas of the disk) is  
displayed.  
Once you have examined this information, press Return to return to the  
main menu.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
122 Administering disks  
Controlling Powerfail Timeout  
Controlling Powerfail Timeout  
Powerfail Timeout is an attribute of a SCSI disk connected to an HP-UX host.  
This is used to detect and handle I/O on non-responding disks.  
See the pfto(7) man page.  
VxVM uses this mechanism in its Powerfail Timeout (pfto) feature. You can  
specify a timeout value for individual VxVM disks using the vxdiskcommand. If  
a disk fails to respond in the specified timeout period, the driver receives a timer  
interrupt.  
You can set the PFTO values on a disk or a set of disks within a disk group using  
the CLI. PFTO helps in preventing system hangs due to non-responding disks.  
However, in some circumstances, it is critical to retry the I/O until it succeeds,  
regardless of how long it takes. In those cases, you can disable PFTO. By default,  
the use of PFTO is enabled.  
Setting the PFTO values  
To set the PFTO value on a disk, use the following command:  
$ vxdisk -g dg_name set disk_name pfto=value  
For example, to set the PFTO value of 50sec on the disk c5t0d6:  
$ vxdisk -g testdg set c5t0d6 pfto=50  
To set the PFTO value on a disk group, use the following command:  
$ vxpfto -g dg_name -t value  
For example, to set the PFTO value to 50 seconds on all disks in the diskgroup  
testdg:  
$ vxpfto -g testdg -t 50  
For more details, see the vxpfto(1M) and vxdisk(1M) manual pages.  
Displaying the PFTO values  
To display the PFTO value and whether PFTO is enabled or disabled for a disk,  
use one of the following commands:  
$ vxprint -g dg_name -l disk_name  
$ vxdisk -g dg_name list disk_name  
The output shows the pftostatefield, which indicates whether PFTO is  
enabled or disabled. The timeoutfield shows the PFTO timeout value. For  
example, the output for the c5t0d6disk shows:  
Device:  
c5t0d6  
devicetag: c5t0d6  
...  
timeout:  
30  
pftostate: disabled  
...  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Administering disks 123  
Controlling Powerfail Timeout  
Enabling or disabling PFTO  
To enable or disable PFTO on a disk, use the following command:  
$ vxdisk -g dg_name set disk_name pftostate={enabled|disabled}  
For example, to disable PFTO on the disk c5t0d6:  
$ vxdisk -g testdg set c5t0d6 pftostate=disabled  
To enable or disable PFTO on a disk group, use the following command:  
$ vxpfto -g dg_name -o pftostate={enabled|disabled}  
For example, to disable PFTO on all disks in the diskgroup testdg:  
$ vxpfto -g testdg -o pftostate=disabled  
For more details, see the vxpfto(1M) and vxdisk(1M) manual pages.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
124 Administering disks  
Controlling Powerfail Timeout  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter  
3
Administering dynamic  
multipathing (DMP)  
The dynamic multipathing (DMP) feature of Veritas Volume Manager (VxVM)  
provides greater availability, reliability and performance by using path failover  
and load balancing. This feature is available for multiported disk arrays from  
various vendors.  
DMP coexists with the native multipathing in HP-UX. For more information, see  
“DMP coexistence with HP-UX native multipathing” on page 130.  
How DMP works  
Multiported disk arrays can be connected to host systems through multiple  
paths. To detect the various paths to a disk, DMP uses a mechanism that is  
specific to each supported array type. DMP can also differentiate between  
different enclosures of a supported array type that are connected to the same  
host system.  
See “Discovering and configuring newly added disk devices” on page 82 for a  
description of how to make newly added disk hardware known to a host system.  
The multipathing policy used by DMP depends on the characteristics of the disk  
array:  
■
An Active/Passive array (A/P array) allows access to its LUNs (logical units;  
real disks or virtual disks created using hardware) via the primary (active)  
path on a single controller (also known as an access port or a storage  
processor) during normal operation.  
In implicit failover mode (or autotrespass mode), an A/P array automatically  
fails over by scheduling I/O to the secondary (passive) path on a separate  
controller if the primary path fails. This passive port is not used for I/O  
until the active port fails. In A/P arrays, path failover can occur for a single  
LUN if I/O fails on the primary path.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                             
126 Administering dynamic multipathing (DMP)  
How DMP works  
For Active/Passive arrays with LUN group failover (A/PG arrays), a group of  
LUNs that are connected through a controller is treated as a single failover  
entity. Unlike A/P arrays, failover occurs at the controller level, and not for  
individual LUNs. The primary and secondary controller are each connected  
to a separate group of LUNs. If a single LUN in the primary controller’s LUN  
group fails, all LUNs in that group fail over to the secondary controller.  
Active/Passive arrays in explicit failover mode (or non-autotrespass mode)  
are termed A/PF arrays. DMP issues the appropriate low-level command to  
make the LUNs fail over to the secondary path.  
The paths of an Active/Passive array are not considered to be on different  
controllers when mirroring across controllers (for example, when creating a  
volume using vxassistmakespecified with the mirror=ctlrattribute).  
A/P-C, A/PF-C and A/PG-C arrays are variants of the A/P, A/PF and A/PG  
array types that support concurrent I/O and load balancing by having  
multiple primary paths into a controller. This functionality is provided by a  
controller with multiple ports, or by the insertion of a SAN hub or switch  
between an array and a controller. Failover to the secondary (passive) path  
occurs only if all the active primary paths fail.  
■
An Active/Active disk array (A/A arrays) permits several paths to be used  
concurrently for I/O. Such arrays allow DMP to provide greater I/O  
throughput by balancing the I/O load uniformly across the multiple paths to  
the LUNs. In the event that one path fails, DMP automatically routes I/O  
over the other available paths.  
A/A-A or Asymmetric Active/Active arrays can be accessed through  
secondary storage paths with little performance degradation. Usually an  
A/A-A array behaves like an A/P array rather than an A/A array. However,  
during failover, an A/A-A array behaves like an A/A array.  
Note: An array support library (ASL) may define additional array types for the  
arrays that it supports.  
VxVM uses DMP metanodes (DMP nodes) to access disk devices connected to the  
system. For each disk in a supported array, DMP maps one node to the set of  
paths that are connected to the disk. Additionally, DMP associates the  
appropriate multipathing policy for the disk array with the node. For disks in an  
unsupported array, DMP maps a separate node to each path that is connected to  
a disk. The raw and block devices for the nodes are created in the directories  
/dev/vx/rdmpand /dev/vx/dmprespectively.  
Figure 3-1 illustrates how DMP sets up a node for a disk in a supported disk  
array.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                               
Administering dynamic multipathing (DMP) 127  
How DMP works  
Figure 3-1  
How DMP represents multiple physical paths to a disk as one node  
VxVM  
Host  
c1  
c2  
Single DMP node  
Multiple paths  
Mapped by DMP  
DMP  
Multiple paths  
Disk  
As described in “Enclosure-based naming” on page 23, VxVM implements a disk  
device naming scheme that allows you to recognize to which array a disk  
belongs. Figure 3-2, shows an example where two paths, c1t99d0and c2t99d0,  
exist to a single disk in the enclosure, but VxVM uses the single DMP node,  
enc0_0, to access it.  
Figure 3-2  
Example of multipathing for a disk enclosure in a SAN environment  
VxVM  
Host  
c1  
c2  
enc0_0  
Mapped by DMP  
DMP  
Fibre Channel hubs  
or switches  
c1t99d0  
c2t99d0  
Disk enclosure enc0  
Disk is c1t99d0 or c2t99d0  
depending on the path  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
128 Administering dynamic multipathing (DMP)  
How DMP works  
See “Changing the disk-naming scheme” on page 91 for details of how to change  
the naming scheme that VxVM uses for disk devices.  
See “Discovering and configuring newly added disk devices” on page 82 for a  
description of how to make newly added disk hardware known to a host system.  
How DMP monitors I/O on paths  
In older releases of VxVM, DMP had one kernel daemon (errord) that performed  
error processing, and another (restored) that performed path restoration  
activities.  
From release 5.0, DMP maintains a pool of kernel threads that are used to  
perform such tasks as error processing, path restoration, statistics collection,  
and SCSI request callbacks. The vxdmpadmstatcommand can be used to provide  
information about the threads. The names errordand restoredhave been  
retained for backward compatibility.  
One kernel thread responds to I/O failures on a path by initiating a probe of the  
host bus adapter (HBA) that corresponds to the path. Another thread then takes  
the appropriate action according to the response from the HBA. The action  
taken can be to retry the I/O request on the path, or to fail the path and  
reschedule the I/O on an alternate path.  
The restore kernel thread is woken periodically (typically every 5 minutes) to  
check the health of the paths, and to resume I/O on paths that have been  
restored. As some paths may suffer from intermittent failure, I/O is only  
resumed on a path if has remained healthy for a given period of time (by default,  
5 minutes). DMP can be configured with different policies for checking the paths  
as described in “Configuring DMP path restoration policies” on page 160.  
The statistics-gathering thread records the start and end time of each I/O  
request, and the number of I/O failures and retries on each path. DMP can be  
configured to use this information to prevent the SCSI driver being flooded by  
I/O requests. This feature is known as I/O throttling.  
If an I/O request relates to a mirrored volume, VxVM specifies the FAILFAST  
flag. In such cases, DMP does not retry failed I/O requests on the path, and  
instead marks the disks on that path as having failed.  
See “Path failover mechanism” on page 128 and “I/O throttling” on page 129 for  
more information about these features of DMP.  
Path failover mechanism  
The DMP feature of VxVM enhances system reliability when used with  
multiported disk arrays. In the event of the loss of a path to a disk array, DMP  
automatically selects the next available path for I/O requests without  
intervention from the administrator.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
               
Administering dynamic multipathing (DMP) 129  
How DMP works  
DMP is also informed when a connection is repaired or restored, and when you  
add or remove devices after the system has been fully booted (provided that the  
operating system recognizes the devices correctly).  
If required, the response of DMP to I/O failure on a path can be tuned for the  
paths to individual arrays. DMP can be configured to time out an I/O request  
either after a given period of time has elapsed without the request succeeding,  
or after a given number of retries on a path have failed.  
For information about how to configure the behavior of DMP in response to I/O  
failure on a path, see “Configuring the response to I/O failures” on page 156.  
I/O throttling  
If I/O throttling is enabled, and the number of outstanding I/O requests builds  
up on a path that has become less responsive, DMP can be configured to prevent  
new I/O requests being sent on the path either when the number of outstanding  
I/O requests has reached a given value, or a given time has elapsed since the last  
successful I/O request on the path. While throttling is applied to a path, the  
outstanding I/O requests on that path are scheduled on other available paths.  
The throttling is removed from the path if the HBA reports no error on the path,  
or if an outstanding I/O request on the path succeeds.  
For information about how to configure I/O throttling on a path, see  
“Configuring the I/O throttling mechanism” on page 157.  
Load balancing  
By default, DMP uses a minimum queue length policy (minimumq) to provide load  
balancing across paths for Active/Active disk arrays. I/O is sent down the path  
that has the minimum number of outstanding I/O requests in the queue for a  
LUN.  
For Active/Passive and Asymmetric Active/Active disk arrays, the default  
round-robinpolicy shares I/O equally between the available paths in a  
round-robin sequence. If the primary path fails, I/O is switched over to another  
available primary or secondary path. As the continuous transfer of ownership of  
LUNs from one controller to another results in severe I/O slowdown, load  
balancing across paths is not performed for Active/Passive disk arrays unless  
they support concurrent I/O.  
You can use the vxdmpadmcommand to change the I/O policy for the paths to an  
enclosure or disk array.  
See “Specifying the I/O policy” on page 147.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
130 Administering dynamic multipathing (DMP)  
How DMP works  
DMP coexistence with HP-UX native multipathing  
The HP-UX 11i v3 release includes support for native multipathing, which can  
coexist with DMP. HP-UX native multipathing creates a persistent (agile) device  
in the /dev/diskand /dev/rdiskdirectories for each disk that can be  
accessed by one or more physical paths. To maintain backward compatibility,  
HP-UX also creates legacy devices in the /dev/dskand /dev/rdskdirectories.  
VxVM recreates disk devices for all paths in the operating system’s hardware  
device tree as DMP nodes in the /dev/vx/dmpand /dev/vx/rdmpdirectories,  
independently of the devices that are listed in the /dev/dskand /dev/rdsk  
directories. VxVM uses a DMP node to represent a disk that can be accessed by  
one or more physical paths. DMP nodes are not used by the native multipathing  
feature of HP-UX.  
VxVM commands display device names according the naming scheme that has  
been selected.  
See “Changing the disk-naming scheme” on page 91.  
By default, VxVM is configured to use DMP metanodes. If you want to use  
HP-UX native multipathing, you must add the HP-UX native multipathing  
metanodes as foreign devices.  
See “Adding foreign devices” on page 89.  
See “Migrating between DMP and HP-UX native multipathing” on page 130.  
For instructions on administering native multipathing with Base-VxVMand  
VxVM-Full, please consult the Release Notes for VxVM 5.0 on 11i v3.  
Migrating between DMP and HP-UX native multipathing  
You can use the vxddladm addforeignand vxddladm rmforeigncommands to  
migrate a system between using DMP and using HP-UX native multipathing.  
These procedures migrate all devices in the /dev/diskand /dev/rdisk  
directories.  
Caution: Before migrating between DMP and HP-UX native multipathing, ensure  
that no applications are accessing VxVM volumes. Migration is not supported  
without first stopping any applications that are using the volumes.  
To migrate from DMP to HP-UX native multipathing  
1
Stop all the volumes in each disk group on the system:  
# vxvol -g diskgroup stopall  
2
Use the following commands to initiate the migration:  
# vxddladm addforeign blockdir=/dev/disk chardir=/dev/rdisk  
# vxconfigd -kr reset  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Administering dynamic multipathing (DMP) 131  
How DMP works  
3
Restart all the volumes in each disk group:  
# vxvol -g diskgroup startall  
The output from the vxdisk listcommand now shows only HP-UX native  
multipathing metanode names, for example:  
# vxdisk list  
DEVICE  
TYPE  
auto:LVM  
auto:LVM  
auto:cdsdisk  
auto:cdsdisk  
auto:cdsdisk  
auto:cdsdisk  
auto:cdsdisk  
auto:cdsdisk  
DISK  
GROUP  
STATUS  
LVM  
LVM  
online  
online  
online  
online  
online  
online  
disk155  
disk156  
disk224  
disk225  
disk226  
disk227  
disk228  
disk229  
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
When HP-UX native multipathing is configured, no DMP metanodes are  
configured for the devices in the /dev/diskand /dev/rdiskdirectories.  
As a result, the vxdisk listcommand only displays the names of the  
HP-UX native multipathing metanodes, and cannot display legacy names  
for the devices.  
To migrate from HP-UX native multipathing to DMP  
1
Stop all the volumes in each disk group on the system:  
# vxvol -g diskgroup stopall  
2
Use the following commands to initiate the migration:  
# vxddladm rmforeign blockdir=/dev/disk chardir=/dev/rdisk  
# vxconfigd -kr reset  
3
Restart all the volumes in each disk group:  
# vxvol -g diskgroup startall  
The output from the vxdisk listcommand now shows DMP metanode  
names according to the current naming scheme. For example, under the  
defaultor legacynaming scheme, vxdisk listdisplays the devices as:  
# vxdisk list  
DEVICE  
c2t0d0  
c3t2d0  
c89t0d0  
c89t0d1  
c89t0d2  
c89t0d3  
c89t0d4  
c89t0d5  
TYPE  
auto:LVM  
auto:LVM  
auto:cdsdisk  
auto:cdsdisk  
auto:cdsdisk  
auto:cdsdisk  
auto:cdsdisk  
auto:cdsdisk  
DISK  
GROUP  
STATUS  
LVM  
LVM  
online  
online  
online  
online  
online  
online  
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Download from Www.Somanuals.com. All Manuals Search And Download.  
132 Administering dynamic multipathing (DMP)  
How DMP works  
and under the newnaming scheme as:  
# vxdisk list  
DEVICE  
TYPE  
auto:LVM  
auto:LVM  
auto:cdsdisk  
auto:cdsdisk  
auto:cdsdisk  
auto:cdsdisk  
auto:cdsdisk  
auto:cdsdisk  
DISK  
GROUP  
STATUS  
LVM  
LVM  
online  
online  
online  
online  
online  
online  
disk155  
disk156  
disk224  
disk225  
disk226  
disk227  
disk228  
disk229  
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
See “Changing the disk-naming scheme” on page 91.  
DMP in a clustered environment  
Note: You need an additional license to use the cluster feature of VxVM.  
In a clustered environment where Active/Passive type disk arrays are shared by  
multiple hosts, all nodes in the cluster must access the disk via the same  
physical path. Accessing a disk via multiple paths simultaneously can severely  
degrade I/O performance (sometimes referred to as the ping-pong effect). Path  
failover on a single cluster node is also coordinated across the cluster so that all  
the nodes continue to share the same physical path.  
Prior to release 4.1 of VxVM, the clustering and DMP features could not handle  
automatic failback in A/P arrays when a path was restored, and did not support  
failback for explicit failover mode arrays. Failback could only be implemented  
manually by running the vxdctlenablecommand on each cluster node after  
the path failure had been corrected. In release 4.1, failback is now an automatic  
cluster-wide operation that is coordinated by the master node. Automatic  
failback in explicit failover mode arrays is also handled by issuing the  
appropriate low-level command. If required, this feature can be disabled by  
selecting the “no failback” option that is defined in the array policy module  
(APM) for an array.  
Note: Support for automatic failback of an A/P array requires that an  
appropriate ASL (and APM, if required) is available for the array, and has been  
installed on the system. See “Administering the Device Discovery Layer” on  
page 85 and “Configuring array policy modules” on page 162.  
For Active/Active type disk arrays, any disk can be simultaneously accessed  
through all available physical paths to it. In a clustered environment, the nodes  
do not all need to access a disk via the same physical path.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Administering dynamic multipathing (DMP) 133  
Disabling and enabling multipathing for specific devices  
Enabling or disabling controllers with shared disk groups  
Prior to release 5.0, VxVM did not allow enabling or disabling of paths or  
controllers connected to a disk that is part of a shared Veritas Volume Manager  
disk group. From VxVM 5.0 onward, such operations are supported on shared  
DMP nodes in a cluster.  
Disabling and enabling multipathing for specific  
devices  
You can use vxdiskadmmenu options 17 and 18 to disable or enable  
multipathing. These menu options also allow you to exclude or exclude devices  
from the view of VxVM.  
See “Disabling multipathing and making devices invisible to VxVM” on  
page 133.  
See “Enabling multipathing and making devices visible to VxVM” on page 134.  
Disabling multipathing and making devices invisible to VxVM  
Note: Some of the operations described in this section require a reboot of the  
system.  
1
2
Select menu task 17 (Prevent multipathing/Suppress devices from  
VxVM’s view) from the vxdiskadmmain menu to prevent a device from  
being multipathed by the VxVM DMP driver (vxdmp), or to exclude a device  
from the view of VxVM. You are prompted to confirm whether you want to  
continue.  
Select the operation you want to perform from the displayed list:  
1 Suppress all paths through a controller from VxVM’s view  
2 Suppress a path from VxVM’s view  
3 Suppress disks from VxVM’s view by specifying a VID:PID  
combination  
4 Suppress all but one paths to a disk  
5 Prevent multipathing of all disks on a controller by VxVM  
6 Prevent multipathing of a disk by VxVM  
7 Prevent multipathing of disks by specifying a VID:PID  
combination  
8 List currently suppressed/non-multipathed devices  
? Display help about menu  
?? Display help about the menuing system  
q Exit from menus  
Help text and examples are provided onscreen for all the menu items.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
134 Administering dynamic multipathing (DMP)  
Disabling and enabling multipathing for specific devices  
◆
Select option 1 to exclude all paths through the specified controller from the  
view of VxVM. These paths remain in the disabled state until the next  
reboot, or until the paths are re-included.  
◆
◆
Select option 2 to exclude specified paths from the view of VxVM.  
Select option 3 to exclude disks from the view of VxVM that match a  
specified Vendor ID and Product ID.  
◆
Select option 4 to define a pathgroup for disks that are not multipathed by  
VxVM. (A pathgroup explicitly defines alternate paths to the same disk.)  
Only one path is made visible to VxVM.  
◆
◆
Select option 5 to disable multipathing for all disks on a specified controller.  
Select option 6 to disable multipathing for specified paths. The disks that  
correspond to a specified path are claimed in the OTHER_DISKS category  
and are not multipathed.  
◆
◆
Select option 7 to disable multipathing for disks that match a specified  
Vendor ID and Product ID. The disks that correspond to a specified Vendor  
ID and Product ID combination are claimed in the OTHER_DISKS category  
and are not multipathed.  
Select option 8 to list the devices that are currently suppressed or not  
multipathed.  
Enabling multipathing and making devices visible to VxVM  
Note: Some of the operations described in this section require a reboot of the  
system.  
1
2
Select menu item 18 (Allow multipathing/Unsuppress devices from  
VxVM’s view) from the vxdiskadmmain menu to re-enable multipathing for  
a device, or to make a device visible to VxVM again. You are prompted to  
confirm whether you want to continue.  
Select the operation you want to perform from the displayed list:  
1 Unsuppress all paths through a controller from VxVM’s view  
2 Unsuppress a path from VxVM’s view  
3 Unsuppress disks from VxVM’s view by specifying a VID:PID  
combination  
4 Remove a pathgroup definition  
5 Allow multipathing of all disks on a controller by VxVM  
6 Allow multipathing of a disk by VxVM  
7 Allow multipathing of disks by specifying a VID:PID  
combination  
8 List currently suppressed/non-multipathed devices  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Administering dynamic multipathing (DMP) 135  
Disabling and enabling multipathing for specific devices  
? Display help about menu  
?? Display help about the menuing system  
q Exit from menus  
◆
Select option 1 to make all paths through a specified controller visible to  
VxVM.  
◆
◆
Select option 2 to make specified paths visible to VxVM.  
Select option 3 to make disks visible to VxVM that match a specified Vendor  
ID and Product ID.  
◆
◆
Select option 4 to remove a pathgroup definition. (A pathgroup explicitly  
defines alternate paths to the same disk.) Once a pathgroup has been  
removed, all paths that were defined in that pathgroup become visible again.  
Select option 5 to enable multipathing for all disks that have paths through  
the specified controller.  
◆
◆
Select option 6 to enable multipathing for specified paths.  
Select option 7 to enable multipathing for disks that match a specified  
Vendor ID and Product ID.  
◆
Select option 8 to list the devices that are currently suppressed or not  
multipathed.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
136 Administering dynamic multipathing (DMP)  
Enabling and disabling I/O for controllers and storage processors  
Enabling and disabling I/O for controllers and  
storage processors  
DMP allows you to turn off I/O for a controller or the array port of a storage  
processor so that you can perform administrative operations. This feature can  
be used for maintenance of HBA controllers on the host, or array ports that are  
attached to disk arrays supported by VxVM. I/O operations to the controller or  
array port can be turned back on after the maintenance task is completed. You  
can accomplish these operations using the vxdmpadmcommand provided with  
VxVM.  
In Active/Active type disk arrays, VxVM uses a balanced path mechanism to  
schedule I/O to multipathed disks. As a result, I/O may go through any available  
path at any given point in time. For example, if a system has an Active/Active  
storage array, and you need to change an interface board that is connected to  
this disk array (if supported by the hardware), you can use the vxdmpadm  
command to list the controllers that are connected to the interface board.  
Disable the controllers to stop further I/O to the disks that are accessed through  
the interface board. You can then replace the board without causing disruption  
to any ongoing I/O to disks in the disk array.  
In Active/Passive type disk arrays, VxVM schedules I/O to use the primary path  
until a failure is encountered. To change the interface card for an array port or  
an HBA controller card on the host (if supported by the hardware) that is  
connected to the disk array, disable I/O operations to the array port or to the  
HBA controller. This shifts all I/O over to an active secondary path or to an  
active primary path on another controller so that you can change the hardware.  
After the operation is over, you can use vxdmpadmto re-enable the paths through  
the controllers.  
See “Disabling I/O for paths, controllers or array ports” on page 153.  
See “Enabling I/O for paths, controllers or array ports” on page 154.  
See “Upgrading disk controller firmware” on page 154.  
Note: From release 5.0 of VxVM, these operations are supported for controllers  
that are used to access disk arrays on which cluster-shareable disk groups are  
configured.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Administering dynamic multipathing (DMP) 137  
Displaying DMP database information  
Displaying DMP database information  
You can use the vxdmpadmcommand to list DMP database information and  
perform other administrative tasks. This command allows you to list all  
controllers that are connected to disks, and other related information that is  
stored in the DMP database. You can use this information to locate system  
hardware, and to help you decide which controllers need to be enabled or  
disabled.  
The vxdmpadmcommand also provides useful information such as disk array  
serial numbers, which DMP devices (disks) are connected to the disk array, and  
which paths are connected to a particular controller, enclosure or array port.  
For more information, see “Administering DMP using vxdmpadm” on page 139.  
Displaying the paths to a disk  
The vxdiskcommand is used to display the multipathing information for a  
particular metadevice. The metadevice is a device representation of a particular  
physical disk having multiple physical paths from one of the system’s HBA  
controllers. In VxVM, all the physical disks in the system are represented as  
metadevices with one or more physical paths.  
You can use the vxdiskpath command to display the relationships between the  
device paths, disk access names, disk media names and disk groups on a system  
as shown here:  
# vxdisk path  
SUBPATH  
c1t0d0  
c4t0d0  
c1t1d0  
c4t1d0  
.
DANAME  
c1t0d0  
c1t0d0  
c1t1d0  
c1t1d0  
DMNAME  
mydg01  
mydg01  
mydg02  
mydg02  
GROUP  
mydg  
mydg  
mydg  
mydg  
STATE  
ENABLED  
ENABLED  
ENABLED  
ENABLED  
.
.
This shows that two paths exist to each of the two disks, mydg01and mydg02,  
and also indicates that each disk is in the ENABLEDstate.  
To view multipathing information for a particular metadevice, use the following  
command:  
# vxdisk list devicename  
For example, to view multipathing information for c1t0d3, use the following  
command:  
# vxdisk list c1t0d3  
Typical output from the vxdisklistcommand is as follows:  
Device:  
c1t0d3  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
138 Administering dynamic multipathing (DMP)  
Displaying the paths to a disk  
devicetag: c1t0d3  
type:  
simple  
hostid:  
disk:  
timeout:  
group:  
info:  
zort  
name=mydg04 id=962923652.362193.zort  
30  
name=mydg id=962212937.1025.zort  
privoffset=128  
flags:  
online ready private autoconfig autoimport imported  
pubpaths: block=/dev/vx/dmp/c1t0d3  
privpaths: char=/dev/vx/rdmp/c1t0d3  
version:  
iosize:  
public:  
private:  
update:  
headers:  
configs:  
logs:  
2.1  
min=1024 (bytes) max=64 (blocks)  
slice=0 offset=1152 len=4101723  
slice=0 offset=128 len=1024  
time=962923719 seqno=0.7  
0 248  
count=1 len=727  
count=1 len=110  
Defined regions:  
config  
disabled  
config  
disabled  
log  
priv 000017-000247[000231]:copy=01 offset=000000  
priv 000249-000744[000496]:copy=01 offset=000231  
priv 000745-000854[000110]:copy=01 offset=000000  
disabled  
lockrgn priv 000855-000919[000065]: part=00 offset=000000  
Multipathing information:  
numpaths:  
2
c1t0d3 state=enabled type=secondary  
c4t1d3 state=disabled type=primary  
In the Multipathing informationsection of this output, the numpathsline  
shows that there are 2 paths to the device, and the following two lines show that  
the path to c1t0d3is active (state=enabled) and that the other path c4t1d3  
has failed (state=disabled).  
The typefield is shown for disks on Active/Passive type disk arrays such as the  
EMC CLARiiON, Hitachi HDS 9200 and 9500, Sun StorEdge 6xxx, and Sun  
StorEdge T3 array. This field indicates the primary and secondary paths to the  
disk.  
The typefield is not displayed for disks on Active/Active type disk arrays such  
as the EMC Symmetrix, Hitachi HDS 99xx and Sun StorEdge 99xx Series, and  
IBM ESS Series. Such arrays have no concept of primary and secondary paths.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Administering dynamic multipathing (DMP) 139  
Administering DMP using vxdmpadm  
Administering DMP using vxdmpadm  
The vxdmpadmutility is a command line administrative interface to the DMP  
feature of VxVM. You can use the vxdmpadmutility to perform the following  
tasks.  
■
■
■
■
■
■
Retrieve the name of the DMP device corresponding to a particular path.  
Display the members of a LUN group.  
List all paths under a DMP device node, HBA controller or array port.  
Display information about the HBA controllers on the host.  
Display information about enclosures.  
Display information about array ports that are connected to the storage  
processors of enclosures.  
■
Display information about devices that are controlled by third-party  
multipathing drivers.  
■
■
■
■
■
■
■
■
■
Gather I/O statistics for a DMP node, enclosure, path or controller.  
Configure the attributes of the paths to an enclosure.  
Set the I/O policy that is used for the paths to an enclosure.  
Enable or disable I/O for a path, HBA controller or array port on the system.  
Upgrade disk controller firmware.  
Rename an enclosure.  
Configure how DMP responds to I/O request failures.  
Configure the I/O throttling mechanism.  
Control the operation of the DMP path restoration thread.  
The following sections cover these tasks in detail along with sample output.  
The vxdmpadmcommand can also be used to change the value of various DMP  
tunables. See “Changing the values of tunables” on page 474.  
For more information about the vxdmpadmcommand, see the vxdmpadm(1M)  
manual page.  
Retrieving information about a DMP node  
The following command displays the DMP node that controls a particular  
physical path:  
# vxdmpadm getdmpnode nodename=c3t2d1  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
140 Administering dynamic multipathing (DMP)  
Administering DMP using vxdmpadm  
The physical path is specified by argument to the nodenameattribute, which  
must be a valid path listed in the /dev/rdskdirectory.  
The above command displays output such as the following:  
NAME  
===============================================================  
c3t2d1 ENABLED ACME enc0  
STATE  
ENCLR-TYPE  
PATHS  
ENBL  
DSBL  
ENCLR-NAME  
2
2
0
Use the enclosureattribute with getdmpnodeto obtain a list of all DMP nodes  
for the specified enclosure.  
# vxdmpadm getdmpnode enclosure=enc0  
NAME  
STATE  
ENCLR-TYPE  
PATHS ENBL  
DSBL  
ENCLR-NAME  
===============================================================  
c2t1d0 ENABLED ACME  
c2t1d1 ENABLED ACME  
c2t1d2 ENABLED ACME  
c2t1d3 ENABLED ACME  
2
2
2
2
2
2
2
2
0
0
0
0
enc0  
enc0  
enc0  
enc0  
Displaying the members of a LUN group  
The following command displays the DMP nodes that are in the same LUN group  
as a specified DMP node:  
# vxdmpadm getlungroup dmpnodename=c11t0d10  
The above command displays output such as the following:  
NAME  
STATE  
ENCLR-TYPE PATHS  
ENBL  
DSBL ENCLR-NAME  
===============================================================  
c11t0d8 ENABLED ACME  
c11t0d9 ENABLED ACME  
c11t0d10 ENABLED ACME  
c11t0d11 ENABLED ACME  
2
2
2
2
2
2
2
2
0
0
0
0
enc1  
enc1  
enc1  
enc1  
Displaying paths controlled by a DMP node, controller or array port  
The vxdmpadm getsubpathscommand combined with the dmpnodename  
attribute displays all the paths to a LUN that are controlled by the specified DMP  
node name from the /dev/vx/rdmpdirectory:  
# vxdmpadm getsubpaths dmpnodename=c2t66d0  
NAME  
STATE[A]  
PATH-TYPE[M]CTLR-NAMEENCLR-TYPE ENCLR-NAME ATTRS  
===================================================================  
c2t66d0ENABLED(A) PRIMARY  
c1t66d0ENABLED PRIMARY  
c2  
c1  
ACME  
ACME  
enc0  
enc0  
-
-
For A/A arrays, all enabled paths that are available for I/O are shown as  
ENABLED(A).  
Download from Www.Somanuals.com. All Manuals Search And Download.  
               
Administering dynamic multipathing (DMP) 141  
Administering DMP using vxdmpadm  
For A/P arrays in which the I/O policy is set to singleactive, only one path is  
shown as ENABLED(A). The other paths are enabled but not available for I/O. If  
the I/O policy is not set to singleactive, DMP can use a group of paths (all  
primary or all secondary) for I/O, which are shown as ENABLED(A). See  
“Specifying the I/O policy” on page 147 for more information.  
Paths that are in the DISABLED state are not available for I/O operations.  
You can use getsubpathsto obtain information about all the paths that are  
connected to a particular HBA controller:  
# vxdmpadm getsubpaths ctlr=c2  
NAME  
STATE[-] PATH-TYPE[-] CTLR-NAME ENCLR-TYPE ENCLR-NAMEATTRS  
===================================================================  
c2t1d0 ENABLED PRIMARY  
c2t2d0 ENABLED PRIMARY  
c2t3d0 ENABLED SECONDARY  
c2t4d0 ENABLED SECONDARY  
c2t1d0  
c2t2d0  
c2t3d0  
c2t4d0  
ACME  
ACME  
ACME  
ACME  
enc0  
enc0  
enc0  
enc0  
-
-
-
-
You can also use getsubpathsto obtain information about all the paths that are  
connected to a port on an array. The array port can be specified by the name of  
the enclosure and the array port ID, or by the worldwide name (WWN) identifier  
of the array port:  
# vxdmpadm getsubpaths enclosure=HDS9500V0portid=1A  
# vxdmpadm getsubpaths pwwn=20:00:00:E0:8B:06:5F:19  
Displaying information about controllers  
The following command lists attributes of all HBA controllers on the system:  
# vxdmpadm listctlr all  
CTLR-NAME  
ENCLR-TYPE  
STATE  
ENCLR-NAME  
===============================================================  
c1  
c2  
c3  
c4  
OTHER  
X1  
ACME  
ACME  
ENABLED  
ENABLED  
ENABLED  
ENABLED  
other0  
jbod0  
enc0  
enc0  
This output shows that the controller c1 is connected to disks that are not in  
any recognized DMP category as the enclosure type is OTHER.  
The other controllers are connected to disks that are in recognized DMP  
categories.  
All the controllers are in the ENABLED state which indicates that they are  
available for I/O operations.  
The state DISABLED is used to indicate that controllers are unavailable for I/O  
operations. The unavailability can be due to a hardware failure or due to I/O  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
142 Administering dynamic multipathing (DMP)  
Administering DMP using vxdmpadm  
operations being disabled on that controller by using the vxdmpadm disable  
command.  
This form of the command lists controllers belonging to a specified enclosure  
and enclosure type:  
# vxdmpadm listctlr enclosure=enc0 type=ACME  
CTLR-NAME  
ENCLR-TYPE  
STATE  
ENCLR-NAME  
===============================================================  
c2  
c3  
ACME  
ACME  
ENABLED  
ENABLED  
enc0  
enc0  
Displaying information about enclosures  
To display the attributes of a specified enclosure, including its enclosure type,  
enclosure serial number, status and array type, use the following command:  
# vxdmpadm listenclosure enc0  
ENCLR_NAME ENCLR_TYPE ENCLR_SNO  
===============================================================  
enc0 A3 60020f20000001a90000 CONNECTED A/P  
STATUS ARRAY_TYPE  
The following command lists attributes for all enclosures in a system:  
# vxdmpadm listenclosure all  
The following is example output from this command:  
ENCLR_NAME ENCLR_TYPE  
ENCLR_SNO  
STATUS ARRAY_TYPE  
===============================================================  
Disk  
ANA0  
enc0  
Disk  
ACME  
A3  
DISKS  
508002000001d660  
60020f20000001a90000 CONNECTED A/P  
CONNECTED Disk  
CONNECTED A/A  
Displaying information about array ports  
To display the attributes of an array port that is accessible via a path, DMP node  
or HBA controller, use one of the following commands:  
# vxdmpadm getportids path=path-name  
# vxdmpadm getportids dmpnodename=dmpnode-name  
# vxdmpadm getportids ctlr=ctlr-name  
The information displayed for an array port includes the name of its enclosure,  
and its ID and worldwide name (WWN) identifier.  
The following form of the command displays information about all of the array  
ports within the specified enclosure:  
# vxdmpadm getportids enclosure=enclr-name  
The following example shows information about the array port that is accessible  
via DMP node c2t66d0:  
# vxdmpadm getportids dmpnodename=c2t66d0  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Administering dynamic multipathing (DMP) 143  
Administering DMP using vxdmpadm  
NAME  
==============================================================  
c2t66d0 HDS9500V0 1A 20:00:00:E0:8B:06:5F:19  
ENCLR-NAME ARRAY-PORT-ID pWWN  
Displaying information about TPD-controlled devices  
The third-party driver (TPD) coexistence feature allows I/O that is controlled by  
third-party multipathing drivers to bypass DMP while retaining the monitoring  
capabilities of DMP. The following commands allow you to display the paths  
that DMP has discovered for a given TPD device, and the TPD device that  
corresponds to a given TPD-controlled node discovered by DMP:  
# vxdmpadm getsubpaths tpdnodename=TPD_node_name  
# vxdmpadm gettpdnode nodename=DMP_node_name  
See “Changing device naming for TPD-controlled enclosures” on page 94 for  
information on how to select whether OS or TPD-based device names are  
displayed.  
For example, consider the following disks in an EMC Symmetrix array controlled  
by PowerPath, which are known to DMP:  
# vxdisk list  
DEVICE  
TYPE  
DISK  
GROUP  
ppdg  
ppdg  
ppdg  
ppdg  
ppdg  
ppdg  
ppdg  
ppdg  
ppdg  
ppdg  
STATUS  
online  
online  
online  
online  
online  
online  
online  
online  
online  
online  
emcpower10  
emcpower11  
emcpower12  
emcpower13  
emcpower14  
emcpower15  
emcpower16  
emcpower17  
emcpower18  
emcpower19  
auto:sliced  
auto:sliced  
auto:sliced  
auto:sliced  
auto:sliced  
auto:sliced  
auto:sliced  
auto:sliced  
auto:sliced  
auto:sliced  
disk1  
disk2  
disk3  
disk4  
disk5  
disk6  
disk7  
disk8  
disk9  
disk10  
The following command displays the paths that DMP has discovered, and which  
correspond to the PowerPath-controlled node, emcpower10:  
# vxdmpadm getsubpaths tpdnodename=emcpower10  
NAME  
TPDNODENAME PATH-TYPE[-]DMP-NODENAME ENCLR-TYPE ENCLR-NAME  
===================================================================  
c7t0d10emcpower10s2 -  
c6t0d10emcpower10s2 -  
emcpower10  
emcpower10  
EMC  
EMC  
EMC0  
EMC0  
Conversely, the next command displays information about the PowerPath node  
that corresponds to the path, c7t0d10, discovered by DMP:  
# vxdmpadm gettpdnode nodename=c7t0d10  
NAME  
===================================================================  
emcpower10s2 ENABLED 2 EMC EMC0  
STATE  
PATHS ENCLR-TYPE  
ENCLR-NAME  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
144 Administering dynamic multipathing (DMP)  
Administering DMP using vxdmpadm  
Gathering and displaying I/O statistics  
You can use the vxdmpadmiostatcommand to gather and display I/O statistics  
for a specified DMP node, enclosure, path or controller.  
To enable the gathering of statistics, enter this command:  
# vxdmpadm iostat start [memory=size]  
To reset the I/O counters to zero, use this command:  
# vxdmpadm iostat reset  
The memoryattribute can be used to limit the maximum amount of memory that  
is used to record I/O statistics for each CPU. The default limit is 32k(32  
kilobytes) per CPU.  
To display the accumulated statistics at regular intervals, use the following  
command:  
# vxdmpadm iostat show {all | dmpnodename=dmp-node | \  
enclosure=enclr-name | pathname=path-name | ctlr=ctlr-name} \  
[interval=seconds [count=N]]  
This command displays I/O statistics for all controllers (all), or for a specified  
DMP node, enclosure, path or controller. The statistics displayed are the CPU  
usage and amount of memory per CPU used to accumulate statistics, the number  
of read and write operations, the number of kilobytes read and written, and the  
average time in milliseconds per kilobyte that is read or written.  
The intervaland countattributes may be used to specify the interval in  
seconds between displaying the I/O statistics, and the number of lines to be  
displayed. The actual interval may be smaller than the value specified if  
insufficient memory is available to record the statistics.  
To disable the gathering of statistics, enter this command:  
# vxdmpadm iostat stop  
Examples of using the vxdmpadm iostat command  
The follow is an example session using the vxdmpadmiostatcommand. The first  
command enables the gathering of I/O statistics:  
# vxdmpadm iostat start  
The next command displays the current statistics including the accumulated  
total numbers of read and write operations and kilobytes read and written, on all  
paths:  
# vxdmpadm iostat show all  
cpu usage = 7952us  
OPERATIONS KBYTES  
READS  
per cpu memory = 8192b  
AVG TIME(ms)  
PATHNAME  
c0t0d0  
c2t118d0  
c3t118d0  
c2t122d0  
c3t122d0  
READS  
WRITES  
WRITES  
READS  
WRITES  
1088  
87  
0
0
0
0
0
0
557056  
44544  
0
0 0.009542 0.000000  
0 0.001194 0.000000  
0 0.000000 0.000000  
0 0.007265 0.000000  
0 0.000000 0.000000  
87  
0
44544  
0
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Administering dynamic multipathing (DMP) 145  
Administering DMP using vxdmpadm  
c2t115d0  
c3t115d0  
c2t103d0  
c3t103d0  
c2t102d0  
c3t102d0  
c2t121d0  
c3t121d0  
c2t112d0  
c3t112d0  
c2t96d0  
87  
0
87  
0
87  
0
87  
0
87  
0
87  
0
87  
0
87  
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
44544  
0 0.001200 0.000000  
0 0.000000 0.000000  
0 0.007315 0.000000  
0 0.000000 0.000000  
0 0.001132 0.000000  
0 0.000000 0.000000  
0 0.000997 0.000000  
0 0.000000 0.000000  
0 0.001559 0.000000  
0 0.000000 0.000000  
0 0.007057 0.000000  
0 0.000000 0.000000  
0 0.007247 0.000000  
0 0.000000 0.000000  
0 0.007235 0.000000  
0 0.000000 0.000000  
0 0.001390 0.000000  
0 0.000000 0.000000  
0
44544  
0
44544  
0
44544  
0
44544  
0
44544  
0
44544  
0
44544  
0
44544  
0
c3t96d0  
c2t106d0  
c3t106d0  
c2t113d0  
c3t113d0  
c2t119d0  
c3t119d0  
87  
0
The following command changes the amount of memory that vxdmpadm can  
use to accumulate the statistics:  
# vxdmpadm iostat start memory=4096  
The displayed statistics can be filtered by path name, DMP node name, and  
enclosure name (note that the per-CPU memory has changed following the  
previous command):  
# vxdmpadm iostat show pathname=c3t115d0  
cpu usage = 8132us  
OPERATIONS BYTES  
WRITES  
per cpu memory = 4096b  
AVG TIME(ms)  
PATHNAME  
c3t115d0  
READS  
0
READS  
0
WRITES  
READS  
WRITES  
0
0 0.000000 0.000000  
# vxdmpadm iostat show dmpnodename=c0t0d0  
cpu usage = 8501us  
per cpu memory = 4096b  
OPERATIONS  
BYTES AVG TIME(ms)  
PATHNAME  
c0t0d0  
READS  
1088  
WRITES  
0
READS  
557056  
WRITES  
READS  
WRITES  
0 0.009542 0.000000  
# vxdmpadm iostat show enclosure=Disk  
cpu usage = 8626us  
per cpu memory = 4096b  
OPERATIONS  
BYTES AVG TIME(ms)  
PATHNAME  
c0t0d0  
READS  
1088  
WRITES  
0
READS  
557056  
WRITES  
READS  
WRITES  
0 0.009542 0.000000  
You can also specify the number of times to display the statistics and the time  
interval. Here the incremental statistics for a path are displayed twice with a  
2-second interval:  
# vxdmpadm iostat show pathname=c3t115d0 interval=2 count=2  
cpu usage = 8195us  
OPERATIONS BYTES  
READS WRITES READS  
per cpu memory = 4096b  
AVG TIME(ms)  
PATHNAME  
WRITES  
READS  
WRITES  
Download from Www.Somanuals.com. All Manuals Search And Download.  
146 Administering dynamic multipathing (DMP)  
Administering DMP using vxdmpadm  
c3t115d0  
0
0
0
0 0.000000 0.000000  
cpu usage = 59us  
OPERATIONS BYTES  
READS WRITES READS  
per cpu memory = 4096b  
AVG TIME(ms)  
PATHNAME  
c3t115d0  
WRITES  
READS  
WRITES  
0
0
0
0 0.000000 0.000000  
Setting the attributes of the paths to an enclosure  
You can use the vxdmpadmsetattrcommand to set the following attributes of  
the paths to an enclosure or disk array:  
■
■
■
active  
Changes a standby (failover) path to an active path. The example below  
specifies an active path for an A/P-C disk array:  
# vxdmpadm setattr path c2t10d0 pathtype=active  
nomanual  
Restores the original primary or secondary attributes of a path. This  
example restores the attributes for a path to an A/P disk array:  
# vxdmpadm setattr path c3t10d0 pathtype=nomanual  
nopreferred  
Restores the normal priority of a path. The following example restores the  
default priority to a path:  
# vxdmpadm setattr path c1t20d0 pathtype=nopreferred  
■
preferred [priority=N]  
Specifies a path as preferred, and optionally assigns a priority number to it.  
If specified, the priority number must be an integer that is greater than or  
equal to one. Higher priority numbers indicate that a path is able to carry a  
greater I/O load.  
Note: Setting a priority for path does not change the I/O policy. The I/O  
policy must be set independently as described in “Specifying the I/O policy”  
on page 147.  
This example first sets the I/O policy to priorityfor an Active/Active disk  
array, and then specifies a preferred path with an assigned priority of 2:  
# vxdmpadm setattr enclosure enc0 iopolicy=priority  
# vxdmpadm setattr path c1t20d0 pathtype=preferred \  
priority=2  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
Administering dynamic multipathing (DMP) 147  
Administering DMP using vxdmpadm  
■
■
■
primary  
Defines a path as being the primary path for an Active/Passive disk array.  
The following example specifies a primary path for an A/P disk array:  
# vxdmpadm setattr path c3t10d0 pathtype=primary  
secondary  
Defines a path as being the secondary path for an Active/Passive disk array.  
This example specifies a secondary path for an A/P disk array:  
# vxdmpadm setattr path c4t10d0 pathtype=secondary  
standby  
Marks a standby (failover) path that it is not used for normal I/O  
scheduling. This path is used if there are no active paths available for I/O.  
The next example specifies a standby path for an A/P-C disk array:  
# vxdmpadm setattr path c2t10d0 pathtype=standby  
Displaying the I/O policy  
To display the current and default settings of the I/O policy for an enclosure,  
array or array type, use the vxdmpadmgetattrcommand.  
The following example displays the default and current setting of iopolicyfor  
JBOD disks:  
# vxdmpadm getattr enclosure Disk iopolicy  
ENCLR_NAME  
---------------------------------------  
Disk MinimumQ Balanced  
DEFAULT  
CURRENT  
The next example displays the setting of partitionsizefor the enclosure  
enc0, on which the balancedI/O policy with a partition size of 2MB has been  
set:  
# vxdmpadm getattr enclosure enc0 partitionsize  
ENCLR_NAME  
DEFAULT  
CURRENT  
---------------------------------------  
enc0  
1024  
2048  
Specifying the I/O policy  
You can use the vxdmpadmsetattrcommand to change the I/O policy for  
distributing I/O load across multiple paths to a disk array or enclosure. You can  
set policies for an enclosure (for example, HDS01), for all enclosures of a  
particular type (such as HDS), or for all enclosures of a particular array type  
(such as A/Afor Active/Active, or A/Pfor Active/Passive).  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                   
148 Administering dynamic multipathing (DMP)  
Administering DMP using vxdmpadm  
Note: Starting with release 4.1 of VxVM, I/O policies are recorded in the file  
/etc/vx/dmppolicy.info, and are persistent across reboots of the system.  
Do not edit this file yourself.  
The following policies may be set:  
■
adaptive  
This policy attempts to maximize overall I/O throughput from/to the disks  
by dynamically scheduling I/O on the paths. It is suggested for use where  
I/O loads can vary over time. For example, I/O from/to a database may  
exhibit both long transfers (table scans) and short transfers (random look  
ups). The policy is also useful for a SAN environment where different paths  
may have different number of hops. No further configuration is possible as  
this policy is automatically managed by DMP.  
In this example, the adaptiveI/O policy is set for the enclosure enc1:  
# vxdmpadm setattr enclosure enc1 iopolicy=adaptive  
■
■
adaptiveminq  
Similar to the adaptivepolicy, except that I/O is scheduled according to the  
length of the I/O queue on each path. The path with the shortest queue is  
assigned the highest priority.  
balanced[partitionsize=size]  
This policy is designed to optimize the use of caching in disk drives and  
RAID controllers. The size of the cache typically ranges from 120KB to  
500KB or more, depending on the characteristics of the particular  
hardware. During normal operation, the disks (or LUNs) are logically  
divided into a number of regions (or partitions), and I/O from/to a given  
region is sent on only one of the active paths. Should that path fail, the  
workload is automatically redistributed across the remaining paths.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Administering dynamic multipathing (DMP) 149  
Administering DMP using vxdmpadm  
You can use the size argument to the partitionsizeattribute to specify the  
partition size. The partition size in blocks is adjustable in powers of 2 from  
2 up to 2^31 as illustrated in the table below:  
Partition size in blocks  
Equivalent size in bytes  
2,048  
2
4
4,096  
8
8,192  
16  
16,384  
32  
32,768  
64  
65,536  
128  
131,072  
256  
262,144  
512  
524,288  
1024 (default)  
2048  
1,048,576  
2,097,152  
4,194,304  
4096  
The default value for the partition size is 1024 blocks (1MB). A value that is  
not a power of 2 is silently rounded down to the nearest acceptable value.  
Specifying a partition size of 0 is equivalent to the default partition size of  
1024 blocks (1MB). For example, the suggested partition size for an Hitachi  
HDS 9960 A/A array is from 16,384 to 65,536 blocks (16MB to 64MB) for an  
I/O activity pattern that consists mostly of sequential reads or writes.  
Note: The benefit of this policy is lost if the value is set larger than the cache  
size.  
The default value can be changed by adjusting the value of a tunable  
parameter (see “dmp_pathswitch_blks_shift” on page 476).  
The next example sets the balancedI/O policy with a partition size of 2048  
blocks (2MB) on the enclosure enc0:  
# vxdmpadm setattr enclosure enc0 iopolicy=balanced \  
partitionsize=2048  
Download from Www.Somanuals.com. All Manuals Search And Download.  
150 Administering dynamic multipathing (DMP)  
Administering DMP using vxdmpadm  
■
minimumq  
This policy sends I/O on paths that have the minimum number of  
outstanding I/O requests in the queue for a LUN. This is suitable for  
low-end disks or JBODs where a significant track cache does not exist. No  
further configuration is possible as DMP automatically determines the path  
with the shortest queue.  
The following example sets the I/O policy to minimumqfor a JBOD:  
# vxdmpadm setattr enclosure Disk iopolicy=minimumq  
This is the default I/O policy for Active/Active (A/A) arrays.  
■
priority  
This policy is useful when the paths in a SAN have unequal performance,  
and you want to enforce load balancing manually. You can assign priorities  
to each path based on your knowledge of the configuration and  
performance characteristics of the available paths, and of other aspects of  
your system. See “Setting the attributes of the paths to an enclosure” on  
page 146 for details of how to assign priority values to individual paths.  
In this example, the I/O policy is set to priorityfor all SENA arrays:  
# vxdmpadm setattr arrayname SENA iopolicy=priority  
■
round-robin  
This policy shares I/O equally between the paths in a round-robin sequence.  
For example, if there are three paths, the first I/O request would use one  
path, the second would use a different path, the third would be sent down  
the remaining path, the fourth would go down the first path, and so on. No  
further configuration is possible as this policy is automatically managed by  
DMP.  
The next example sets the I/O policy to round-robinfor all Active/Active  
arrays:  
# vxdmpadm setattr arraytype A/A iopolicy=round-robin  
This is the default I/O policy for Active/Passive (A/P) and Asymmetric  
Active/Active (A/A-A) arrays.  
■
singleactive  
This policy routes I/O down the single active path. This policy can be  
configured for A/P arrays with one active path per controller, where the  
other paths are used in case of failover. If configured for A/A arrays, there  
is no load balancing across the paths, and the alternate paths are only used  
to provide high availability (HA). If the currently active path fails, I/O is  
switched to an alternate active path. No further configuration is possible as  
the single active path is selected by DMP.  
The following example sets the I/O policy to singleactivefor JBOD disks:  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Administering dynamic multipathing (DMP) 151  
Administering DMP using vxdmpadm  
# vxdmpadm setattr arrayname DISK iopolicy=singleactive  
Scheduling I/O on the paths of an Asymmetric Active/Active  
array  
You can specify the use_all_pathsattribute in conjunction with the adaptive,  
balanced, minimumq, priorityand round-robinI/O policies to specify whether  
I/O requests are to be scheduled on the secondary paths in addition to the  
primary paths of an Asymmetric Active/Active (A/A-A) array. Depending on the  
characteristics of the array, the consequent improved load balancing can  
increase the total I/O throughput. However, this feature should only be enabled  
if recommended by the array vendor. It has no effect for array types other than  
A/A-A.  
For example, the following command sets the balancedI/O policy with a  
partition size of 2048 blocks (2MB) on the enclosure enc0, and allows  
scheduling of I/O requests on the secondary paths:  
# vxdmpadm setattr enclosure enc0 iopolicy=balanced \  
partitionsize=2048 use_all_paths=no  
The default setting for this attribute is use_all_paths=no.  
Example of applying load balancing in a SAN  
This example describes how to configure load balancing in a SAN environment  
where there are multiple primary paths to an Active/Passive device through  
several SAN switches. As can be seen in this sample output from the vxdisk  
listcommand, the device c3t2d15has eight primary paths:  
# vxdisk list c3t2d15  
Device: c3t2d15  
...  
numpaths: 8  
c2t0d15 state=enabled type=primary  
c2t1d15 state=enabled type=primary  
c3t1d15 state=enabled type=primary  
c3t2d15 state=enabled type=primary  
c4t2d15 state=enabled type=primary  
c4t3d15 state=enabled type=primary  
c5t3d15 state=enabled type=primary  
c5t4d15 state=enabled type=primary  
In addition, the device is in the enclosure ENC0, belongs to the disk group mydg,  
and contains a simple concatenated volume myvol1.  
The first step is to enable the gathering of DMP statistics:  
# vxdmpadm iostat start  
Next the ddcommand is used to apply an input workload from the volume:  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
152 Administering dynamic multipathing (DMP)  
Administering DMP using vxdmpadm  
# dd if=/dev/vx/rdsk/mydg/myvol1 of=/dev/null &  
By running the vxdmpadmiostatcommand to display the DMP statistics for the  
device, it can be seen that all I/O is being directed to one path, c5t4d15:  
# vxdmpadm iostat show dmpnodename=c3t2d15 interval=5 count=2  
...  
cpu usage = 11294us per cpu memory = 32768b  
OPERATIONS  
READS WRITES  
KBYTES  
WRITES  
AVG TIME(ms)  
PATHNAME  
c2t0d15  
c2t1d15  
c3t1d15  
c3t2d15  
c4t2d15  
c4t3d15  
c5t3d15  
c5t4d15  
READS  
READS  
WRITES  
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.000000  
0.000000  
0.000000  
0.000000  
0.000000  
0.000000  
0.000000  
0.411069  
0.000000  
0.000000  
0.000000  
0.000000  
0.000000  
0.000000  
0.000000  
0.000000  
5493  
5493  
The vxdmpadmcommand is used to display the I/O policy for the enclosure that  
contains the device:  
# vxdmpadm getattr enclosure ENC0 iopolicy  
ENCLR_NAME  
============================================  
ENC0 Round-Robin Single-Active  
DEFAULT  
CURRENT  
This shows that the policy for the enclosure is set to singleactive, which  
explains why all the I/O is taking place on one path.  
To balance the I/O load across the multiple primary paths, the policy is set to  
round-robinas shown here:  
# vxdmpadm setattr enclosure ENC0 iopolicy=round-robin  
# vxdmpadm getattr enclosure ENC0 iopolicy  
ENCLR_NAME  
============================================  
ENC0 Round-Robin Round-Robin  
DEFAULT  
CURRENT  
The DMP statistics are now reset:  
# vxdmpadm iostat reset  
With the workload still running, the effect of changing the I/O policy to balance  
the load across the primary paths can now be seen.  
# vxdmpadm iostat show dmpnodename=c3t2d15 interval=5 \  
count=2  
...  
cpu usage = 14403us per cpu memory = 32768b  
OPERATIONS  
READS WRITES  
KBYTES  
WRITES  
AVG TIME(ms)  
READS WRITES  
PATHNAME  
c2t0d15  
c2t1d15  
c3t1d15  
c3t2d15  
READS  
1021  
947  
1004  
1027  
1021  
947  
1004  
1027  
0
0
0
0
0
0
0
0
0.396670  
0.391763  
0.393426  
0.402142  
0.000000  
0.000000  
0.000000  
0.000000  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Administering dynamic multipathing (DMP) 153  
Administering DMP using vxdmpadm  
c4t2d15  
c4t3d15  
c5t3d15  
c5t4d15  
1086  
1048  
1036  
1021  
0
0
0
0
1086  
1048  
1036  
1021  
0
0
0
0
0.390424  
0.391221  
0.390927  
0.392752  
0.000000  
0.000000  
0.000000  
0.000000  
The enclosure can be returned to the single active I/O policy by entering the  
following command:  
# vxdmpadm setattr enclosure ENC0 iopolicy=singleactive  
Disabling I/O for paths, controllers or array ports  
Note: From release 5.0 of VxVM, this operation is supported for controllers that  
are used to access disk arrays on which cluster-shareable disk groups are  
configured.  
Disabling I/O through a path, HBA controller or array port prevents DMP from  
issuing I/O requests through the specified path, or the paths that are connected  
to the specified controller or array port. The command blocks until all pending  
I/O requests issued through the paths are completed.  
To disable I/O for a path, use the following command:  
# vxdmpadm [-c|-f] disable path=path_name  
To disable I/O for the paths connected to an HBA controller, use the following  
command:  
# vxdmpadm [-c|-f] disable ctlr=ctlr_name  
To disable I/O for the paths connected to an array port, use one of the following  
commands:  
# vxdmpadm [-c|-f] disable enclosure=enclr_name \  
portid=array_port_ID  
# vxdmpadm [-c|-f] disable pwwn=array_port_WWN  
where the array port is specified either by the enclosure name and the array port  
ID, or by the array port’s worldwide name (WWN) identifier.  
The following are examples of using the command to disable I/O on an array  
port:  
# vxdmpadm disable enclosure=HDS9500V0portid=1A  
# vxdmpadm disable pwwn=20:00:00:E0:8B:06:5F:19  
You can use the -c option to check if there is only a single active path to the disk.  
If so, the disablecommand fails with an error message unless you use the -f  
option to forcibly disable the path.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
154 Administering dynamic multipathing (DMP)  
Administering DMP using vxdmpadm  
The disableoperation fails if it is issued to a controller that is connected to the  
root disk through a single path, and there are no root disk mirrors configured on  
alternate paths. If such mirrors exist, the command succeeds.  
Enabling I/O for paths, controllers or array ports  
Note: This operation is not supported for controllers that are used to access disk  
arrays on which cluster-shareable disk groups are configured.  
Enabling a controller allows a previously disabled path, HBA controller or array  
port to accept I/O again. This operation succeeds only if the path, controller or  
array port is accessible to the host, and I/O can be performed on it. When  
connecting Active/Passive disk arrays, the enableoperation results in failback  
of I/O to the primary path. The enableoperation can also be used to allow I/O to  
the controllers on a system board that was previously detached.  
To enable I/O for a path, use the following command:  
# vxdmpadm enable path=path_name  
To enable I/O for the paths connected to an HBA controller, use the following  
command:  
# vxdmpadm enable ctlr=ctlr_name  
To enable I/O for the paths connected to an array port, use one of the following  
commands:  
# vxdmpadm enable enclosure=enclr_nameportid=array_port_ID  
# vxdmpadm [-f] disable pwwn=array_port_WWN  
where the array port is specified either by the enclosure name and the array port  
ID, or by the array port’s worldwide name (WWN) identifier.  
The following are examples of using the command to enable I/O on an array  
port:  
# vxdmpadm enable enclosure=HDS9500V0portid=1A  
# vxdmpadm enable pwwn=20:00:00:E0:8B:06:5F:19  
Upgrading disk controller firmware  
You can upgrade disk controller firmware without performing a system reboot  
or unloading the VxVM drivers.  
First obtain the appropriate firmware upgrades from your disk drive vendor.  
You can usually download the appropriate files and documentation from the  
vendor’s support website.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Administering dynamic multipathing (DMP) 155  
Administering DMP using vxdmpadm  
For a system with a volume mirrored across 2 controllers on one HBA, set up the  
configuration as follows:  
1
Disable the plex that is associated with the disk device:  
# /opt/VRTS/bin/vxplex -g diskgroup det plex  
2
Stop I/O to all disks through one controller of the HBA:  
# /opt/VRTS/bin/vxdmpadm disable ctlr=first_cntlr  
For the other controller on the HBA, enter:  
# /opt/VRTS/bin/vxdmpadm -f disable ctlr=second_cntlr  
3
4
Upgrade the firmware on those disks for which the controllers have been  
disabled using the procedures that you obtained from the disk drive vendor.  
After doing the upgrade, re-enable all the controllers:  
# /opt/VRTS/bin/vxdmpadm enable ctlr=first_cntlr  
# /opt/VRTS/bin/vxdmpadm enable ctlr=second_cntlr  
5
Re-enable the plex associated with the device:  
# /opt/VRTS/bin/vxplex -g diskgroup att volume plex  
This command takes some time depending upon the size of the mirror set.  
Renaming an enclosure  
The vxdmpadm setattrcommand can be used to assign a meaningful name to an  
existing enclosure, for example:  
# vxdmpadm setattr enclosure enc0 name=GRP1  
This example changes the name of an enclosure from enc0to GRP1.  
Note: The maximum length of the enclosure name prefix is 25 characters. The  
name must not contain an underbar character (_).  
The following command shows the changed name:  
# vxdmpadm listenclosure all  
ENCLR_NAME ENCLR_TYPE  
ENCLR_SNO  
STATUS  
============================================================  
other0  
jbod0  
GRP1  
OTHER  
X1  
ACME  
OTHER_DISKS  
X1_DISKS  
60020f20000001a90000  
CONNECTED  
CONNECTED  
CONNECTED  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
156 Administering dynamic multipathing (DMP)  
Administering DMP using vxdmpadm  
Configuring the response to I/O failures  
By default, DMP is configured to retry a failed I/O request up to 5 times for a  
single path. To display the current settings for handling I/O request failures that  
are applied to the paths to an enclosure, array name or array type, use the  
vxdmpadmgetattrcommand:  
# vxdmpadm getattr \  
{enclosure enc-name|arrayname name|arraytype type} \  
recoveryoption  
See “Displaying recoveryoption values” on page 159 for more information.  
The following example displays the I/O request failure setting for the paths to  
the enclosure enc0:  
# vxdmpadm getattr enclosure enc0 recoveryoption  
The vxdmpadm setattr command can be used to configure how DMP responds  
to failed I/O requests on the paths to a specified enclosure, disk array name, or  
type of array.  
The following form of the command sets a limit for the number of times that  
DMP will attempt to retry sending an I/O request on a path:  
# vxdmpadm setattr \  
{enclosure enc-name|arrayname name|arraytype type} \  
recoveryoption=fixedretry retrycount==n  
The value of the argument to retrycountspecifies the number of retries to be  
attempted before DMP reschedules the I/O request on another available path, or  
fails the request altogether.  
As an alternative to specifying a fixed number of retries, the following version of  
the command specifies how long DMP should allow an I/O request to be retried  
on a path:  
# vxdmpadm setattr \  
{enclosure enc-name|arrayname name|arraytype type} \  
recoveryoption=timebound iotimeout==seconds  
The value of the argument to iotimeoutspecifies the time in seconds that DMP  
waits for an outstanding I/O request to succeed before it reschedules the request  
on another available path, or fails the I/O request altogether. The effective  
number of retries is the value of iotimeoutdivided by the sum of the times  
taken for each retry attempt. DMP abandons retrying to send the I/O request  
before the specified time limit has expired if it predicts that the next retry will  
take the total elapsed time over this limit.  
The default value of iotimeoutis 10 seconds. For some applications, such as  
Oracle, it may be desirable to set iotimeoutto a larger value, such as 60 seconds.  
Note: The fixedretryand timeboundsettings are mutually exclusive.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Administering dynamic multipathing (DMP) 157  
Administering DMP using vxdmpadm  
The following example configures time-bound recovery for the enclosure enc0,  
and sets the value of iotimeoutto 60 seconds:  
# vxdmpadm setattr enclosure enc0 recoveryoption=timebound \  
iotimeout=60  
The next example sets a fixed-retry limit of 10 for the paths to all Active/Active  
arrays:  
# vxdmpadm setattr arraytype A/A recoveryoption=fixedretry \  
retrycount=10  
Specifying recoveryoption=defaultresets DMP to the default settings  
corresponding to recoveryoption=fixedretryretrycount=30, for example:  
# vxdmpadm setattr arraytype A/A recoveryoption=default  
This command also has the effect of configuring I/O throttling with an I/O  
timeout value of 10 seconds on the paths.  
See “Configuring the I/O throttling mechanism” on page 157 for details.  
Note: The response to I/O failure settings is persistent across reboots of the  
system.  
Configuring the I/O throttling mechanism  
By default, II/O throttling is turned on for all paths with DMP configured to wait  
for a maximum of 10 seconds for outstanding I/O requests to succeed. To display  
the current settings for I/O throttling that are applied to the paths to an  
enclosure, array name or array type, use the vxdmpadmgetattrcommand:  
# vxdmpadm getattr \  
{enclosure enc-name|arrayname name|arraytype type}\  
recoveryoption  
See “Displaying recoveryoption values” on page 159 for more information.  
The following example displays the I/O throttling setting for the paths to the  
enclosure enc0:  
# vxdmpadm getattr enclosure enc0 recoveryoption  
If enabled, I/O throttling imposes a small overhead on CPU and memory usage  
because of the activity of the statistics-gathering daemon. If I/O throttling is  
disabled, the daemon no longer collects statistics, and remains inactive until I/O  
throttling is re-enabled.  
To turn off I/O throttling, use the following form of the vxdmpadmsetattr  
command:  
# vxdmpadm setattr \  
{enclosure enc-name|arrayname name|arraytype type}\  
recoveryoption=nothrottle  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
158 Administering dynamic multipathing (DMP)  
Administering DMP using vxdmpadm  
The following example shows how to disable I/O throttling for the paths to the  
enclosure enc0:  
# vxdmpadm setattr enclosure enc0 recoveryoption=nothrottle  
The vxdmpadm setattr command can be used to enable I/O throttling on the  
paths to a specified enclosure, disk array name, or type of array:  
# vxdmpadm setattr \  
{enclosure enc-name|arrayname name|arraytype type}\  
recoveryoption=throttle {iotimeout=seconds|queuedepth=n}  
If the iotimeoutattribute is specified, its argument specifies the time in  
seconds that DMP waits for an outstanding I/O request to succeed before  
invoking I/O throttling on the path. The default value of iotimeoutis 10  
seconds. Setting iotimeoutto a larger value potentially causes more I/O  
requests to become queued up in the SCSI driver before I/O throttling is invoked.  
If the queuedepthattribute is specified, its argument specifies the number of I/O  
requests that can be outstanding on a path before DMP invokes I/O throttling.  
The default value of queuedepthis 40. Setting queuedepthto a larger value  
allows more I/O requests to become queued up in the SCSI driver before I/O  
throttling is invoked.  
Note: The iotimeoutand queuedepthattributes are mutually exclusive.  
The following example sets the value of iotimeoutto 60 seconds for the  
enclosure enc0:  
# vxdmpadm setattr enclosure enc0 recoveryoption=throttle \  
iotimeout=60  
The next example sets the value of queuedepthto 30 for the paths to all  
Active/Active arrays:  
# vxdmpadm setattr arraytype A/A recoveryoption=throttle \  
queuedepth=30  
Specifying recoveryoption=defaultresets I/O throttling to the default settings  
corresponding to recoveryoption=throttleiotimeout=10, for example:  
# vxdmpadm setattr arraytype A/A recoveryoption=default  
This command also has the effect of configuring a fixed-retry limit of 30 on the  
paths.  
See “Configuring the response to I/O failures” on page 156 for details.  
Note: The I/O throttling settings are persistent across reboots of the system.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Administering dynamic multipathing (DMP) 159  
Administering DMP using vxdmpadm  
Displaying recoveryoption values  
The following example shows the vxdmpadmgetattrcommandbeing used to  
display the recoveryoptionoption values that are set on an enclosure.  
# vxdmpadm getattr enclosure HDS9500-ALUA0 recoveryoption  
ENCLR-NAME  
RECOVERY-OPTION DEFAULT[VAL]  
CURRENT[VAL]  
===============================================================  
HDS9500-ALUA0 Throttle  
HDS9500-ALUA0 Error-Retry  
Timebound[10] Queuedepth[60]  
Fixed-Retry[30]Timebound[20]  
This shows the default and current policy options and their values. The possible  
option settings are summarized in the following table.  
Recovery option  
Possible settings  
Description  
Error-Retry settings:  
recoveryoption=fixedretry  
Fixed-Retry (retrycount)  
DMP retries a failed I/O  
request for the specified  
number of times if I/O fails.  
recoveryoption=timebound Timebound (iotimeout)  
Throttle settings:  
DMP retries a failed I/O  
request after the specified  
time in seconds if I/O fails.  
recoveryoption=nothrottle  
recoveryoption=throttle  
None  
Not applicable  
Queuedepth (queuedepth)  
DMP throttles the path if  
the specified number of  
queued I/O requests is  
exceeded.  
Timebound (iotimeout)  
DMP throttles the path if an  
I/O request does not return  
within the specified time in  
seconds.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
160 Administering dynamic multipathing (DMP)  
Administering DMP using vxdmpadm  
Configuring DMP path restoration policies  
DMP maintains a kernel thread that re-examines the condition of paths at a  
specified interval. The type of analysis that is performed on the paths depends  
on the checking policy that is configured.  
Note: The DMP path restoration thread does not change the disabled state of the  
path through a controller that you have disabled using vxdmpadmdisable.  
Use the startrestorecommand to configure one of the following policies:  
■ check_all  
The path restoration thread analyzes all paths in the system and revives the  
paths that are back online, as well as disabling the paths that are  
inaccessible. The command to configure this policy is:  
# vxdmpadm start restore [interval=seconds] policy=check_all  
■ check_alternate  
The path restoration thread checks that at least one alternate path is  
healthy. It generates a notification if this condition is not met. This policy  
avoids inquiry commands on all healthy paths, and is less costly than  
check_allin cases where a large number of paths are available. This  
policy is the same as check_allif there are only two paths per DMP node.  
The command to configure this policy is:  
# vxdmpadm start restore [interval=seconds] \  
policy=check_alternate  
■ check_disabled  
This is the default path restoration policy. The path restoration thread  
checks the condition of paths that were previously disabled due to hardware  
failures, and revives them if they are back online. The command to  
configure this policy is:  
# vxdmpadm start restore [interval=seconds] \  
policy=check_disabled  
■ check_periodic  
The path restoration thread performs check_allonce in a given number  
of cycles, and check_disabledin the remainder of the cycles. This policy  
may lead to periodic slowing down (due to check_all) if there is a large  
number of paths available. The command to configure this policy is:  
# vxdmpadm start restore interval=seconds \  
policy=check_periodic [period=number]  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                   
Administering dynamic multipathing (DMP) 161  
Administering DMP using vxdmpadm  
The intervalattribute must be specified for this policy. The default  
number of cycles between running the check_allpolicy is 10.  
The interval attribute specifies how often the path restoration thread examines  
the paths. For example, after stopping the path restoration thread, the polling  
interval can be set to 400 seconds using the following command:  
# vxdmpadm start restore interval=400  
Note: The default interval is 300 seconds. Decreasing this interval can adversely  
affect system performance.  
To change the interval or policy, first stop the path restoration thread as  
described in “Stopping the DMP path restoration thread” on page 161, and then  
restart it with new attributes.  
See the vxdmpadm(1M) manual page for more information about DMP restore  
policies.  
Stopping the DMP path restoration thread  
Use the following command to stop the DMP path restoration thread:  
# vxdmpadm stop restore  
Note: Automatic path failback stops if the path restoration thread is stopped.  
Displaying the status of the DMP path restoration thread  
Use the following command to display the status of the automatic path  
restoration kernel thread, its polling interval, and the policy that it uses to check  
the condition of paths:  
# vxdmpadm stat restored  
This produces output such as the following:  
The number of daemons running : 1  
The interval of daemon: 300  
The policy of daemon: check_disabled  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
162 Administering dynamic multipathing (DMP)  
Administering DMP using vxdmpadm  
Displaying information about the DMP error-handling thread  
To display information about the kernel thread that handles DMP errors, use the  
following command:  
# vxdmpadm stat errord  
One daemon should be shown as running.  
Configuring array policy modules  
An array policy module (APM) is a dynamically loadable kernel module that may  
be provided by some vendors for use in conjunction with an array. An APM  
defines procedures to:  
■
Select an I/O path when multiple paths to a disk within the array are  
available.  
■
■
■
■
Select the path failover mechanism.  
Select the alternate path in the case of a path failure.  
Put a path change into effect.  
Respond to SCSI reservation or release requests.  
DMP supplies default procedures for these functions when an array is  
registered. An APM may modify some or all of the existing procedures that are  
provided by DMP or by another version of the APM.  
You can use the following command to display all the APMs that are configured  
for a system:  
# vxdmpadm listapm all  
The output from this command includes the file name of each module, the  
supported array type, the APM name, the APM version, and whether the module  
is currently in use (loaded). To see detailed information for an individual  
module, specify the module name as the argument to the command:  
# vxdmpadm listapm module_name  
To add and configure an APM, use the following command:  
# vxdmpadm -a cfgapm module_name [attr1=value1 \  
[attr2=value2 ...]]  
The optional configuration attributes and their values are specific to the APM  
for an array. Consult the documentation that is provided by the array vendor for  
details.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
Administering dynamic multipathing (DMP) 163  
Administering DMP using vxdmpadm  
Note: By default, DMP uses the most recent APM that is available. Specify the -u  
option instead of the -aoption if you want to force DMP to use an earlier  
version of the APM. The current version of an APM is replaced only if it is not in  
use.  
Specifying the -roption allows you to remove an APM that is not currently  
loaded:  
# vxdmpadm -r cfgapm module_name  
For more information about configuring APMs, see the vxdmpadm(1M) manual  
page.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
164 Administering dynamic multipathing (DMP)  
Administering DMP using vxdmpadm  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter  
4
Creating and  
administering disk groups  
This chapter describes how to create and manage disk groups. Disk groups are  
named collections of disks that share a common configuration. Volumes are  
created within a disk group and are restricted to using disks within that disk  
group.  
Note: In releases of Veritas Volume Manager (VxVM) prior to 4.0, a system  
installed with VxVM was configured with a default disk group, rootdg, that had  
to contain at least one disk. By default, operations were directed to the rootdg  
disk group. From release 4.0 onward, VxVM can function without any disk group  
having been configured. Only when the first disk is placed under VxVM control  
must a disk group be configured. There is no longer a requirement that you  
name any disk group rootdg, and any disk group that is named rootdghas no  
special properties because of this name. See “Specifying a disk group to  
commands” on page 167 for more information about using disk group names  
that are reserved for special purposes.  
Additionally, prior to VxVM 4.0, some commands such as vxdiskwere able to  
deduce the disk group if the name of an object was uniquely defined in one disk  
group among all the imported disk groups. Resolution of a disk group in this way  
is no longer supported for any command.  
For a discussion of disk groups that are compatible with the Cross-platform Data  
Sharing (CDS) feature of Veritas Volume Manager, see the Veritas Storage  
Foundation Cross-Platform Data Sharing Administrator’s Guide. The CDS  
feature allows you to move VxVM disks and objects between machines that are  
running under different operating systems.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
166 Creating and administering disk groups  
As system administrator, you can create additional disk groups to arrange your  
system’s disks for different purposes. Many systems do not use more than one  
disk group, unless they have a large number of disks. Disks can be initialized,  
reserved, and added to disk groups at any time. Disks need not be added to disk  
groups until the disks are needed to create VxVM objects.  
When a disk is added to a disk group, it is given a name (for example, mydg02).  
This name identifies a disk for operations such as volume creation or mirroring.  
The name also relates directly to the underlying physical disk. If a physical disk  
is moved to a different target address or to a different controller, the name  
mydg02continues to refer to it. Disks can be replaced by first associating a  
different physical disk with the name of the disk to be replaced and then  
recovering any volume data that was stored on the original disk (from mirrors or  
backup copies).  
Having disk groups that contain many disks and VxVM objects causes the  
private region to fill. In the case of large disk groups that are expected to contain  
more than several hundred disks and VxVM objects, disks should be set up with  
larger private areas. A major portion of a private region provides space for a disk  
group configuration database that contains records for each VxVM object in that  
disk group. Because each configuration record takes up approximately 256  
bytes, the number of records that can be created in a disk group can be  
estimated from the configuration database copy size. The copy size in blocks can  
be obtained from the output of the command vxdglistdiskgroup as the value of  
the permlenparameter on the line starting with the string “config:”. This  
value is the smallest of the lenvalues for all copies of the configuration  
database in the disk group. The amount of remaining free space in the  
configuration database is shown as the value of the freeparameter. An  
example is shown in “Displaying disk group information” on page 169. One way  
to overcome the problem of running out of free space is to split the affected disk  
group into two separate disk groups. See “Reorganizing the contents of disk  
groups” on page 195 for details.  
For information on backing up and restoring disk group configurations, see  
“Backing up and restoring disk group configuration data” on page 213.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Creating and administering disk groups 167  
Specifying a disk group to commands  
Specifying a disk group to commands  
Note: Most VxVM commands require superuser or equivalent privileges.  
Many VxVM commands allow you to specify a disk group using the -goption.  
For example, the following command creates a volume in the disk group, mktdg:  
# vxassist -g mktdg make mktvol 5g  
The block special device corresponding to this volume is:  
/dev/vx/dsk/mktdg/mktvol  
System-wide reserved disk groups  
The following disk group names are reserved, and cannot be used to name any  
disk groups that you create:  
bootdg  
Specifes the boot disk group. This is an alias for the disk group that  
contains the volumes that are used to boot the system. VxVM sets  
bootdgto the appropriate disk group if it takes control of the root  
disk. Otherwise, bootdgis set to nodg(no disk group; see below).  
Caution: Do not attempt to change the assigned value of bootdg. Doing so may  
render your system unbootable.  
defaultdg Specifies the default disk group. This is an alias for the disk group  
name that should be assumed if the -goption is not specified to a  
command, or if the VXVM_DEFAULTDGenvironment variable is  
undefined. By default, defaultdgis set to nodg(no disk group;  
see below).  
nodg  
Specifies to an operation that no disk group has been defined. For  
example, if the root disk is not under VxVM control, bootdgis set  
to nodg.  
Note: If you have upgraded your system, you may find it convenient to continue  
to configure a disk group named rootdgas the default disk group (defaultdg).  
There is no requirement that both defaultdgand bootdgrefer to the same  
disk group, nor that either the default disk group or the boot disk group be  
named rootdg.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
               
168 Creating and administering disk groups  
Specifying a disk group to commands  
Rules for determining the default disk group  
It is recommended that you use the -goption to specify a disk group to VxVM  
commands that accept this option. If you do not specify the disk group, VxVM  
applies rules in the following order until it determines a disk group name:  
■
■
■
Use the default disk group name that is specified by the environment  
variable VXVM_DEFAULTDG. This variable can also be set to one of the  
reserved system-wide disk group names: bootdg, defaultdg, or nodg. If  
the variable is undefined, the following rule is applied.  
Use the disk group that has been assigned to the system-wide default disk  
group alias, defaultdg. See “Displaying and specifying the system-wide  
default disk group” on page 168. If this alias is undefined, the following rule  
is applied.  
If the operation can be performed without requiring a disk group name (for  
example, an edit operation on disk access records), do so.  
If none of these rules succeeds, the requested operation fails.  
Caution: In releases of VxVM prior to 4.0, a subset of commands attempted to  
deduce the disk group by searching for the object name that was being operated  
upon by a command. This functionality is no longer supported. Scripts that rely  
on deducing the disk group from an object name may fail.  
Displaying the system-wide boot disk group  
To display the currently defined system-wide boot disk group, use the following  
command:  
# vxdg bootdg  
See the vxdg(1M) manual page for more information.  
Displaying and specifying the system-wide default disk group  
To display the currently defined system-wide default disk group, use the  
following command:  
# vxdg defaultdg  
If a default disk group has not been defined, nodgis displayed. Alternatively,  
you can use the following command to display the default disk group:  
# vxprint -Gng defaultdg 2>/dev/null  
In this case, if there is no default disk group, nothing is displayed.  
Use the following command to specify the name of the disk group that is aliased  
by defaultdg:  
# vxdctl defaultdg diskgroup  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                   
Creating and administering disk groups 169  
Displaying disk group information  
If bootdgis specified as the argument to this command, the default disk group  
is set to be the same as the currently defined system-wide boot disk group.  
If nodgis specified as the argument to the vxdctldefaultdgcommand, the  
default disk group is undefined.  
Note: The specified diskgroup need not currently exist on the system.  
See the vxdctl(1M) and vxdg(1M) manual pages for more information.  
Displaying disk group information  
To display information on existing disk groups, enter the following command:  
# vxdg list  
NAME  
rootdg enabled  
newdg enabled  
STATE  
ID  
730344554.1025.tweety  
731118794.1213.tweety  
To display more detailed information on a specific disk group, use the following  
command:  
# vxdg list diskgroup  
The output from this command when applied to a disk group named mydgis  
similar to the following:  
# vxdg list mydg  
Group: mydg  
dgid: 962910960.1025.bass  
import-id: 0.1  
flags:  
version: 140  
local-activation: read-write  
alignment : 512 (bytes)  
ssb: on  
detach-policy: local  
copies: nconfig=default nlog=default  
config: seqno=0.1183 permlen=3448 free=3428 templen=12  
loglen=522  
config disk c0t10d0 copy 1 len=3448 state=clean online  
config disk c0t11d0 copy 1 len=3448 state=clean online  
log disk c0t10d0 copy 1 len=522  
log disk c0t11d0 copy 1 len=522  
To verify the disk group ID and name associated with a specific disk (for  
example, to import the disk group), use the following command:  
# vxdisk -s list devicename  
This command provides output that includes the following information for the  
specified disk. For example, output for disk c0t12d0as follows:  
Disk: c0t12d0  
type:  
simple  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
170 Creating and administering disk groups  
Creating a disk group  
flags: online ready private autoconfig autoimport imported  
diskid: 963504891.1070.bass  
dgname: newdg  
dgid:  
963504895.1075.bass  
hostid: bass  
info:  
privoffset=128  
Displaying free space in a disk group  
Before you add volumes and file systems to your system, make sure you have  
enough free disk space to meet your needs.  
To display free space in the system, use the following command:  
# vxdg free  
The following is example output:  
GROUP DISK  
DEVICE  
TAG  
OFFSET  
LENGTH  
FLAGS  
mydg mydg01  
mydg mydg02  
newdg newdg01  
newdg newdg02  
oradg oradg01  
c0t10d0  
c0t11d0  
c0t12d0  
c0t13d0  
c0t14d0  
c0t10d0  
c0t11d0  
c0t12d0  
c0t13d0  
c0t14d0  
0
0
0
0
0
4444228 -  
4443310 -  
4443310 -  
4443310 -  
4443310 -  
To display free space for a disk group, use the following command:  
# vxdg -g diskgroup free  
where -gdiskgroup optionally specifies a disk group.  
For example, to display the free space in the disk group, mydg, use the following  
command:  
# vxdg -g mydg free  
The following example output shows the amount of free space in sectors:  
DISK  
DEVICE  
TAG  
OFFSET  
LENGTH  
FLAGS  
mydg01 c0t10d0  
mydg02 c0t11d0  
c0t10d0  
c0t11d0  
0
0
4444228  
4443310  
-
-
Creating a disk group  
Data related to a particular set of applications or a particular group of users may  
need to be made accessible on another system. Examples of this are:  
■
■
A system has failed and its data needs to be moved to other systems.  
The work load must be balanced across a number of systems.  
Disks must be placed in one or more disk groups before VxVM can use the disks  
for volumes. It is important that you locate data related to particular  
applications or users on an identifiable set of disks. When you need to move  
these disks, this allows you to move only the application or user data that should  
be moved.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Creating and administering disk groups 171  
Adding a disk to a disk group  
A disk group must have at least one disk associated with it. A new disk group can  
be created when you use menu item 1 (Add or initialize one or more  
disks)of the vxdiskadmcommand to add disks to VxVM control, as described  
in “Adding a disk to VxVM” on page 97. The disks to be added to a disk group  
must not belong to an existing disk group.  
You can also use the vxdiskaddcommand to create a new disk group:  
# vxdiskadd c1t0d0  
where c1t0d0in this example is the device name of a disk that is not currently  
assigned to a disk group. The command dialog is similar to that described for the  
vxdiskadmcommand in “Adding a disk to VxVM” on page 97.  
Disk groups can also be created by using the vxdginitcommand:  
# vxdg init diskgroup [cds=on|off] diskname=devicename  
For example, to create a disk group named mktdgon device c1t0d0:  
# vxdg init mktdg mktdg01=c1t0d0  
The disk specified by the device name, c1t0d0, must have been previously  
initialized with vxdiskaddor vxdiskadm, and must not currently belong to a  
disk group.  
You can use the cdsattribute with the vxdginitcommand to specify whether a  
new disk group is compatible with the Cross-platform Data Sharing (CDS)  
feature. In Veritas Volume Manager 4.0 and later releases, newly created disk  
groups are compatible with CDS by default (equivalent to specifying cds=on). If  
you want to change this behavior, edit the file /etc/default/vxdg, and set  
the attribute-value pair cds=offin this file before creating a new disk group.  
Alternatively, you can use the following command to set this attribute for a disk  
group:  
# vxdg -g diskgroup set cds=on|off  
Adding a disk to a disk group  
To add a disk to an existing disk group, use menu item 1 (Add or  
initialize one or more disks)of the vxdiskadmcommand. For details  
of this procedure, see “Adding a disk to VxVM” on page 97.  
You can also use the vxdiskaddcommand to add a disk to a disk group, for  
example:  
# vxdiskadd c1t1d0  
where c1t1d0is the device name of a disk that is not currently assigned to a  
disk group. The command dialog is similar to that described for the vxdiskadm  
command in “Adding a disk to VxVM” on page 97.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                       
172 Creating and administering disk groups  
Removing a disk from a disk group  
Removing a disk from a disk group  
Note: Before you can remove the last disk from a disk group, you must disable  
the disk group as described in “Disabling a disk group” on page 207.  
Alternatively, you can destroy the disk group as described in “Destroying a disk  
group” on page 208.  
A disk that contains no subdisks can be removed from its disk group with this  
command:  
# vxdg [-g diskgroup] rmdisk diskname  
For example, to remove mydg02from the disk group, mydg, use this command:  
# vxdg -g mydg rmdisk mydg02  
If the disk has subdisks on it when you try to remove it, the following error  
message is displayed:  
VxVM vxdg ERROR V-5-1-552 Disk diskname is used by one or more  
subdisks  
Use -k to remove device assignment.  
Using the -koption allows you to remove the disk even if subdisks are present.  
For more information, see the vxdg(1M) manual page.  
Caution: Use of the -koption to vxdgcan result in data loss.  
Once the disk has been removed from its disk group, you can (optionally) remove  
it from VxVM control completely, as follows:  
# vxdiskunsetup devicename  
For example, to remove c1t0d0from VxVM control, use these commands:  
# vxdiskunsetup c1t0d0  
You can remove a disk on which some subdisks of volumes are defined. For  
example, you can consolidate all the volumes onto one disk. If you use  
vxdiskadmto remove a disk, you can choose to move volumes off that disk. To  
do this, run vxdiskadmand select item 2(Remove a disk) from the main  
menu.  
If the disk is used by some volumes, this message is displayed:  
VxVM ERROR V-5-2-369 The following volumes currently use part of  
disk mydg02:  
home usrvol  
Volumes must be moved from mydg02 before it can be removed.  
Move volumes to other disks? [y,n,q,?] (default: n)  
If you choose y, then all volumes are moved off the disk, if possible. Some  
volumes may not be movable. The most common reasons why a volume may not  
be movable are as follows:  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                 
Creating and administering disk groups 173  
Deporting a disk group  
■
■
There is not enough space on the remaining disks.  
Plexes or striped subdisks cannot be allocated on different disks from  
existing plexes or striped subdisks in the volume.  
If vxdiskadmcannot move some volumes, you may need to remove some plexes  
from some disks to free more space before proceeding with the disk removal  
operation.  
Deporting a disk group  
Deporting a disk group disables access to a disk group that is currently enabled  
(imported) by the system. Deport a disk group if you intend to move the disks in  
a disk group to another system. Also, deport a disk group if you want to use all of  
the disks remaining in a disk group for a new purpose.  
To deport a disk group  
1
Stop all activity by applications to volumes that are configured in the disk  
group that is to be deported. Unmount file systems and shut down databases  
that are configured on the volumes.  
Note: Deportation fails if the disk group contains volumes that are in use  
(for example, by mounted file systems or databases).  
2
3
4
Use the following command to stop the volumes in the disk group:  
# vxvol -g diskgroup stopall  
Select menu item 8 (Remove access to (deport) a disk group)  
from the vxdiskadmmain menu.  
At the following prompt, enter the name of the disk group to be deported (in  
this example, newdg):  
Remove access to (deport) a disk group  
Menu: VolumeManager/Disk/DeportDiskGroup  
Use this menu operation to remove access to  
a disk group that is currently enabled (imported) by this  
system.  
Deport a disk group if you intend to move the disks in a disk  
group to another system. Also, deport a disk group if you  
want to use all of the disks remaining in a disk group for some  
new purpose.  
You will be prompted for the name of a disk group. You will  
also be asked if the disks should be disabled (offlined). For  
removable disk devices on some systems, it is important to  
disable all access to the disk before removing the disk.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
174 Creating and administering disk groups  
Importing a disk group  
Enter name of disk group [<group>,list,q,?] (default: list)  
newdg  
5
At the following prompt, enter yif you intend to remove the disks in this  
disk group:  
VxVM INFO V-5-2-377 The requested operation is to disable  
access to the removable disk group named newdg. This disk  
group is stored on the following disks:  
newdg01 on device c1t1d0  
You can choose to disable access to (also known as “offline”)  
these disks. This may be necessary to prevent errors if you  
actually remove any of the disks from the system.  
Disable (offline) the indicated disks? [y,n,q,?]  
(default: n) y  
6
7
At the following prompt, press Return to continue with the operation:  
Continue with operation? [y,n,q,?] (default: y)  
Once the disk group is deported, the vxdiskadmutility displays the  
following message:  
VxVM INFO V-5-2-269 Removal of disk group newdg was  
successful.  
At the following prompt, indicate whether you want to disable another disk  
group (y) or return to the vxdiskadmmain menu (n):  
Disable another disk group? [y,n,q,?] (default: n)  
Alternatively, you can use the vxdgcommand to deport a disk group:  
# vxdg deport diskgroup  
Importing a disk group  
Importing a disk group enables access by the system to a disk group. To move a  
disk group from one system to another, first disable (deport) the disk group on  
the original system, and then move the disk between systems and enable  
(import) the disk group.  
To import a disk group  
1
Use the following command to ensure that the disks in the deported disk  
group are online:  
# vxdisk -s list  
2
3
Select menu item 7 (Enable access to (import) a disk group)  
from the vxdiskadmmain menu.  
At the following prompt, enter the name of the disk group to import (in this  
example, newdg):  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Creating and administering disk groups 175  
Handling disks with duplicated identifiers  
Enable access to (import) a disk group  
Menu: VolumeManager/Disk/EnableDiskGroup  
Use this operation to enable access to a  
disk group. This can be used as the final part of moving a disk  
group from one system to another. The first part of moving a  
disk group is to use the “Remove access to (deport) a disk  
group” operation on the original host.  
A disk group can be imported from another host that failed  
without first deporting the disk group. Be sure that all disks  
in the disk group are moved between hosts.  
If two hosts share a SCSI bus, be very careful to ensure that  
the other host really has failed or has deported the disk  
group.  
If two active hosts import a disk group at the same time, the  
disk group will be corrupted and will become unusable.  
Select disk group to import [<group>,list,q,?] (default: list)  
newdg  
Once the import is complete, the vxdiskadmutility displays the following  
success message:  
VxVM INFO V-5-2-374 The import of newdg was successful.  
4
At the following prompt, indicate whether you want to import another disk  
group (y) or return to the vxdiskadmmain menu (n):  
Select another disk group? [y,n,q,?] (default: n)  
Alternatively, you can use the vxdgcommand to import a disk group:  
# vxdg import diskgroup  
Handling disks with duplicated identifiers  
Advanced disk arrays provide hardware tools that you can use to create clones  
of existing disks outside the control of VxVM. For example, these disks may have  
been created as hardware snapshots or mirrors of existing disks in a disk group.  
As a result, the VxVM private region is also duplicated on the cloned disk. When  
the disk group containing the original disk is subsequently imported, VxVM  
detects multiple disks that have the same disk identifier that is defined in the  
private region. In releases prior to 5.0, if VxVM could not determine which disk  
was the original, it would not import such disks into the disk group. The  
duplicated disks would have to be re-initialized before they could be imported.  
From release 5.0, a unique disk identifier (UDID) is added to the disk’s private  
region when the disk is initialized or when the disk is imported into a disk group  
(if this identifier does not already exist). Whenever a disk is brought online, the  
current UDID value that is known to the Device Discovery Layer (DDL) is  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
176 Creating and administering disk groups  
Handling disks with duplicated identifiers  
compared with the UDID that is set in the disk’s private region. If the UDID  
values do not match, the udid_mismatchflag is set on the disk. This flag can be  
viewed with the vxdisklistcommand.  
A new set of vxdiskand vxdgoperations are provided to handle such disks;  
either by writing the DDL value of the UDID to a disk’s private region, or by  
tagging a disk and specifying that it is a cloned disk to the vxdgimport  
operation.  
The following is sample output from the vxdisklistcommand showing that  
disks c2t66d0, c2t67d0and c2t68d0are marked with the udid_mismatch  
flag:  
# vxdisk list  
DEVICE  
c0t06d0  
c0t16d0  
...  
TYPE  
auto:cdsdisk  
auto:cdsdisk  
DISK  
-
-
GROUP STATUS  
-
-
online  
online  
c2t64d0  
c2t65d0  
c2t66d0  
c2t67d0  
c2t68d0  
auto:cdsdisk  
auto:cdsdisk  
auto:cdsdisk  
auto:cdsdisk  
auto:cdsdisk  
-
-
-
-
-
-
-
-
-
-
online  
online  
online udid_mismatch  
online udid_mismatch  
online udid_mismatch  
Writing a new UDID to a disk  
You can use the following command to update the unique disk identifier (UDID)  
for one or more disks:  
# vxdisk [-f] [-g diskgroup] updateudid disk ...  
This command uses the current value of the UDID that is stored in the Device  
Discovery Layer (DDL) database to correct the value in the private region. The -f  
option must be specified if VxVM has not set the udid_mismatchflag for a  
disk.  
For example, the following command updates the UDIDs for the disks c2t66d0  
and c2t67d0:  
# vxdisk updateudid c2t66d0 c2t67d0  
Importing a disk group containing cloned disks  
By default, disks on which the udid_mismatchflag or the clone_diskflag  
has been set are not imported by the vxdgimportcommand unless all disks in  
the disk group have at least one of these flags set, and no two of the disks have  
the same UDID. You can then import the cloned disks by specifying the  
-ouseclonedev=onoption to the vxdgimportcommand, as shown in this  
example:  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Creating and administering disk groups 177  
Handling disks with duplicated identifiers  
# vxdg -o useclonedev=on [-o updateid] import mydg  
Note: This form of the command allows only cloned disks to be imported. All  
non-cloned disks remain unimported.  
If the clone_diskflag is set on a disk, this indicates the disk was previously  
imported into a disk group with the udid_mismatchflag set.  
The -o updateidoption can be specified to write new identification attributes  
to the disks, and to set the clone_diskflag on the disks. (The vxdisk set  
clone=oncommand can also be used to set the flag.) However, the import fails if  
multiple copies of one or more cloned disks exist. In this case, you can either  
update the UDIDs of the cloned disks as described in “Writing a new UDID to a  
disk” on page 176, or you can use the following command to tag all the disks in  
the disk group that are to be imported:  
# vxdisk [-g diskgroup] settag tagname disk ...  
where tagname is a string of up to 128 characters, not including spaces or tabs.  
For example, the following command sets the tag, my_tagged_disks, on  
several disks that are to be imported together:  
# vxdisk settag my_tagged_disks c2t66d0 c2t67d0  
To check which disks are tagged, use the vxdisklisttagcommand:  
# vxdisk listtag  
DANAME  
c0t06d0  
c0t16d0  
...  
DMNAME  
mydg01  
mydg02  
NAME  
-
-
VALUE  
-
-
c2t64d0  
c2t65d0  
c2t66d0  
c2t67d0  
c2t68d0  
mydg05  
mydg06  
mydg07  
mydg08  
mydg09  
my_tagged_disks  
my_tagged_disks  
my_tagged_disks  
my_tagged_disks  
-
-
-
-
-
-
The configuration database in a VM disk’s private region contains persistent  
configuration data (or metadata) about the objects in a disk group. This database  
is consulted by VxVM when the disk group is imported. At least one of the  
cloned disks that are being imported must contain a copy of the current  
configuration database in its private region.  
You can use the following command to ensure that a copy of the metadata is  
placed on a disk, regardless of the placement policy for the disk group:  
# vxdisk [-g diskgroup] set disk keepmeta=always  
Alternatively, use the following command to place a copy of the configuration  
database and kernel log on all disks in a disk group that share a given tag:  
# vxdg [-g diskgroup] set tagmeta=on tag=tagname nconfig=all \  
nlog=all  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
178 Creating and administering disk groups  
Handling disks with duplicated identifiers  
To check which disks in a disk group contain copies of this configuration  
information, use the vxdglistmetacommand:  
# vxdg [-q] listmeta diskgroup  
The -qoption can be specified to suppress detailed configuration information  
from being displayed.  
The tagged disks in the disk group may be imported by specifying the tag to the  
vxdgimportcommand in addition to the -ouseclonedev=onoption:  
# vxdg -o useclonedev=on -o tag=my_tagged_disks import mydg  
If you have already imported the non-cloned disks in a disk group, you can use  
the -nand -toption to specify a temporary name for the disk group containing  
the cloned disks:  
# vxdg -t -n clonedg -o useclonedev=on -o tag=my_tagged_disks \  
import mydg  
See “Renaming a disk group” on page 183 for more information.  
To remove a tag from a disk, use the vxdiskrmtagcommand, as shown in the  
following example:  
# vxdisk rmtag tag=my_tagged_disks c2t67d0  
For further information about the use of the vxdiskand vxdgcommands to tag  
disks, and handle duplicate UDIDs, see the vxdisk(1M) and vxdg(1M) manual  
pages.  
Sample cases of operations on cloned disks  
The following sections contain examples of operations that can be used with  
cloned disks:  
■
■
■
Enabling configuration database copies on tagged disks  
Importing cloned disks without tags  
Importing cloned disks with tags  
Enabling configuration database copies on tagged disks  
In this example, the following commands have been used to tag some of the  
disks in an Hitachi TagmaStore array:  
# vxdisk settag TagmaStore-USP0_28 t1=v1  
# vxdisk settag TagmaStore-USP0_28 t2=v2  
# vxdisk settag TagmaStore-USP0_24 t2=v2  
# vxdisk settag TagmaStore-USP0_25 t1=v1  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Creating and administering disk groups 179  
Handling disks with duplicated identifiers  
These tags can be viewed by using the vxdisklisttagcommand:  
# vxdisk listtag  
DEVICE  
NAME  
VALUE  
v2  
v1  
v1  
v2  
TagmaStore-USP0_24 t2  
TagmaStore-USP0_25 t1  
TagmaStore-USP0_28 t1  
TagmaStore-USP0_28 t2  
The following command ensures that configuration database copies and kernel  
log copies are maintained for all disks in the disk group mydgthat are tagged as  
t1:  
# vxdg -g mydg set tagmeta=on tag=t1 nconfig=all nlog=all  
The disks for which such metadata is maintained can be seen by using this  
command:  
# vxdisk -o alldgs list  
DEVICE  
TYPE  
DISK  
GROUP STATUS  
TagmaStore-USP0_10 auto:cdsdisk -  
-
online  
TagmaStore-USP0_24 auto:cdsdisk mydg02 mydg  
TagmaStore-USP0_25 auto:cdsdisk mydg03 mydg  
online  
online tagmeta  
online  
TagmaStore-USP0_26 auto:cdsdisk -  
TagmaStore-USP0_27 auto:cdsdisk -  
-
-
online  
TagmaStore-USP0_28 auto:cdsdisk mydg01 mydg  
online tagmeta  
Alternatively, the following command can be used to ensure that a copy of the  
metadata is kept with a disk:  
# vxdisk set TagmaStore-USP0_25 keepmeta=always  
# vxdisk -o alldgs list  
DEVICE  
TYPE  
DISK  
GROUP STATUS  
TagmaStore-USP0_10 auto:cdsdisk -  
TagmaStore-USP0_22 auto:cdsdisk -  
TagmaStore-USP0_23 auto:cdsdisk -  
-
-
-
online  
online  
online  
TagmaStore-USP0_24 auto:cdsdisk mydg02 mydg  
TagmaStore-USP0_25 auto:cdsdisk mydg03 mydg  
TagmaStore-USP0_28 auto:cdsdisk mydg01 mydg  
online  
online keepmeta  
online  
Importing cloned disks without tags  
In the first example, cloned disks (shadow image devices) from an Hitachi  
TagmaStore array are to be imported. The primary (non-cloned) disks, mydg01,  
mydg02and mydg03, are already imported, and the cloned disks are not tagged.  
# vxdisk -o alldgs list  
DEVICE  
TYPE  
DISK  
GROUP STATUS  
TagmaStore-USP0_3 auto:cdsdisk -  
TagmaStore-USP0_23 auto:cdsdisk mydg02  
TagmaStore-USP0_25 auto:cdsdisk mydg03  
TagmaStore-USP0_30 auto:cdsdisk -  
TagmaStore-USP0_31 auto:cdsdisk -  
TagmaStore-USP0_32 auto:cdsdisk mydg01  
(mydg) online udid_mismatch  
mydg online  
mydg online  
(mydg) online udid_mismatch  
(mydg) online udid_mismatch  
mydg online  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
180 Creating and administering disk groups  
Handling disks with duplicated identifiers  
To import the cloned disks, they must be assigned a new disk group name, and  
their UDIDs must be updated:  
# vxdg -n newdg -o useclonedev=on -o updateid import mydg  
# vxdisk -o alldgs list  
DEVICE  
TYPE  
DISK  
GROUP STATUS  
TagmaStore-USP0_3 auto:cdsdisk mydg03 newdg online clone_disk  
TagmaStore-USP0_23 auto:cdsdisk mydg02 mydg  
TagmaStore-USP0_25 auto:cdsdisk mydg03 mydg  
online  
online  
TagmaStore-USP0_30 auto:cdsdisk mydg02 newdg online clone_disk  
TagmaStore-USP0_31 auto:cdsdisk mydg01 newdg online clone_disk  
TagmaStore-USP0_32 auto:cdsdisk mydg01 mydg  
online  
Note that the state of the imported cloned disks has changed from online  
udid_mismatchto onlineclone_disk.  
In the next example, none of the disks (neither cloned nor non-cloned) have  
been imported:  
# vxdisk -o alldgs list  
DEVICE  
TYPE  
DISK  
GROUP STATUS  
TagmaStore-USP0_3 auto:cdsdisk -  
TagmaStore-USP0_23 auto:cdsdisk -  
TagmaStore-USP0_25 auto:cdsdisk -  
TagmaStore-USP0_30 auto:cdsdisk -  
TagmaStore-USP0_31 auto:cdsdisk -  
TagmaStore-USP0_32 auto:cdsdisk -  
(mydg) online udid_mismatch  
(mydg) online  
(mydg) online  
(mydg) online udid_mismatch  
(mydg) online udid_mismatch  
(mydg) online  
To import only the cloned disks into the mydgdisk group:  
# vxdg -o useclonedev=on -o updateid import mydg  
# vxdisk -o alldgs list  
DEVICE  
TYPE  
DISK  
GROUP STATUS  
online clone_disk  
(mydg) online  
(mydg) online  
TagmaStore-USP0_3 auto:cdsdisk mydg03 mydg  
TagmaStore-USP0_23 auto:cdsdisk -  
TagmaStore-USP0_25 auto:cdsdisk -  
TagmaStore-USP0_30 auto:cdsdisk mydg02 mydg  
TagmaStore-USP0_31 auto:cdsdisk mydg01 mydg  
online clone_disk  
online clone_disk  
TagmaStore-USP0_32 auto:cdsdisk -  
(mydg) online  
In the next example, a cloned disk (BCV device) from an EMC Symmetrix DMX  
array is to be imported. Before the cloned disk, EMC0_27, has been split off from  
the disk group, the vxdisklistcommand shows that it is in the error  
udid_mismatchstate:  
# vxdisk -o alldgs list  
DEVICE  
EMC0_1  
EMC0_27  
TYPE  
auto:cdsdisk EMC0_1 mydg  
auto  
DISK  
GROUP STATUS  
online  
-
-
error udid_mismatch  
The symmircommand is used to split off the BCV device:  
# /usr/symcli/bin/symmir -g mydg split DEV001  
After updating VxVM’s information about the disk by running the vxdisk  
scandiskscommand, the cloned disk is in the onlineudid_mismatchstate:  
# vxdisk -o alldgs list  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Creating and administering disk groups 181  
Handling disks with duplicated identifiers  
DEVICE  
EMC0_1  
EMC0_27  
TYPE  
DISK  
GROUP STATUS  
auto:cdsdisk EMC0_1 mydg  
auto:cdsdisk -  
online  
-
online udid_mismatch  
The following command imports the cloned disk into the new disk group newdg,  
and updates the disk’s UDID:  
# vxdg -n newdg -o useclonedev=on -o updateid import mydg  
The state of the cloned disk is now shown as onlineclone_disk:  
# vxdisk -o alldgs list  
DEVICE  
EMC0_1  
EMC0_27  
TYPE  
auto:cdsdisk EMC0_1 mydg  
auto:cdsdisk EMC0_1 newdg online clone_disk  
DISK  
GROUP STATUS  
online  
Importing cloned disks with tags  
In this example, cloned disks (BCV devices) from an EMC Symmetrix DMX array  
are to be imported. The primary (non-cloned) disks, mydg01, mydg02and  
mydg03, are already imported, and the cloned disks with the tag t1are to be  
imported.  
# vxdisk -o alldgs list  
DEVICE  
EMC0_4  
EMC0_6  
EMC0_8  
EMC0_15  
EMC0_18  
EMC0_24  
TYPE  
DISK  
GROUP STATUS  
auto:cdsdisk mydg01 mydg  
auto:cdsdisk mydg02 mydg  
auto:cdsdisk -  
auto:cdsdisk -  
auto:cdsdisk mydg03 mydg  
auto:cdsdisk - (mydg) online udid_mismatch  
online  
online  
(mydg) online udid_mismatch  
(mydg) online udid_mismatch  
online  
The disks are tagged as follows:  
# vxdisk listtag  
DEVICE  
EMC0_4  
EMC0_4  
EMC0_6  
EMC0_8  
EMC0_15  
EMC0_18  
EMC0_24  
EMC0_24  
NAME  
t2  
t1  
t2  
t1  
t2  
t1  
t1  
t2  
VALUE  
v2  
v1  
v2  
v1  
v2  
v1  
v1  
v2  
To import the cloned disks that are tagged as t1, they must be assigned a new  
disk group name, and their UDIDs must be updated:  
# vxdg -n newdg -o useclonedev=on -o tag=t1 -o updateid import mydg  
# vxdisk -o alldgs list  
DEVICE  
EMC0_4  
EMC0_6  
EMC0_8  
EMC0_15  
EMC0_18  
EMC0_24  
TYPE  
DISK  
GROUP STATUS  
auto:cdsdisk mydg01 mydg  
auto:cdsdisk mydg02 mydg  
auto:cdsdisk mydg03 newdg online clone_disk  
auto:cdsdisk - (mydg) online udid_mismatch  
auto:cdsdisk mydg03 mydg online  
auto:cdsdisk mydg01 newdg online clone_disk  
online  
online  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
182 Creating and administering disk groups  
Handling disks with duplicated identifiers  
As the cloned disk EMC0_15is not tagged as t1, it is not imported. Note that the  
state of the imported cloned disks has changed from onlineudid_mismatch  
to onlineclone_disk.  
By default, the state of imported cloned disks is shown as onlineclone_disk.  
This can be removed by using the vxdisk setcommand as shown here:  
# vxdisk set EMC0_8 clone=off  
# vxdisk -o alldgs list  
DEVICE  
EMC0_4  
EMC0_6  
EMC0_8  
EMC0_15  
EMC0_18  
EMC0_24  
TYPE  
DISK  
GROUP STATUS  
auto:cdsdisk mydg01 mydg  
auto:cdsdisk mydg02 mydg  
auto:cdsdisk mydg03 newdg online  
auto:cdsdisk -  
auto:cdsdisk mydg03 mydg  
online  
online  
(mydg) online udid_mismatch  
online  
auto:cdsdisk mydg01 newdg online clone_disk  
In the next example, none of the disks (neither cloned nor non-cloned) have  
been imported:  
# vxdisk -o alldgs list  
DEVICE  
EMC0_4  
EMC0_6  
EMC0_8  
EMC0_15  
EMC0_18  
EMC0_24  
TYPE  
DISK  
GROUP STATUS  
(mydg) online  
(mydg) online  
(mydg) online udid_mismatch  
(mydg) online udid_mismatch  
(mydg) online  
auto:cdsdisk -  
auto:cdsdisk -  
auto:cdsdisk -  
auto:cdsdisk -  
auto:cdsdisk -  
auto:cdsdisk -  
(mydg) online udid_mismatch  
To import only the cloned disks that have been tagged as t1into the mydgdisk  
group:  
# vxdg -o useclonedev=on -o tag=t1 -o updateid import mydg  
# vxdisk -o alldgs list  
DEVICE  
EMC0_4  
EMC0_6  
EMC0_8  
EMC0_15  
EMC0_18  
EMC0_24  
TYPE  
auto:cdsdisk -  
auto:cdsdisk -  
auto:cdsdisk mydg03 mydg  
auto:cdsdisk -  
auto:cdsdisk -  
DISK  
GROUP STATUS  
(mydg) online  
(mydg) online  
online clone_disk  
(mydg) online udid_mismatch  
(mydg) online  
auto:cdsdisk mydg01 mydg  
online clone_disk  
As in the previous example, the cloned disk EMC0_15is not tagged as t1, and so  
it is not imported.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Creating and administering disk groups 183  
Renaming a disk group  
Renaming a disk group  
Only one disk group of a given name can exist per system. It is not possible to  
import or deport a disk group when the target system already has a disk group of  
the same name. To avoid this problem, VxVM allows you to rename a disk group  
during import or deport.  
To rename a disk group during import, use the following command:  
# vxdg [-t] -n newdg import diskgroup  
If the -toption is included, the import is temporary and does not persist across  
reboots. In this case, the stored name of the disk group remains unchanged on  
its original host, but the disk group is known by the name specified by newdgto  
the importing host. If the -toption is not used, the name change is permanent.  
For example, this command temporarily renames the disk group, mydg, as  
mytempdgon import:  
# vxdg -t -n mytempdg import mydg  
To rename a disk group during deport, use the following command:  
# vxdg [-h hostname] -n newdg deport diskgroup  
When renaming on deport, you can specify the -hhostname option to assign a  
lock to an alternate host. This ensures that the disk group is automatically  
imported when the alternate host reboots.  
For example, this command renames the disk group, mydg, as myexdg, and  
deports it to the host, jingo:  
# vxdg -h jingo -n myexdg deport mydg  
Note: You cannot use this method to rename the active boot disk group because  
it contains volumes that are in use by mounted file systems (such as /). To  
rename the boot disk group, boot the system from an LVM root disk instead of  
from the VxVM root disk. You can then use the above methods to rename the  
boot disk group. See the sections under “Rootability” on page 102 for more  
information.  
To temporarily move the boot disk group, bootdg, from one host to another  
(for repair work on the root volume, for example) and then move it back  
1
On the original host, identify the disk group ID of the bootdgdisk group to  
be imported with the following command:  
# vxdisk -g bootdg -s list  
This command results in output such as the following:  
dgname: rootdg  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
184 Creating and administering disk groups  
Moving disks between disk groups  
dgid:  
774226267.1025.tweety  
Note: In this example, the administrator has chosen to name the boot disk  
group as rootdg. The ID of this disk group is 774226267.1025.tweety.  
This procedure assumes that all the disks in the boot disk group are  
accessible by both hosts.  
2
3
Shut down the original host.  
On the importing host, import and rename the rootdgdisk group with this  
command:  
# vxdg -tC -n newdg import diskgroup  
The -toption indicates a temporary import name, and the -Coption clears  
import locks. The -noption specifies an alternate name for the rootdg  
being imported so that it does not conflict with the existing rootdg.  
diskgroup is the disk group ID of the disk group being imported (for  
example, 774226267.1025.tweety).  
If a reboot or crash occurs at this point, the temporarily imported disk  
group becomes unimported and requires a reimport.  
4
After the necessary work has been done on the imported disk group, deport  
it back to its original host with this command:  
# vxdg -h hostname deport diskgroup  
Here hostname is the name of the system whose rootdgis being returned  
(the system name can be confirmed with the command uname-n).  
This command removes the imported disk group from the importing host  
and returns locks to its original host. The original host can then  
automatically import its boot disk group at the next reboot.  
Moving disks between disk groups  
To move a disk between disk groups, remove the disk from one disk group and  
add it to the other. For example, to move the physical disk c0t3d0(attached  
with the disk name salesdg04) from disk group salesdgand add it to disk  
group mktdg, use the following commands:  
# vxdg -g salesdg rmdisk salesdg04  
# vxdg -g mktdg adddisk mktdg02=c0t3d0  
Caution: This procedure does not save the configurations nor data on the disks.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Creating and administering disk groups 185  
Moving disk groups between systems  
You can also move a disk by using the vxdiskadmcommand. Select item 3  
(Remove a disk)from the main menu, and then select item 1 (Add or  
initialize a disk).  
See “Moving objects between disk groups” on page 203 for an alternative and  
preferred method of moving disks between disk groups. This method preserves  
VxVM objects, such as volumes, that are configured on the disks.  
Moving disk groups between systems  
An important feature of disk groups is that they can be moved between systems.  
If all disks in a disk group are moved from one system to another, then the disk  
group can be used by the second system. You do not have to re-specify the  
configuration.  
To move a disk group between systems  
1
On the first system, stop all volumes in the disk group, then deport (disable  
local access to) the disk group with the following command:  
# vxdg deport diskgroup  
2
Move all the disks to the second system and perform the steps necessary  
(system-dependent) for the second system and VxVM to recognize the new  
disks.  
This can require a reboot, in which case the vxconfigddaemon is  
restarted and recognizes the new disks. If you do not reboot, use the  
command vxdctlenableto restart the vxconfigdprogram so VxVM also  
recognizes the disks.  
3
4
Import (enable local access to) the disk group on the second system with this  
command:  
# vxdg import diskgroup  
Caution: All disks in the disk group must be moved to the other system. If  
they are not moved, the import fails.  
After the disk group is imported, start all volumes in the disk group with  
this command:  
# vxrecover -g diskgroup -sb  
You can also move disks from a system that has crashed. In this case, you cannot  
deport the disk group from the first system. When a disk group is created or  
imported on a system, that system writes a lock on all disks in the disk group.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
186 Creating and administering disk groups  
Moving disk groups between systems  
Caution: The purpose of the lock is to ensure that dual-ported disks (disks that  
can be accessed simultaneously by two systems) are not used by both systems at  
the same time. If two systems try to access the same disks at the same time, this  
must be managed using software such as the clustering functionality of VxVM.  
Otherwise, configuration information stored on the disk may be corrupted, and  
the data on the disk may become unusable.  
Handling errors when importing disks  
When you move disks from a system that has crashed or that failed to detect the  
group before the disk was moved, the locks stored on the disks remain and must  
be cleared. The system returns the following error message:  
VxVM vxdg ERROR V-5-1-587 disk group groupname: import failed:  
Disk is in use by another host  
The next message indicates that the disk group does not contains any valid disks  
(not that it does not contains any disks):  
VxVM vxdg ERROR V-5-1-587 Disk group groupname: import failed:  
No valid disk found containing disk group  
The disks may be considered invalid due to a mismatch between the host ID in  
their configuration copies and that stored in the /etc/vx/volbootfile.  
To clear locks on a specific set of devices, use the following command:  
# vxdisk clearimport devicename ...  
To clear the locks during import, use the following command:  
# vxdg -C import diskgroup  
Caution: Be careful when using the vxdiskclearimportor vxdg-Cimport  
command on systems that have dual-ported disks. Clearing the locks allows  
those disks to be accessed at the same time from multiple hosts and can result in  
corrupted data.  
A disk group can be imported successfully if all the disks are accessible that  
were visible when the disk group was last imported successfully. However,  
sometimes you may need to specify the -foption to forcibly import a disk group  
if some disks are not available. If the importoperation fails, an error message is  
displayed.  
The following error message indicates a fatal error that requires hardware  
repair or the creation of a new disk group, and recovery of the disk group  
configuration and data:  
VxVM vxdg ERROR V-5-1-587 Disk group groupname: import failed:  
Disk group has no valid configuration copies  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                 
Creating and administering disk groups 187  
Moving disk groups between systems  
The following error message indicates a recoverable error.  
VxVM vxdg ERROR V-5-1-587 Disk group groupname: import failed:  
Disk for disk group not found  
If some of the disks in the disk group have failed, you can force the disk group to  
be imported by specifying the -foption to the vxdgimportcommand:  
# vxdg -f import diskgroup  
Caution: Be careful when using the -foption. It can cause the same disk group  
to be imported twice from different sets of disks. This can cause the disk group  
configuration to become inconsistent.  
See “Handling conflicting configuration copies” on page 190.  
As using the -foption to force the import of an incomplete disk group counts as  
a successful import, an incomplete disk group may be imported subsequently  
without this option being specified. This may not be what you expect.  
These operations can also be performed using the vxdiskadmutility. To deport  
a disk group using vxdiskadm, select menu item 8 (Remove access to  
(deport) a disk group). To import a disk group, select item 7 (Enable  
access to (import) a disk group). The vxdiskadmimport operation  
checks for host import locks and prompts to see if you want to clear any that are  
found. It also starts volumes in the disk group.  
Reserving minor numbers for disk groups  
A device minor number uniquely identifies some characteristic of a device to the  
device driver that controls that device. It is often used to identify some  
characteristic mode of an individual device, or to identify separate devices that  
are all under the control of a single controller. VxVM assigns unique device  
minor numbers to each object (volume, plex, subdisk, disk, or disk group) that it  
controls.  
When you move a disk group between systems, it is possible for the minor  
numbers that it used on its previous system to coincide (or collide) with those of  
objects known to VxVM on the new system. To get around this potential  
problem, you can allocate separate ranges of minor numbers for each disk  
group. VxVM uses the specified range of minor numbers when it creates volume  
objects from the disks in the disk group. This guarantees that each volume has  
the same minor number across reboots or reconfigurations. Disk groups may  
then be moved between machines without causing device number collisions.  
VxVM chooses minor device numbers for objects created from this disk group  
starting at the base minor number base_minor. Minor numbers can range from  
this value up to 16,777,215. Try to leave a reasonable number of unallocated  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
188 Creating and administering disk groups  
Moving disk groups between systems  
minor numbers near the top of this range to allow for temporary device number  
remapping in the event that a device minor number collision may still occur.  
VxVM reserves the range of minor numbers from 0 to 999 for use with volumes  
in the boot disk group. For example, the rootvolvolume is always assigned  
minor number 0.  
If you do not specify the base of the minor number range for a disk group, VxVM  
chooses one at random. The number chosen is at least 1000, is a multiple of  
1000, and yields a usable range of 1000 device numbers. The chosen number  
also does not overlap within a range of 1000 of any currently imported disk  
groups, and it does not overlap any currently allocated volume device numbers.  
Note: The default policy ensures that a small number of disk groups can be  
merged successfully between a set of machines. However, where disk groups are  
merged automatically using failover mechanisms, select ranges that avoid  
overlap.  
To view the base minor number for an existing disk group, use the vxprint  
command as shown in the following examples for the disk group, mydg:  
# vxprint -l mydg | egrep minors  
minors: >=45000  
# vxprint -g mydg -m | egrep base_minor  
base_minor=45000  
To set a base volume device minor number for a disk group that is being created,  
use the following command:  
# vxdg init diskgroup minor=base_minor disk_access_name ...  
For example, the following command creates the disk group, newdg, that  
includes the specified disks, and has a base minor number of 30000:  
# xvdg init newdg minor=30000 c1d0t0 c1t1d0  
If a disk group already exists, you can use the vxdgreminorcommand to change  
its base minor number:  
# vxdg -g diskgroup reminor new_base_minor  
For example, the following command changes the base minor number to 30000  
for the disk group, mydg:  
# vxprint -g mydg reminor 30000  
If a volume is open, its old device number remains in effect until the system is  
rebooted or until the disk group is deported and re-imported. If you close the  
open volume, you can run vxdgreminoragain to allow the renumbering to take  
effect without rebooting or re-importing.  
An example of where it is necessary to change the base minor number is for a  
cluster-shareable disk group. The volumes in a shared disk group must have the  
same minor number on all the nodes. If there is a conflict between the minor  
numbers when a node attempts to join the cluster, the join fails. You can use the  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Creating and administering disk groups 189  
Moving disk groups between systems  
reminoroperation on the nodes that are in the cluster to resolve the conflict. In  
a cluster where more than one node is joined, use a base minor number which  
does not conflict on any node.  
For further information on minor number reservation, see the vxdg(1M)  
manual page.  
Compatibility of disk groups between platforms  
For disk groups that support the Cross-platform Data Sharing (CDS) feature, the  
upper limit on the minor number range is restricted on AIX, HP-UX, Linux (with  
a 2.6 or later kernel) and Solaris to 65,535 to ensure portability between these  
operating systems.  
On a Linux platform with a pre-2.6 kernel, the number of minor numbers per  
major number is limited to 256 with a base of 0. This has the effect of limiting  
the number of volumes and disks that can be supported system-wide to a smaller  
value than is allowed on other operating system platforms. The number of disks  
that are supported by a pre-2.6 Linux kernel is typically limited to a few  
hundred. With the extended major numbering scheme that was implemented in  
VxVM 4.0 on Linux, a maximum of 4079 volumes could be configured, provided  
that a contiguous block of 15 extended major numbers was available.  
VxVM 4.1 runs on a 2.6 version Linux kernel, which increases the number of  
minor devices that are configurable from 256 to 65,536 per major device  
number. This allows a large number of volumes and disk devices to be  
configured on a system. The theoretical limit on the number of DMP and volume  
devices that can be configured on such a system are 65,536 and 1,048,576  
respectively. However, in practice, the number of VxVM devices that can be  
configured in a single disk group is limited by the size of the private region.  
When a CDS-compatible disk group is imported on a Linux system with a pre-2.6  
kernel, VxVM attempts to reassign the minor numbers of the volumes, and fails  
if this is not possible.  
To help ensure that a CDS-compatible disk group is portable between operating  
systems, including Linux with a pre-2.6 kernel, use the following command to  
set the maxdevattribute on the disk group:  
# vxdg -g diskgroup set maxdev=4079  
Note: Such a disk group may still not be importable by VxVM 4.0 on Linux with a  
pre-2.6 kernel if it would increase the number of minor numbers on the system  
that are assigned to volumes to more than 4079, or if the number of available  
extended major numbers is smaller than 15.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
190 Creating and administering disk groups  
Handling conflicting configuration copies  
You can use the following command to discover the maximum number of  
volumes that are supported by VxVM on a Linux host:  
# cat /proc/sys/vxvm/vxio/vol_max_volumes  
4079  
See the vxdg(1M) manual page for more information.  
Handling conflicting configuration copies  
If an incomplete disk group is imported on several different systems, this can  
create inconsistencies in the disk group configuration copies that you may need  
to resolve manually. This section and following sections describe how such a  
condition can occur, and how to correct it. (When the condition occurs in a  
cluster that has been split, it is usually referred to as a serial split brain  
condition).  
Note: The procedures given here require that the version number of the disk  
group is at least 110.  
Example of a serial split brain condition in a cluster  
Note: This section presents an example of how a serial split brain condition  
might occur for a shared disk group in a cluster. For more information about  
shared disk groups in clusters, see “Administering cluster functionality” on  
page 397. Conflicts between configuration copies can also occur for private disk  
groups in clustered and non-clustered configurations where the disk groups  
have been partially imported on different systems. The procedure in “Correcting  
conflicting configuration information” on page 194 describes how to correct  
such problems.  
A campus cluster (also known as a stretch cluster or remote mirror  
configuration) typically consists of a 2-node cluster where each component  
(server, switch and storage) of the cluster exists in a separate building. Figure 4-  
1 illustrates a 2-node cluster with node 0, a fibre channel switch and disk  
enclosure enc0in building A, and node 1, another switch and enclosure enc1in  
building B. The fibre channel connectivity is multiply redundant to implement  
redundant-loop access between each node and each enclosure. As usual, the two  
nodes are also linked by a redundant private network.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Creating and administering disk groups 191  
Handling conflicting configuration copies  
Figure 4-1  
Typical arrangement of a 2-node campus cluster  
Redundant private  
network  
Node 0  
Node 1  
Fibre Channel  
switches  
Disk enclosures  
enc0  
enc1  
Building A  
Building B  
A serial split brain condition typically arises in a cluster when a private (non-  
shared) disk group is imported on Node 0 with Node 1 configured as the failover  
node.  
If the network connections between the nodes are severed, both nodes think that  
the other node has died. (This is the usual cause of the split brain condition in  
clusters). If a disk group is spread across both enclosure enc0and enc1, each  
portion loses connectivity to the other portion of the disk group. Node 0  
continues to update to the disks in the portion of the disk group that it can  
access. Node 1, operating as the failover node, imports the other portion of the  
disk group (with the -foption set), and starts updating the disks that it can see.  
When the network links are restored, attempting to reattach the missing disks  
to the disk group on Node 0, or to re-import the entire disk group on either node,  
fails. This serial split brain condition arises because VxVM increments the serial  
ID in the disk media record of each imported disk in all the disk group  
configuration databases on those disks, and also in the private region of each  
imported disk. The value that is stored in the configuration database represents  
the serial ID that the disk group expects a disk to have. The serial ID that is  
stored in a disk’s private region is considered to be its actual value.  
If some disks went missing from the disk group (due to physical disconnection  
or power failure) and those disks were imported by another host, the serial IDs  
Download from Www.Somanuals.com. All Manuals Search And Download.  
192 Creating and administering disk groups  
Handling conflicting configuration copies  
for the disks in their copies of the configuration database, and also in each disk’s  
private region, are updated separately on that host. When the disks are  
subsequently re-imported into the original shared disk group, the actual serial  
IDs on the disks do not agree with the expected values from the configuration  
copies on other disks in the disk group.  
Depending on what happened to the different portions of the split disk group,  
there are two possibilities for resolving inconsistencies between the  
configuration databases:  
■
If the other disks in the disk group were not imported on another host,  
VxVM resolves the conflicting values of the serial IDs by using the version  
of the configuration database from the disk with the greatest value for the  
updated ID (shown as the value of update_tidin the output from the  
vxprint -mdiskgroup | grep update_tidcommand). This case is  
illustrated below.  
Figure 4-2  
Example of a serial split brain condition that can be resolved  
automatically  
Disk B not imported  
Partial disk group  
imported on host X  
Disk A  
Disk B  
1.Disk A is imported on a separate  
host. Disk B is not imported. The  
actual and expected serial IDs are  
updated only on disk A.  
Actual A = 1  
Actual B = 0  
Configuration  
database  
Configuration  
database  
Expected A = 1  
Expected B = 0  
Expected A = 0  
Expected B = 0  
Imported shared disk group  
2.The disk group is re- imported on  
the cluster. The configuration copy  
on disk A is used to correct the  
configuration copy on disk B as the  
actual value of the updated ID on  
disk A is greatest.  
Disk A  
Disk B  
Actual A = 1  
Actual B = 0  
Configuration  
database  
Configuration  
database  
Expected A = 1  
Expected B = 0  
Expected A = 1  
Expected B = 0  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Creating and administering disk groups 193  
Handling conflicting configuration copies  
■
If the other disks were also imported on another host, no disk can be  
considered to have a definitive copy of the configuration database. The  
figure below illustrates how this condition can arise for two disks.  
Figure 4-3  
Exampleofatrueserialsplitbrainconditionthatcannotberesolved  
automatically  
Partial disk group  
imported on host Y  
Partial disk group  
imported on host X  
1.Disks A and B are imported  
independently on separate hosts.  
The actual and expected serial IDs  
are updated independently on  
each disk.  
Disk A  
Disk B  
Actual A = 1  
Actual B = 1  
Configuration  
database  
Configuration  
database  
Expected A = 1  
Expected B = 0  
Expected A = 0  
Expected B = 1  
Shared disk group fails to import  
2.The disk group cannot be  
reimported on the cluster. This is  
because the databases do not  
agree on the actual and expected  
serial IDs. You must choose which  
configuration database to use.  
Disk A  
Disk B  
Actual A = 1  
Actual B = 1  
Configuration  
database  
Configuration  
database  
Expected A = 1  
Expected B = 0  
Expected A = 0  
Expected B = 1  
This is a true serial split brain condition, which VxVM cannot correct  
automatically. In this case, the disk group import fails, and the vxdgutility  
outputs error messages similar to the following before exiting:  
VxVM vxconfigd NOTICE V-5-0-33 Split Brain. da id is 0.1, while  
dm id is 0.0 for DM mydg01  
VxVM vxdg ERROR V-5-1-587 Disk group newdg: import failed:  
Serial Split Brain detected. Run vxsplitlines  
The import does not succeed even if you specify the -fflag to vxdg.  
Although it is usually possible to resolve this conflict by choosing the  
version of the configuration database with the highest valued configuration  
ID (shown as the value of seqnoin the output from the vxdglistdiskgroup  
| grep configcommand), this may not be the correct thing to do in all  
circumstances.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
194 Creating and administering disk groups  
Handling conflicting configuration copies  
The following section, “Correcting conflicting configuration information,”  
describes how to fix this condition.  
For more information on how to set up and maintain a remote mirror  
configuration, see “Administering sites and remote mirrors” on page 431.  
Correcting conflicting configuration information  
Note: This procedure requires that the disk group has a version number of at  
least 110. See “Upgrading a disk group” on page 208 for more information about  
disk group version numbers.  
To resolve conflicting configuration information, you must decide which disk  
contains the correct version of the disk group configuration database. To assist  
you in doing this, you can run the vxsplitlinescommand to show the actual  
serial ID on each disk in the disk group and the serial ID that was expected from  
the configuration database. For each disk, the command also shows the vxdg  
command that you must run to select the configuration database copy on that  
disk as being the definitive copy to use for importing the disk group.  
The following is sample output from running vxsplitlineson the disk group  
newdg:  
# vxsplitlines -g newdg  
The following splits were found in disk group newdg  
They are listed in da(dm) name pairs.  
Pool 0.  
c2t5d0 ( c2t5d0 ), c2t6d0 ( c2t6d0 ),  
The configuration from any of the disks in this split should appear  
to be be the same.  
To see the configuration from any of the disks in this split, run:  
/etc/vx/diag.d/vxprivutil dumpconfig /dev/vx/dmp/c2t5d0  
To import the dg with the configuration from this split, run:  
/usr/sbin/vxdg -o selectcp=1045852127.32.olancha import newdg  
To get more information about this particular configuration, run:  
/usr/sbin/vxsplitlines -g newdg -c c2t5d0  
Split 1.  
c2t7d0 ( c2t7d0 ), c2t8d0 ( c2t8d0 ),  
The configuration from any of the disks in this split should appear  
to be be the same.  
To see the configuration from any of the disks in this split, run:  
/etc/vx/diag.d/vxprivutil dumpconfig /dev/vx/dmp/c2t7d0  
To import the dg with the configuration from this split, run:  
/usr/sbin/vxdg -o selectcp=1045852127.33.olancha import newdg  
To get more information about this particular configuration, run:  
/usr/sbin/vxsplitlines -g newdg -c c2t7d0  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Creating and administering disk groups 195  
Reorganizing the contents of disk groups  
In this example, the disk group has four disks, and is split so that two disks  
appear to be on each side of the split.  
You can specify the -coption to vxsplitlinesto print detailed information  
about each of the disk IDs from the configuration copy on a disk specified by its  
disk access name:  
# vxsplitlines -g newdg -c c2t6d0  
DANAME(DMNAME)  
|| Actual SSB  
|| Expected SSB  
c2t5d0( c2t5d0 ) || 0.1  
c2t6d0( c2t6d0 ) || 0.1  
c2t7d0( c2t7d0 ) || 0.1  
c2t8d0( c2t8d0 ) || 0.1  
|| 0.0 ssb ids don’t match  
|| 0.1 ssb ids match  
|| 0.1 ssb ids match  
|| 0.0 ssb ids don’t match  
Please note that even though some disks ssb ids might match  
that does not necessarily mean that those disks’ config copies  
have all the changes. From some other configuration copies,  
those disks’ ssb ids might not match.  
To see the configuration from this disk, run  
/etc/vx/diag.d/vxprivutil dumpconfig /dev/vx/dmp/c2t6d0  
Based on your knowledge of how the serial split brain condition came about, you  
must choose one disk’s configuration to be used to import the disk group. For  
example, the following command imports the disk group using the  
configuration copy that is on side 0 of the split:  
# /usr/sbin/vxdg -o selectcp=1045852127.32.olancha import newdg  
When you have selected a preferred configuration copy, and the disk group has  
been imported, VxVM resets the serial IDs to 0 for the imported disks. The  
actual and expected serial IDs for any disks in the disk group that are not  
imported at this time remain unaltered.  
Reorganizing the contents of disk groups  
TM  
Note: You need a Veritas FlashSnap license to use this feature.  
There are several circumstances under which you might want to reorganize the  
contents of your existing disk groups:  
■
■
To group volumes or disks differently as the needs of your organization  
change. For example, you might want to split disk groups to match the  
boundaries of separate departments, or to join disk groups when  
departments are merged.  
To reduce the size of a disk group’s configuration database in the event that  
its private region is nearly full. This is a much simpler solution than the  
alternative of trying to grow the private region.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
196 Creating and administering disk groups  
Reorganizing the contents of disk groups  
■
■
To perform online maintenance and upgrading of fault-tolerant systems  
that can be split into separate hosts for this purpose, and then rejoined.  
To isolate volumes or disks from a disk group, and process them  
independently on the same host or on a different host. This allows you to  
implement off-host processing solutions for the purposes of backup or  
decision support. This is discussed further in “Configuring off-host  
processing” on page 369.  
You can use either the Veritas Enterprise Administrator (VEA) or the vxdg  
command to reorganize your disk groups. For more information about using the  
graphical user interface, see the Veritas Enterprise Administrator User’s Guide  
and VEA online help. This section describes how to use the vxdgcommand.  
The vxdgcommand provides the following operations for reorganizing disk  
groups:  
■
move—moves a self-contained set of VxVM objects between imported disk  
groups. This operation fails if it would remove all the disks from the source  
disk group. Volume states are preserved across the move. The move  
operation is illustrated in Figure 4-4.  
Figure 4-4  
Disk group move operation  
Source disk group  
Target disk group  
move  
Source disk group  
Target disk group  
After  
move  
■
split—removes a self-contained set of VxVM objects from an imported disk  
group, and moves them to a newly created target disk group. This operation  
fails if it would remove all the disks from the source disk group, or if an  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Creating and administering disk groups 197  
Reorganizing the contents of disk groups  
imported disk group exists with the same name as the target disk group. An  
existing deported disk group is destroyed if it has the same name as the  
target disk group (as is the case for the vxdginitcommand). The split  
operation is illustrated in Figure 4-5.  
Figure 4-5  
Disk group split operation  
Source disk group  
Disks to be split into new disk group  
Source disk group  
New target disk group  
After  
split  
■
join—removes all VxVM objects from an imported disk group and moves  
them to an imported target disk group. The source disk group is removed  
when the join is complete. The joinoperation is illustrated in Figure 4-6.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
198 Creating and administering disk groups  
Reorganizing the contents of disk groups  
Figure 4-6  
Disk group join operation  
Source disk group  
Target disk group  
join  
Target disk group  
After  
join  
These operations are performed on VxVM objects such as disks or top-level  
volumes, and include all component objects such as sub-volumes, plexes and  
subdisks. The objects to be moved must be self-contained, meaning that the disks  
that are moved must not contain any other objects that are not intended for the  
move.  
If you specify one or more disks to be moved, all VxVM objects on the disks are  
moved. You can use the -oexpandoption to ensure that vxdgmoves all disks on  
which the specified objects are configured. Care should be taken when doing this  
as the result may not always be what you expect. You can use the listmove  
operation with vxdgto help you establish what is the self-contained set of  
objects that corresponds to a specified set of objects.  
Caution: Before moving volumes between disk groups, stop all applications that  
are accessing the volumes, and unmount all file systems that are configured on  
these volumes.  
If the system crashes or a hardware subsystem fails, VxVM attempts to complete  
or reverse an incomplete disk group reconfiguration when the system is  
restarted or the hardware subsystem is repaired, depending on how far the  
reconfiguration had progressed. If one of the disk groups is no longer available  
because it has been imported by another host or because it no longer exists, you  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Creating and administering disk groups 199  
Reorganizing the contents of disk groups  
must recover the disk group manually as described in the section “Recovery  
from Incomplete Disk Group Moves” in the chapter “Recovery from Hardware  
Failure” of the Veritas Volume Manager Troubleshooting Guide.  
Limitations of disk group split and join  
The disk group split and join feature has the following limitations:  
■
Disk groups involved in a move, split or join must be version 90 or greater.  
See “Upgrading a disk group” on page 208 for more information on disk  
group versions.  
■
■
■
■
The reconfiguration must involve an integral number of physical disks.  
Objects to be moved must not contain open volumes.  
Disks cannot be moved between CDS and non-CDS compatible disk groups.  
Moved volumes are initially disabled following a disk group move, split or  
join. Use the vxrecover-mand vxvolstartallcommands to recover and  
restart the volumes.  
■
■
■
Data change objects (DCOs) and snap objects that have been dissociated by  
Persistent FastResync cannot be moved between disk groups.  
Veritas Volume Replicator (VVR) objects cannot be moved between disk  
groups.  
For a disk group move to succeed, the source disk group must contain at  
least one disk that can store copies of the configuration database after the  
move.  
■
■
For a disk group split to succeed, both the source and target disk groups  
must contain at least one disk that can store copies of the configuration  
database after the split.  
For a disk group move or join to succeed, the configuration database in the  
target disk group must be able to accommodate information about all the  
objects in the enlarged disk group.  
■
■
■
■
Splitting or moving a volume into a different disk group changes the  
volume’s record ID.  
The operation can only be performed on the master node of a cluster if  
either the source disk group or the target disk group is shared.  
In a cluster environment, disk groups involved in a move or join must both  
be private or must both be shared.  
When used with objects that have been created using the Veritas Intelligent  
Storage Provisioning (ISP) feature, only complete storage pools may be split  
or moved from a disk group. Individual objects such as application volumes  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
200 Creating and administering disk groups  
Reorganizing the contents of disk groups  
within storage pools may not be split or moved. See the Veritas Storage  
Foundation Intelligent Storage Provisioning Administrator’s Guide for a  
description of ISP and storage pools.  
■
If a cache object or volume set that is to be split or moved uses ISP volumes,  
the storage pool that contains these volumes must also be specified.  
The following sections describe how to use the vxdgcommand to reorganize disk  
groups. For more information about the vxdgcommand, see the vxdg(1M)  
manual page.  
Listing objects potentially affected by a move  
To display the VxVM objects that would be moved for a specified list of objects,  
use the following command:  
# vxdg [-o expand] listmove sourcedg targetdg object ...  
The following example lists the objects that would be affected by moving volume  
vol1from disk group mydgto newdg:  
# vxdg listmove mydg newdg vol1  
mydg01 c0t1d0 mydg05 c1t96d0 vol1 vol1-01 vol1-02 mydg01-01  
mydg05-01  
However, the following command produces an error because only a part of the  
volume vol1is configured on the disk mydg01:  
# vxdg listmove mydg newdg mydg01  
VxVM vxdg ERROR V-5-2-4597 vxdg listmove mydg newdg failed  
VxVM vxdg ERROR V-5-2-3091 mydg05 : Disk not moving, but  
subdisks on it are  
Specifying the -oexpandoption, as shown below, ensures that the list of objects  
to be moved includes the other disks (in this case, mydg05) that are configured  
in vol1:  
# vxdg -o expand listmove mydg newdg mydg01  
mydg01 c0t1d0 mydg05 c1t96d0 vol1 vol1-01 vol1-02 mydg01-01  
mydg05-01  
Moving DCO volumes between disk groups  
When you move the parent volume (such as a snapshot volume) to a different  
disk group, its DCO volume must accompany it. If you use the vxassistaddlog,  
vxmakeor vxdcocommands to set up a DCO for a volume, you must ensure that  
the disks that contain the plexes of the DCO volume accompany their parent  
volume during the move. You can use the vxprintcommand on a volume to  
examine the configuration of its associated DCO volume.  
If you use the vxassistcommand or the Veritas Enterprise Administrator (VEA)  
to create both a volume and its DCO, or the vxsnappreparecommand to add a  
DCO to a volume, the DCO plexes are automatically placed on different disks  
from the data plexes of the parent volume. In previous releases, version 0 DCO  
Download from Www.Somanuals.com. All Manuals Search And Download.  
               
Creating and administering disk groups 201  
Reorganizing the contents of disk groups  
plexes were placed on the same disks as the data plexes for convenience when  
performing disk group split and move operations. As version 20 DCOs support  
dirty region logging (DRL) in addition to Persistent FastResync, it is preferable  
for the DCO plexes to be separated from the data plexes. This improves the  
performance of I/O from/to the volume, and provides resilience for the DRL  
logs.  
Figure 4-7 illustrates some instances in which it is not be possible to split a disk  
group because of the location of the DCO plexes on the disks of the disk group.  
For more information about relocating DCO plexes, see “Specifying storage for  
version 0 DCO plexes” on page 357 and “Specifying storage for version 20 DCO  
plexes” on page 276.  
For more information about the layout of DCO volumes and their use with  
volume snapshots, see and “FastResync” on page 66. For more information  
about the administration of volume snapshots, see “Volume snapshots” on  
page 63 and “Administering volume snapshots” on page 303.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
202 Creating and administering disk groups  
Reorganizing the contents of disk groups  
Figure 4-7  
Examples of disk groups that can and cannot be split  
The disk group can be split as the DCO  
Volume  
data plexes  
Snapshot  
plexes are on dedicated disks, and can  
plex  
therefore accompany the disks that contain  
the volume data.  
Split  
Volume DCO  
plexes  
Snapshot  
DCO plex  
The disk group cannot be split as the DCO  
plexes cannot accompany their volumes.  
One solution is to relocate the DCO plexes.  
In this example, use an additional disk in the  
disk group as an intermediary to swap the  
misplaced DCO plexes. Alternatively, to  
improve DRL performance and resilience,  
allocate the DCO plexes to dedicated disks.  
Volume  
Snapshot  
plex  
data plexes  
X
Volume DCO  
plex  
Volume DCO  
plex  
Snapshot  
DCO plex  
The disk group can be split as the  
Snapshot  
Volume  
data plexes  
DCO plexes can accompany their  
volumes. However, you may not  
wish the data in the portions of  
the disks marked “?” to be moved  
as well.  
plex  
?
Split  
?
Snapshot  
DCO plex  
Volume DCO  
plexes  
The disk group cannot  
Snapshot  
Volume 1  
data plexes  
Volume 2  
data plexes  
be split as this would  
separate the disks  
containing Volume 2’s  
data plexes. Possible  
solutions are to  
plex  
relocate the snapshot  
DCO plex to the  
snapshot plex disk, or  
to another suitable  
disk that can be  
X
?
?
Snapshot  
DCO plex  
Volume 1 DCO  
plexes  
moved.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Creating and administering disk groups 203  
Reorganizing the contents of disk groups  
Moving objects between disk groups  
To move a self-contained set of VxVM objects from an imported source disk  
group to an imported target disk group, use the following command:  
# vxdg [-o expand] [-o override|verify] move sourcedg targetdg \  
object ...  
The -oexpandoption ensures that the objects that are actually moved include  
all other disks containing subdisks that are associated with the specified objects  
or with objects that they contain.  
The default behavior of vxdgwhen moving licensed disks in an EMC array is to  
perform an EMC disk compatibility check for each disk involved in the move. If  
the compatibility checks succeed, the move takes place. vxdgthen checks again  
to ensure that the configuration has not changed since it performed the  
compatibility check. If the configuration has changed, vxdgattempts to perform  
the entire move again.  
The -ooverrideoption enables the move to take place without any EMC  
checking.  
The -overifyoption returns the access names of the disks that would be moved  
but does not perform the move.  
Note: The -ooverrideand -overify optionsrequire a valid EMC license.  
See “Moving objects between disk groups” on page 424 for information on how  
to move objects between disk groups in a cluster.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
204 Creating and administering disk groups  
Reorganizing the contents of disk groups  
For example, the following output from vxprintshows the contents of disk  
groups rootdgand mydg:  
# vxprint  
Disk group: rootdg  
TY NAME  
ASSOC  
KSTATE  
LENGTH  
PLOFFS  
STATE TUTIL0  
PUTIL0  
dg rootdg  
rootdg  
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
dm rootdg02  
dm rootdg03  
dm rootdg04  
dm rootdg06  
c1t97d0  
c1t112d0  
c1t114d0  
c1t98d0  
17678493  
17678493 -  
17678493  
17678493  
-
-
Disk group: mydg  
TY NAME  
dg mydg  
ASSOC  
mydg  
KSTATE  
LENGTH  
-
17678493  
17678493  
17678493  
17678493  
PLOFFS  
STATE TUTIL0  
PUTIL0  
-
-
-
-
-
-
-
-
-
-
-
-
0
-
0
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
dm mydg01  
dm mydg05  
dm mydg07  
dm mydg08  
v vol1  
pl vol1-01  
sd mydg01-01  
pl vol1-02  
sd mydg05-01  
c0t1d0  
c1t96d0  
c1t99d0  
c1t100d0  
fsgen  
vol1  
vol1-01  
vol1  
ENABLED 2048  
ENABLED 3591  
ENABLED 3591  
ENABLED 3591  
ENABLED 3591  
ACTIVE -  
ACTIVE -  
-
-
ACTIVE -  
vol1-02  
-
-
The following command moves the self-contained set of objects implied by  
specifying disk mydg01from disk group mydgto rootdg:  
# vxdg -o expand move mydg rootdg mydg01  
The moved volumes are initially disabled following the move. Use the following  
commands to recover and restart the volumes in the target disk group:  
# vxrecover -g targetdg -m [volume ...]  
# vxvol -g targetdg startall  
The output from vxprintafter the move shows that not only mydg01but also  
volume vol1and mydg05have moved to rootdg, leaving only mydg07and  
mydg08in disk group mydg:  
# vxprint  
Disk group: rootdg  
TY NAME  
dg rootdg  
dm mydg01  
ASSOC  
KSTATE  
LENGTH  
-
17678493  
17678493  
PLOFFS  
STATE TUTIL0  
PUTIL0  
rootdg  
c0t1d0  
c1t97d0  
c1t112d0  
c1t114d0  
c1t96d0  
c1t98d0  
fsgen  
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
dm rootdg02  
dm rootdg03  
dm rootdg04  
dm mydg05  
dm rootdg06  
v vol1  
pl vol1-01  
sd mydg01-01  
pl vol1-02  
sd mydg05-01  
17678493 -  
17678493  
17678493  
17678493  
-
-
-
-
-
0
-
0
ENABLED 2048  
ENABLED 3591  
ENABLED 3591  
ENABLED 3591  
ENABLED 3591  
ACTIVE -  
ACTIVE -  
vol1  
vol1-01  
vol1  
-
-
ACTIVE -  
vol1-02  
-
-
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Creating and administering disk groups 205  
Reorganizing the contents of disk groups  
Disk group: mydg  
TY NAME  
dg mydg  
dm mydg07  
dm mydg08  
ASSOC  
mydg  
c1t99d0  
c1t100d0  
KSTATE  
LENGTH  
-
17678493  
17678493  
PLOFFS  
STATE TUTIL0  
PUTIL0  
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
The following commands would also achieve the same result:  
# vxdg move mydg rootdg mydg01 mydg05  
# vxdg move mydg rootdg vol1  
Splitting disk groups  
To remove a self-contained set of VxVM objects from an imported source disk  
group to a new target disk group, use the following command:  
# vxdg [-o expand] [-o override|verify] split sourcedg targetdg \  
object ...  
For a description of the -oexpand, -ooverride, and -overifyoptions, see  
“Moving objects between disk groups” on page 203.  
See “Splitting disk groups” on page 424 for more information on splitting  
shared disk groups in clusters.  
For example, the following output from vxprintshows the contents of disk  
group rootdg:  
# vxprint  
Disk group: rootdg  
TY NAME  
dg rootdg  
ASSOC  
rootdg  
c0t1d0  
KSTATE  
LENGTH  
-
17678493  
17678493  
PLOFFS  
STATE TUTIL0  
PUTIL0  
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
dm rootdg01  
dm rootdg02  
dm rootdg03  
dm rootdg04  
dm rootdg05  
dm rootdg06  
dm rootdg07  
dm rootdg08  
v vol1  
pl vol1-01  
sd rootdg01-01  
pl vol1-02  
sd rootdg05-01  
c1t97d0  
c1t112d0  
c1t114d0  
c1t96d0  
c1t98d0  
c1t99d0  
c1t100d0  
fsgen  
17678493 -  
17678493  
17678493  
17678493  
17678493  
17678493  
-
-
-
-
-
-
-
0
-
0
ENABLED 2048  
ENABLED 3591  
ACTIVE -  
ACTIVE -  
vol1  
vol1-01 ENABLED 3591  
vol1  
-
-
ENABLED 3591  
ACTIVE -  
vol1-02 ENABLED 3591  
-
-
The following command removes disks rootdg07and rootdg08from rootdg  
to form a new disk group, mydg:  
# vxdg -o expand split rootdg mydg rootdg07 rootdg08  
The moved volumes are initially disabled following the split. Use the following  
commands to recover and restart the volumes in the new target disk group:  
# vxrecover -g targetdg -m [volume ...]  
# vxvol -g targetdg startall  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
206 Creating and administering disk groups  
Reorganizing the contents of disk groups  
The output from vxprintafter the split shows the new disk group, mydg:  
# vxprint  
Disk group: rootdg  
TY NAME  
dg rootdg  
ASSOC  
KSTATE  
LENGTH  
-
17678493  
17678493  
PLOFFS  
STATE TUTIL0  
PUTIL0  
rootdg  
c0t1d0  
c1t97d0  
c1t112d0  
c1t114d0  
c1t96d0  
c1t98d0  
fsgen  
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
dm rootdg01  
dm rootdg02  
dm rootdg03  
dm rootdg04  
dm rootdg05  
dm rootdg06  
v vol1  
pl vol1-01  
sd rootdg01-01  
pl vol1-02  
sd rootdg05-01  
17678493 -  
17678493  
17678493  
17678493  
-
-
-
-
-
0
-
0
ENABLED 2048  
ENABLED 3591  
ACTIVE -  
ACTIVE -  
vol1  
vol1-01 ENABLED 3591  
vol1  
-
-
ENABLED 3591  
ACTIVE -  
vol1-02 ENABLED 3591  
-
-
Disk group: mydg  
TY NAME  
ASSOC  
KSTATE  
LENGTH  
PLOFFS  
STATE TUTIL0  
PUTIL0  
dg mydg  
dm rootdg07  
dm rootdg08  
mydg  
c1t99d0  
c1t100d0  
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
17678493  
17678493  
Joining disk groups  
To remove all VxVM objects from an imported source disk group to an imported  
target disk group, use the following command:  
# vxdg [-o override|verify] join sourcedg targetdg  
Note: You cannot specify rootdgas the source disk group for a joinoperation.  
For a description of the -ooverrideand -overifyoptions, see “Moving objects  
between disk groups” on page 203.  
See “Joining disk groups” on page 424 for information on joining disk groups in  
a cluster.  
For example, the following output from vxprintshows the contents of the disk  
group rootdgand mydg:  
# vxprint  
Disk group: rootdg  
TY NAME  
dg rootdg  
ASSOC  
rootdg  
c0t1d0  
c1t97d0  
c1t112d0  
c1t114d0  
c1t99d0  
c1t100d0  
KSTATE  
LENGTH  
-
17678493  
17678493  
PLOFFS  
STATE TUTIL0  
PUTIL0  
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
dm rootdg01  
dm rootdg02  
dm rootdg03  
dm rootdg04  
dm rootdg07  
dm rootdg08  
17678493 -  
17678493  
17678493  
17678493  
-
-
-
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Creating and administering disk groups 207  
Disabling a disk group  
Disk group: mydg  
TY NAME  
dg mydg  
ASSOC  
mydg  
c1t96d0  
c1t98d0  
fsgen  
vol1  
vol1-01  
vol1  
KSTATE  
LENGTH  
-
17678493  
17678493  
PLOFFS  
STATE TUTIL0  
PUTIL0  
-
-
-
-
-
-
-
-
0
-
0
-
-
-
-
-
-
-
-
-
-
-
-
-
-
dm mydg05  
dm mydg06  
v vol1  
pl vol1-01  
sd mydg01-01  
pl vol1-02  
sd mydg05-01  
ENABLED 2048  
ENABLED 3591  
ENABLED 3591  
ENABLED 3591  
ENABLED 3591  
ACTIVE -  
ACTIVE -  
-
-
ACTIVE -  
vol1-02  
-
-
The following command joins disk group mydgto rootdg:  
# vxdg join mydg rootdg  
The moved volumes are initially disabled following the join. Use the following  
commands to recover and restart the volumes in the target disk group:  
# vxrecover -g targetdg -m [volume ...]  
# vxvol -g targetdg startall  
The output from vxprintafter the join shows that disk group mydghas been  
removed:  
# vxprint  
Disk group: rootdg  
TY NAME  
dg rootdg  
dm mydg01  
ASSOC  
KSTATE  
LENGTH  
-
17678493  
17678493  
PLOFFS  
STATE TUTIL0  
PUTIL0  
rootdg  
c0t1d0  
c1t97d0  
c1t112d0  
c1t114d0  
c1t96d0  
c1t98d0  
c1t99d0  
c1t100d0  
fsgen  
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
dm rootdg02  
dm rootdg03  
dm rootdg04  
dm mydg05  
dm rootdg06  
dm rootdg07  
dm rootdg08  
v vol1  
pl vol1-01  
sd mydg01-01  
pl vol1-02  
sd mydg05-01  
17678493 -  
17678493  
17678493  
17678493  
17678493  
17678493  
-
-
-
-
-
-
-
0
-
0
ENABLED 2048  
ENABLED 3591  
ENABLED 3591  
ENABLED 3591  
ENABLED 3591  
ACTIVE -  
ACTIVE -  
vol1  
vol1-01  
vol1  
-
-
ACTIVE -  
vol1-02  
-
-
Disabling a disk group  
To disable a disk group, unmount and stop any volumes in the disk group, and  
then use the following command to deport it:  
# vxdg deport diskgroup  
Deporting a disk group does not actually remove the disk group. It disables use  
of the disk group by the system. Disks in a deported disk group can be reused,  
reinitialized, added to other disk groups, or imported for use on other systems.  
Use the vxdgimportcommand to re-enable access to the disk group.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
208 Creating and administering disk groups  
Destroying a disk group  
Destroying a disk group  
The vxdgcommand provides a destroyoption that removes a disk group from  
the system and frees the disks in that disk group for reinitialization:  
# vxdg destroy diskgroup  
Caution: This command destroys all data on the disks.  
When a disk group is destroyed, the disks that are released can be re-used in  
other disk groups.  
Recovering a destroyed disk group  
If a disk group has been accidentally destroyed, you can recover it, provided that  
the disks that were in the disk group have not been modified or reused  
elsewhere.  
To recover a destroyed disk group  
1
Enter the following command to find out the disk group ID (dgid) of one of  
the disks that was in the disk group:  
# vxdisk -s list disk_access_name  
The disk must be specified by its disk access name, such as c0t12d0.  
Examine the output from the command for a line similar to the following  
that specifies the disk group ID.  
dgid:  
963504895.1075.bass  
2
Use the disk group ID to import the disk group:  
# vxdg import dgid  
Upgrading a disk group  
Note: On some platforms, the first release of Veritas Volume Manager was 3.0 or  
3.2.  
Prior to the release of Veritas Volume Manager 3.0, the disk group version was  
automatically upgraded (if needed) when the disk group was imported.  
From release 3.0 of Veritas Volume Manager, the two operations of importing a  
disk group and upgrading its version are separate. You can import a disk group  
from a previous version and use it without upgrading it.  
When you want to use new features, the disk group can be upgraded. The  
upgrade is an explicit operation. Once the upgrade occurs, the disk group  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                     
Creating and administering disk groups 209  
Upgrading a disk group  
becomes incompatible with earlier releases of VxVM that do not support the  
new version.  
Before the imported disk group is upgraded, no changes are made to the disk  
group to prevent its use on the release from which it was imported until you  
explicitly upgrade it to the current release.  
Until completion of the upgrade, the disk group can be used “as is” provided  
there is no attempt to use the features of the current version. Attempts to use a  
feature of the current version that is not a feature of the version from which the  
disk group was imported results in an error message similar to this:  
VxVM vxedit ERROR V-5-1-2829 Disk group version doesn’t support  
feature  
To use any of the new features, you must run the vxdg upgradecommand to  
explicitly upgrade the disk group to a version that supports those features.  
All disk groups have a version number associated with them. Veritas Volume  
Manager releases support a specific set of disk group versions. VxVM can import  
and perform operations on a disk group of that version. The operations are  
limited by what features and operations the disk group version supports.  
The table, “Disk group version assignments,” summarizes the Veritas Volume  
Manager releases that introduce and support specific disk group versions:  
Table 4-1  
Disk group version assignments  
VxVM release  
Introduces disk group version Supports disk group versions  
1.2  
10  
10  
1.3  
15  
15  
2.0  
20  
20  
2.2  
30  
30  
2.3  
40  
40  
2.5  
50  
50  
3.0  
60  
20-40, 60  
20-70  
20-80  
20-90  
20-110  
20-120  
20-140  
3.1  
70  
3.1.1  
3.2, 3.5  
4.0  
80  
90  
110  
120  
140  
4.1  
5.0  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
210 Creating and administering disk groups  
Upgrading a disk group  
Importing the disk group of a previous version on a Veritas Volume Manager  
system prevents the use of features introduced since that version was released.  
The table, “Features supported by disk group versions,” summarizes the  
features that are supported by disk group versions 20 through 140:  
Table 4-2  
Features supported by disk group versions  
Disk  
New features supported  
Previous version  
group  
version  
features supported  
140  
■
■
■
■
Data Migration Features  
DMP Enhancements  
20, 30, 40, 50, 60, 70,  
80, 90, 110, 120, 130  
Import of Cloned Disks  
Intelligent Storage Provisioning (ISP)  
Enhancements  
■
■
Remote Mirror (Campus Cluster)  
Veritas Volume Replicator (VVR)  
Enhancements  
130  
120  
■
VVR Enhancements  
20, 30, 40, 50, 60, 70,  
80, 90, 110, 120  
■
■
■
■
Automatic Cluster-wide Failback for A/P arrays 20, 30, 40, 50, 60, 70,  
80, 90, 110  
Migration of Volumes to ISP  
Persistent DMP Policies  
Shared Disk Group Failure Policy  
110  
■
■
■
■
■
Cross-platform Data Sharing (CDS)  
Device Discovery Layer (DDL) 2.0  
Disk Group Configuration Backup and Restore  
Elimination of rootdgas a Special Disk Group  
Full-Sized and Space-Optimized Instant  
Snapshots  
20, 30, 40, 50, 60, 70,  
80, 90  
■
■
■
Intelligent Storage Provisioning (ISP)  
Serial Split Brain Detection  
Volume Sets (Multiple Device Support for VxFS)  
90  
■
■
■
■
■
■
■
Cluster Support for Oracle Resilvering  
Disk Group Move, Split and Join  
Device Discovery Layer (DDL) 1.0  
Layered Volume Support in Clusters  
Ordered Allocation  
20, 30, 40, 50, 60, 70,  
80  
OS Independent Naming Support  
Persistent FastResync  
80  
■
VVR Enhancements  
20, 30, 40, 50, 60, 70  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Creating and administering disk groups 211  
Upgrading a disk group  
Table 4-2  
Features supported by disk group versions  
New features supported  
Disk  
Previous version  
group  
version  
features supported  
70  
■
■
■
■
Non-Persistent FastResync  
Sequential DRL  
20, 30, 40, 50, 60  
Unrelocate  
VVR Enhancements  
60  
50  
■
■
Online Relayout  
20, 30, 40  
20, 30, 40  
Safe RAID-5 Subdisk Moves  
■
SRVM (now known as Veritas Volume  
Replicator or VVR)  
40  
30  
20  
■
■
Hot-Relocation  
20, 30  
20  
VxSmartSync Recovery Accelerator  
■
■
■
■
■
■
Dirty Region Logging (DRL)  
Disk Group Configuration Copy Limiting  
Mirrored Volumes Logging  
New-Style Stripes  
RAID-5 Volumes  
Recovery Checkpointing  
To list the version of a disk group, use this command:  
# vxdg list dgname  
You can also determine the disk group version by using the vxprintcommand  
with the -lformat option.  
To upgrade a disk group to the highest version supported by the release of  
VxVM that is currently running, use this command:  
# vxdg upgrade dgname  
By default, VxVM creates a disk group of the highest version supported by the  
release. For example, Veritas Volume Manager 5.0 creates disk groups with  
version 140.  
It may sometimes be necessary to create a disk group for an older version. The  
default disk group version for a disk group created on a system running Veritas  
Volume Manager 5.0 is 140. Such a disk group cannot be imported on a system  
running Veritas Volume Manager 4.1, as that release only supports up to version  
120. Therefore, to create a disk group on a system running Veritas Volume  
Manager 5.0 that can be imported by a system running Veritas Volume Manager  
4.1, the disk group must be created with a version of 120 or less.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
212 Creating and administering disk groups  
Managing the configuration daemon in VxVM  
To create a disk group with a previous version, specify the -Tversion option to  
the vxdginitcommand.  
For example, to create a disk group with version 120 that can be imported by a  
system running VxVM 4.1, use the following command:  
# vxdg -T 120 init newdg newdg01=c0t3d0  
This creates a disk group, newdg, which can be imported by Veritas Volume  
Manager 4.1. Note that while this disk group can be imported on the VxVM 4.1  
system, attempts to use features from Veritas Volume Manager 5.0 will fail.  
Managing the configuration daemon in VxVM  
The VxVM configuration daemon (vxconfigd) provides the interface between  
VxVM commands and the kernel device drivers. vxconfigdhandles  
configuration change requests from VxVM utilities, communicates the change  
requests to the VxVM kernel, and modifies configuration information stored on  
disk. vxconfigdalso initializes VxVM when the system is booted.  
The vxdctlcommand is the command-line interface to the vxconfigddaemon.  
You can use vxdctlto:  
■
■
Control the operation of the vxconfigddaemon.  
Change the system-wide definition of the default disk group.  
Note: In VxVM 4.0 and later releases, disk access records are no longer stored in  
the /etc/vx/volbootfile. Non-persistent disk access records are created by  
scanning the disks at system startup. Persistent disk access records for simple  
and noprivdisks are permanently stored in the /etc/vx/darecsfile in the  
rootfile system. The vxconfigddaemon reads the contents of this file to locate  
the disks and the configuration databases for their disk groups. (The /etc/vx/  
darecsfile is also used to store definitions of foreign devices that are not  
autoconfigurable. Such entries may be added by using the vxddladmaddforeign  
command. See the vxddladm(1M) manual page for more information.)  
You can also use vxdctlto:  
■
■
■
Reconfigure the DMP database to include disk devices newly attached to, or  
removed from the system.  
Create DMP device nodes in the directories /dev/vx/dmpand /dev/vx/  
rdmp.  
Update the DMP database with changes in path type for active/passive disk  
arrays. Use the utilities provided by the disk-array vendor to change the  
path type between primary and secondary.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Creating and administering disk groups 213  
Backing up and restoring disk group configuration data  
For more information about how to use vxdctl, refer to the vxdctl(1M) manual  
page.  
Backing up and restoring disk group configuration  
data  
The disk group configuration backup and restoration feature allows you to back  
up and restore all configuration data for disk groups, and for VxVM objects such  
as volumes that are configured within the disk groups. The vxconfigbackupd  
daemon monitors changes to the VxVM configuration and automatically records  
any configuration changes that occur. Two utilities, vxconfigbackupand  
vxconfigrestore, are provided for backing up and restoring a VxVM  
configuration for a disk group.  
For information on backing up and restoring disk group configurations, see the  
“Backing Up and Restoring Disk Group Configurations” chapter in the Veritas  
Volume Manager Troubleshooting Guide, and the vxconfigbackup(1M) and  
vxconfigrestore(1M) manual pages.  
Using vxnotify to monitor configuration changes  
You can use the vxnotifyutility to display events relating to disk and  
configuration changes that are managed by the vxconfigdconfiguration  
daemon. If vxnotifyis running on a system where the VxVM clustering feature  
is active, it displays events that are related to changes in the cluster state of the  
system on which it is running. The vxnotifyutility displays the requested event  
types until you kill it, until it has received a specified number of events, or until  
a specified period of time has elapsed.  
Examples of configuration events that can be detected include disabling and  
enabling of controllers, paths and DMP nodes, RAID-5 volumes entering  
degraded mode, detachment of disks, plexes and volumes, and nodes joining and  
leaving a cluster.  
For example, the following vxnotifycommand displays information about all  
disk, plex, and volume detachments as they occur:  
# vxnotify -f  
The following command provides information about cluster configuration  
changes, including the import and deport of shared disk groups:  
# vxnotify -s -i  
For more information about the vxnotifyutility, and the types of configuration  
events that it can report, see the vxnotify(1M) manual page.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
214 Creating and administering disk groups  
Using vxnotify to monitor configuration changes  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter  
5
Creating and  
administering subdisks  
This chapter describes how to create and maintain subdisks. Subdisks are the  
low-level building blocks in a Veritas Volume Mananger (VxVM) configuration  
that are required to create plexes and volumes.  
Note: Most VxVM commands require superuser or equivalent privileges.  
Creating subdisks  
Note: Subdisks are created automatically if you use the vxassistcommand or  
the Veritas Enterprise Administrator (VEA) to create volumes. For more  
information, see “Creating a volume” on page 238.  
Use the vxmakecommand to create VxVM objects, such as subdisks:  
# vxmake [-g diskgroup] sd subdisk diskname,offset,length  
where: subdisk is the name of the subdisk, diskname is the disk name, offset is  
the starting point (offset) of the subdisk within the disk, and length is the length  
of the subdisk.  
For example, to create a subdisk named mydg02-01in the disk group, mydg,  
that starts at the beginning of disk mydg02and has a length of 8000 sectors, use  
the following command:  
# vxmake -g mydg sd mydg02-01 mydg02,0,8000  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
216 Creating and administering subdisks  
Displaying subdisk information  
Note: As for all VxVM commands, the default size unit is s, representing a  
sector. Add a suffix, such as kfor kilobyte, mfor megabyte or gfor gigabyte, to  
change the unit of size. For example, 500mwould represent 500 megabytes.  
If you intend to use the new subdisk to build a volume, you must associate the  
subdisk with a plex (see “Associating subdisks with plexes” on page 218).  
Subdisks for all plex layouts (concatenated, striped, RAID-5) are created the  
same way.  
Displaying subdisk information  
The vxprintcommand displays information about VxVM objects. To display  
general information for all subdisks, use this command:  
# vxprint -st  
The -soption specifies information about subdisks. The -toption prints a  
single-line output record that depends on the type of object being listed.  
The following is example output:  
SD NAME  
SV NAME  
PLEX  
PLEX  
DISK  
DISKOFFS LENGTH [COL/]OFF DEVICE  
MODE  
MODE  
VOLNAME NVOLLAYR LENGTH [COL/]OFF AM/NM  
sd mydg01-01 vol1-01 mydg01 0  
sd mydg02-01 vol2-01 mydg02 0  
102400 0  
102400 0  
c2t0d1 ENA  
c2t1d1 ENA  
You can display complete information about a particular subdisk by using this  
command:  
# vxprint [-g diskgroup] -l subdisk  
For example, the following command displays all information for subdisk  
mydg02-01in the disk group, mydg:  
# vxprint -g mydg -l mydg02-01  
This command provides the following output:  
Disk group: mydg  
Subdisk: mydg02-01  
info:  
disk=mydg02 offset=0 len=205632  
vol=mvol plex=mvol-02 (offset=0)  
enabled  
assoc:  
flags:  
device:  
device=c2t1d1 path=/dev/vx/dmp/c2t1d1 diskdev=32/68  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Creating and administering subdisks 217  
Moving subdisks  
Moving subdisks  
Moving a subdisk copies the disk space contents of a subdisk onto one or more  
other subdisks. If the subdisk being moved is associated with a plex, then the  
data stored on the original subdisk is copied to the new subdisks. The old  
subdisk is dissociated from the plex, and the new subdisks are associated with  
the plex. The association is at the same offset within the plex as the source  
subdisk. To move a subdisk, use the following command:  
# vxsd [-g diskgroup] mv old_subdisk new_subdisk [new_subdisk ...]  
For example, if mydg03in the disk group, mydg, is to be evacuated, and mydg12  
has enough room on two of its subdisks, use the following command:  
# vxsd -g mydg mv mydg03-01 mydg12-01 mydg12-02  
For the subdisk move to work correctly, the following conditions must be met:  
■
■
The subdisks involved must be the same size.  
The subdisk being moved must be part of an active plex on an active  
(ENABLED) volume.  
■
The new subdisk must not be associated with any other plex.  
See “Configuring hot-relocation to use only spare disks” on page 390 for  
information about manually relocating subdisks after hot-relocation.  
Splitting subdisks  
Splitting a subdisk divides an existing subdisk into two separate subdisks. To  
split a subdisk, use the following command:  
# vxsd [-g diskgroup] –s size split subdisk newsd1 newsd2  
where subdisk is the name of the original subdisk, newsd1 is the name of the  
first of the two subdisks to be created and newsd2 is the name of the second  
subdisk to be created.  
The -soption is required to specify the size of the first of the two subdisks to be  
created. The second subdisk occupies the remaining space used by the original  
subdisk.  
If the original subdisk is associated with a plex before the task, upon completion  
of the split, both of the resulting subdisks are associated with the same plex.  
To split the original subdisk into more than two subdisks, repeat the previous  
command as many times as necessary on the resulting subdisks.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                 
218 Creating and administering subdisks  
Joining subdisks  
For example, to split subdisk mydg03-02, with size 2000 megabytes into  
subdisks mydg03-02, mydg03-03, mydg03-04and mydg03-05, each with size  
500 megabytes, all in the disk group, mydg, use the following commands:  
# vxsd -g mydg -s 1000m split mydg03-02 mydg03-02 mydg03-04  
# vxsd -g mydg -s 500m split mydg03-02 mydg03-02 mydg03-03  
# vxsd -g mydg -s 500m split mydg03-04 mydg03-04 mydg03-05  
Joining subdisks  
Joining subdisks combines two or more existing subdisks into one subdisk. To  
join subdisks, the subdisks must be contiguous on the same disk. If the selected  
subdisks are associated, they must be associated with the same plex, and be  
contiguous in that plex. To join several subdisks, use the following command:  
# vxsd [-g diskgroup] join subdisk1 subdisk2 ... new_subdisk  
For example, to join the contiguous subdisks mydg03-02, mydg03-03, mydg03-  
04and mydg03-05as subdisk mydg03-02in the disk group, mydg, use the  
following command:  
# vxsd -g mydg join mydg03-02 mydg03-03 mydg03-04 mydg03-05 \  
mydg03-02  
Associating subdisks with plexes  
Associating a subdisk with a plex places the amount of disk space defined by the  
subdisk at a specific offset within the plex. The entire area that the subdisk fills  
must not be occupied by any portion of another subdisk. There are several ways  
that subdisks can be associated with plexes, depending on the overall state of  
the configuration.  
If you have already created all the subdisks needed for a particular plex, to  
associate subdisks at plex creation, use the following command:  
# vxmake [-g diskgroup] plex plex sd=subdisk,...  
For example, to create the plex home-1and associate subdisks mydg02-01,  
mydg02-00, and mydg02-02with plex home-1, all in the disk group, mydg, use  
the following command:  
# vxmake -g mydg plex home-1 sd=mydg02-01,mydg02-00,mydg02-02  
Subdisks are associated in order starting at offset 0. If you use this type of  
command, you do not have to specify the multiple commands needed to create  
the plex and then associate each of the subdisks with that plex. In this example,  
the subdisks are associated to the plex in the order they are listed (after sd=).  
The disk space defined as mydg02-01is first, mydg02-00is second, and  
mydg02-02is third. This method of associating subdisks is convenient during  
initial configuration.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
               
Creating and administering subdisks 219  
Associating subdisks with plexes  
Subdisks can also be associated with a plex that already exists. To associate one  
or more subdisks with an existing plex, use the following command:  
# vxsd [-g diskgroup] assoc plex subdisk1 [subdisk2 subdisk3 ...]  
For example, to associate subdisks named mydg02-01, mydg02-00, and  
mydg02-02with a plex named home-1, use the following command:  
# vxsd -g mydg assoc home-1 mydg02-01 mydg02-00 mydg02-01  
If the plex is not empty, the new subdisks are added after any subdisks that are  
already associated with the plex, unless the -loption is specified with the  
command. The -loption associates subdisks at a specific offset within the plex.  
The -loption is required if you previously created a sparse plex (that is, a plex  
with portions of its address space that do not map to subdisks) for a particular  
volume, and subsequently want to make the plex complete. To complete the  
plex, create a subdisk of a size that fits the hole in the sparse plex exactly. Then,  
associate the subdisk with the plex by specifying the offset of the beginning of  
the hole in the plex, using the following command:  
# vxsd [-g diskgroup] -l offset assocsparse_plex exact_size_subdisk  
For example, the following command would insert the subdisk, mydg15-01, in  
the plex, vol10-01, starting at an offset of 4096 blocks:  
# vxsd -g mydg -l 4096b assoc vol10-01 mydg15-01  
Note: The subdisk must be exactly the right size. VxVM does not allow the space  
defined for two subdisks to overlap within a plex.  
For striped or RAID-5 plexes, use the following command to specify a column  
number and column offset for the subdisk to be added:  
# vxsd [-g diskgroup] -l column_#/offset assoc plex subdisk ...  
If only one number is specified with the -loption for striped plexes, the number  
is interpreted as a column number and the subdisk is associated at the end of the  
column.  
For example, the following command would add the subdisk, mydg11-01, to the  
end of column 1 of the plex, vol02-01:  
# vxsd -g mydg -l 1 assoc vol02-01 mydg11-01  
Alternatively, to add M subdisks at the end of each of the N columns in a striped  
or RAID-5 volume, you can use the following form of the vxsdcommand:  
# vxsd [-g diskgroup] assoc plex subdisk1:0 ... subdiskM:N-1  
The following example shows how to append three subdisk to the ends of the  
three columns in a striped plex, vol-01, in the disk group, mydg:  
# vxsd -g mydg assoc vol01-01 mydg10-01:0 mydg11-01:1 \  
mydg12-01:2  
If a subdisk is filling a “hole” in the plex (that is, some portion of the volume  
logical address space is mapped by the subdisk), the subdisk is considered stale.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
220 Creating and administering subdisks  
Associating log subdisks  
If the volume is enabled, the association operation regenerates data that belongs  
on the subdisk. Otherwise, it is marked as stale and is recovered when the  
volume is started.  
Associating log subdisks  
Note: The version 20 DCO volume layout includes space for a DRL. Do not apply  
the procedure described in this section to a volume that has a version 20 DCO  
volume associated with it. See “Preparing a volume for DRL and instant  
snapshots” on page 275 for more information.  
Log subdisks are defined and added to a plex that is to become part of a volume  
on which dirty region logging (DRL) is enabled. DRL is enabled for a volume  
when the volume is mirrored and has at least one log subdisk.  
For a description of DRL, see “Dirty region logging” on page 60. Log subdisks are  
ignored as far as the usual plex policies are concerned, and are only used to hold  
the dirty region log.  
Note: Only one log subdisk can be associated with a plex. Because this log  
subdisk is frequently written, care should be taken to position it on a disk that is  
not heavily used. Placing a log subdisk on a heavily-used disk can degrade  
system performance.  
To add a log subdisk to an existing plex, use the following command:  
# vxsd [-g diskgroup] aslog plex subdisk  
where subdisk is the name to be used for the log subdisk. The plex must be  
associated with a mirrored volume before dirty region logging takes effect.  
For example, to associate a subdisk named mydg02-01with a plex named  
vol01-02,which is already associated with volume vol01in the disk group,  
mydg, use the following command:  
# vxsd -g mydg aslog vol01-02 mydg02-01  
You can also add a log subdisk to an existing volume with the following  
command:  
# vxassist [-g diskgroup] addlog volume disk  
This command automatically creates a log subdisk within a log plex on the  
specified disk for the specified volume.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Creating and administering subdisks 221  
Dissociating subdisks from plexes  
Dissociating subdisks from plexes  
To break an established connection between a subdisk and the plex to which it  
belongs, the subdisk is dissociated from the plex. A subdisk is dissociated when  
the subdisk is removed or used in another plex. To dissociate a subdisk, use the  
following command:  
# vxsd [-g diskgroup] [-o force] dis subdisk  
For example, to dissociate a subdisk named mydg02-01from the plex with  
which it is currently associated in the disk group, mydg, use the following  
command:  
# vxsd -g mydg dis mydg02-01  
You can additionally remove the dissociated subdisks from VxVM control using  
the following form of the command:  
# vxsd [-g diskgroup] -o rm dis subdisk  
Caution: If the subdisk maps a portion of a volume’s address space, dissociating  
it places the volume in DEGRADED mode. In this case, the disoperation prints a  
warning and must be forced using the -oforceoption to succeed. Also, if  
removing the subdisk makes the volume unusable, because another subdisk in  
the same stripe is unusable or missing and the volume is not DISABLED and  
empty, the operation is not allowed.  
Removing subdisks  
To remove a subdisk, use the following command:  
# vxedit [-g diskgroup] rm subdisk  
For example, to remove a subdisk named mydg02-01from the disk group, mydg,  
use the following command:  
# vxedit -g mydg rm mydg02-01  
Changing subdisk attributes  
Caution: Change subdisk attributes with extreme care.  
The vxeditcommand changes attributes of subdisks and other VxVM objects.  
To change subdisk attributes, use the following command:  
# vxedit [-g diskgroup] set attribute=value ... subdisk ...  
Subdisk fields that can be changed using the vxeditcommand include:  
■ name  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                         
222 Creating and administering subdisks  
Changing subdisk attributes  
■ putiln  
■ tutiln  
■ len  
■ comment  
The putiln field attributes are maintained on reboot; tutiln fields are  
temporary and are not retained on reboot. VxVM sets the putil0and tutil0  
utility fields. Other Symantec products, such as the Veritas Enterprise  
Administrator (VEA), set the putil1and tutil1fields. The putil2and  
tutil2are available for you to use for site-specific purposes. The length field,  
len, can only be changed if the subdisk is dissociated.  
For example, to change the comment field of a subdisk named mydg02-01in the  
disk group, mydg, use the following command:  
# vxedit -g mydg set comment=“subdisk comment” mydg02-01  
To prevent a particular subdisk from being associated with a plex, set the  
putil0field to a non-null string, as shown in the following command:  
# vxedit -g mydg set putil0=”DO-NOT-USE” mydg02-01  
See the vxedit(1M) manual page for more information about using the vxedit  
command to change the attribute fields of VxVM objects.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Chapter  
6
Creating and  
administering plexes  
This chapter describes how to create and maintain plexes. Plexes are logical  
groupings of subdisks that create an area of disk space independent of physical  
disk size or other restrictions. Replication (mirroring) of disk data is set up by  
creating multiple data plexes for a single volume. Each data plex in a mirrored  
volume contains an identical copy of the volume data. Because each data plex  
must reside on different disks from the other plexes, the replication provided by  
mirroring prevents data loss in the event of a single-point disk-subsystem  
failure. Multiple data plexes also provide increased data integrity and reliability.  
Note: Most VxVM commands require superuser or equivalent privileges.  
Creating plexes  
Note: Plexes are created automatically if you use the vxassistcommand or the  
Veritas Enterprise Administrator (VEA) to create volumes. For more  
information, see “Creating a volume” on page 238.  
Use the vxmakecommand to create VxVM objects, such as plexes. When creating  
a plex, identify the subdisks that are to be associated with it:  
To create a plex from existing subdisks, use the following command:  
# vxmake [-g diskgroup] plex plex sd=subdisk1[,subdisk2,...]  
For example, to create a concatenated plex named vol01-02from two existing  
subdisks named mydg02-01and mydg02-02in the disk group, mydg, use the  
following command:  
# vxmake -g mydg plex vol01-02 sd=mydg02-01,mydg02-02  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
224 Creating and administering plexes  
Creating a striped plex  
Creating a striped plex  
To create a striped plex, you must specify additional attributes. For example, to  
create a striped plex named pl-01in the disk group, mydg, with a stripe width  
of 32 sectors and 2 columns, use the following command:  
# vxmake -g mydg plex pl-01 layout=stripe stwidth=32 ncolumn=2 \  
sd=mydg01-01,mydg02-01  
To use a plex to build a volume, you must associate the plex with the volume. For  
more information, see the section, “Attaching and associating plexes” on  
page 229.  
Displaying plex information  
Listing plexes helps identify free plexes for building volumes. Use the plex (–p)  
option to the vxprintcommand to list information about all plexes.  
To display detailed information about all plexes in the system, use the following  
command:  
# vxprint -lp  
To display detailed information about a specific plex, use the following  
command:  
# vxprint [-g diskgroup] -l plex  
The -toption prints a single line of information about the plex. To list free  
plexes, use the following command:  
# vxprint -pt  
The following section describes the meaning of the various plex states that may  
be displayed in the STATE field of vxprintoutput.  
Plex states  
Plex states reflect whether or not plexes are complete and are consistent copies  
(mirrors) of the volume contents. VxVM utilities automatically maintain the  
plex state. However, if a volume should not be written to because there are  
changes to that volume and if a plex is associated with that volume, you can  
modify the state of the plex. For example, if a disk with a particular plex located  
on it begins to fail, you can temporarily disable that plex.  
Note: A plex does not have to be associated with a volume. A plex can be created  
with the vxmakeplexcommand and be attached to a volume later.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                   
Creating and administering plexes 225  
Displaying plex information  
VxVM utilities use plex states to:  
■
■
■
■
indicate whether volume contents have been initialized to a known state  
determine if a plex contains a valid copy (mirror) of the volume contents  
track whether a plex was in active use at the time of a system failure  
monitor operations on plexes  
This section explains the individual plex states in detail. For more information  
about the possible transitions between plex states and how these are applied  
during volume recovery, see the chapter “Understanding the Plex State Cycle”  
in the section “Recovery from Hardware Failure” in the Veritas Volume Manager  
Troubleshooting Guide.  
Plexes that are associated with a volume have one of the following states:  
ACTIVE plex state  
A plex can be in the ACTIVE state in two ways:  
■
when the volume is started and the plex fully participates in normal volume  
I/O (the plex contents change as the contents of the volume change)  
■
when the volume is stopped as a result of a system crash and the plex is  
ACTIVE at the moment of the crash  
In the latter case, a system failure can leave plex contents in an inconsistent  
state. When a volume is started, VxVM does the recovery action to guarantee  
that the contents of the plexes marked as ACTIVE are made identical.  
Note: On a system running well, ACTIVE should be the most common state you  
see for any volume plexes.  
CLEAN plex state  
A plex is in a CLEAN state when it is known to contain a consistent copy (mirror)  
of the volume contents and an operation has disabled the volume. As a result,  
when all plexes of a volume are clean, no action is required to guarantee that the  
plexes are identical when that volume is started.  
DCOSNP plex state  
This state indicates that a data change object (DCO) plex attached to a volume  
can be used by a snapshot plex to create a DCO volume during a snapshot  
operation.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
226 Creating and administering plexes  
Displaying plex information  
EMPTY plex state  
Volume creation sets all plexes associated with the volume to the EMPTY state  
to indicate that the plex is not yet initialized.  
IOFAIL plex state  
The IOFAIL plex state is associated with persistent state logging. When the  
vxconfigddaemon detects an uncorrectable I/O failure on an ACTIVE plex, it  
places the plex in the IOFAIL state to exclude it from the recovery selection  
process at volume start time.  
This state indicates that the plex is out-of-date with respect to the volume, and  
that it requires complete recovery. It is likely that one or more of the disks  
associated with the plex should be replaced.  
LOG plex state  
The state of a dirty region logging (DRL) or RAID-5 log plex is always set to LOG.  
OFFLINE plex state  
The vxmendofftask indefinitely detaches a plex from a volume by setting the  
plex state to OFFLINE. Although the detached plex maintains its association  
with the volume, changes to the volume do not update the OFFLINE plex. The  
plex is not updated until the plex is put online and reattached with the vxplex  
atttask. When this occurs, the plex is placed in the STALE state, which causes  
its contents to be recovered at the next vxvolstartoperation.  
SNAPATT plex state  
This state indicates a snapshot plex that is being attached by the snapstart  
operation. When the attach is complete, the state for the plex is changed to  
SNAPDONE. If the system fails before the attach completes, the plex and all of  
its subdisks are removed.  
SNAPDIS plex state  
This state indicates a snapshot plex that is fully attached. A plex in this state can  
be turned into a snapshot volume with the vxplexsnapshotcommand. If the  
system fails before the attach completes, the plex is dissociated from the  
volume. See the vxplex(1M) manual page for more information.  
SNAPDONE plex state  
The SNAPDONE plex state indicates that a snapshot plex is ready for a snapshot  
to be taken using vxassist snapshot.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
Creating and administering plexes 227  
Displaying plex information  
SNAPTMP plex state  
The SNAPTMP plex state is used during a vxassistsnapstartoperation when a  
snapshot is being prepared on a volume.  
STALE plex state  
If there is a possibility that a plex does not have the complete and current  
volume contents, that plex is placed in the STALE state. Also, if an I/O error  
occurs on a plex, the kernel stops using and updating the contents of that plex,  
and the plex state is set to STALE.  
A vxplexattoperation recovers the contents of a STALE plex from an ACTIVE  
plex. Atomic copy operations copy the contents of the volume to the STALE  
plexes. The system administrator can force a plex to the STALE state with a  
vxplexdetoperation.  
TEMP plex state  
Setting a plex to the TEMP state eases some plex operations that cannot occur in  
a truly atomic fashion. For example, attaching a plex to an enabled volume  
requires copying volume contents to the plex before it can be considered fully  
attached.  
A utility sets the plex state to TEMP at the start of such an operation and to an  
appropriate state at the end of the operation. If the system fails for any reason, a  
TEMP plex state indicates that the operation is incomplete. A later vxvol  
startdissociates plexes in the TEMP state.  
TEMPRM plex state  
A TEMPRM plex state is similar to a TEMP state except that at the completion of  
the operation, the TEMPRM plex is removed. Some subdisk operations require a  
temporary plex. Associating a subdisk with a plex, for example, requires  
updating the subdisk with the volume contents before actually associating the  
subdisk. This update requires associating the subdisk with a temporary plex,  
marked TEMPRM, until the operation completes and removes the TEMPRM  
plex.  
If the system fails for any reason, the TEMPRM state indicates that the  
operation did not complete successfully. A later operation dissociates and  
removes TEMPRM plexes.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
228 Creating and administering plexes  
Displaying plex information  
TEMPRMSD plex state  
The TEMPRMSD plex state is used by vxassistwhen attaching new data  
plexes to a volume. If the synchronization operation does not complete, the plex  
and its subdisks are removed.  
Plex condition flags  
vxprintmay also display one of the following condition flags in the STATE  
field:  
IOFAIL plex condition  
The plex was detached as a result of an I/O failure detected during normal  
volume I/O. The plex is out-of-date with respect to the volume, and in need of  
complete recovery. However, this condition also indicates a likelihood that one  
of the disks in the system should be replaced.  
NODAREC plex condition  
No physical disk was found for one of the subdisks in the plex. This implies  
either that the physical disk failed, making it unrecognizable, or that the  
physical disk is no longer attached through a known access path. The plex  
cannot be used until this condition is fixed, or the affected subdisk is  
dissociated.  
NODEVICE plex condition  
A physical device could not be found corresponding to the disk ID in the disk  
media record for one of the subdisks associated with the plex. The plex cannot  
be used until this condition is fixed, or the affected subdisk is dissociated.  
RECOVER plex condition  
A disk corresponding to one of the disk media records was replaced, or was  
reattached too late to prevent the plex from becoming out-of-date with respect  
to the volume. The plex required complete recovery from another plex in the  
volume to synchronize its contents.  
REMOVED plex condition  
Set in the disk media record when one of the subdisks associated with the plex is  
removed. The plex cannot be used until this condition is fixed, or the affected  
subdisk is dissociated.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
               
Creating and administering plexes 229  
Attaching and associating plexes  
Plex kernel states  
The plex kernel state indicates the accessibility of the plex to the volume driver  
which monitors it.  
Note: No user intervention is required to set these states; they are maintained  
internally. On a system that is operating properly, all plexes are enabled.  
The following plex kernel states are defined:  
DETACHED plex kernel state  
Maintenance is being performed on the plex. Any write request to the volume is  
not reflected in the plex. A read request from the volume is not satisfied from  
the plex. Plex operations and ioctlfunction calls are accepted.  
DISABLED plex kernel state  
The plex is offline and cannot be accessed.  
ENABLED plex kernel state  
The plex is online. A write request to the volume is reflected in the plex. A read  
request from the volume is satisfied from the plex. If a plex is sparse, this is  
indicated by the SPARSEmodifier being displayed in the output from the  
vxprint -tcommand.  
Attaching and associating plexes  
A plex becomes a participating plex for a volume by attaching it to a volume.  
(Attaching a plex associates it with the volume and enables the plex for use.) To  
attach a plex to an existing volume, use the following command:  
# vxplex [-g diskgroup] att volume plex  
For example, to attach a plex named vol01-02to a volume named vol01in the  
disk group, mydg, use the following command:  
# vxplex -g mydg att vol01 vol01-02  
If the volume does not already exist, a plex (or multiple plexes) can be associated  
with the volume when it is created using the following command:  
# vxmake [-g diskgroup] -U usetype vol volume plex=plex1[,plex2...]  
For example, to create a mirrored, fsgen-type volume named home,and to  
associate two existing plexes named home-1and home-2with home, use the  
following command:  
# vxmake -g mydg -U fsgen vol home plex=home-1,home-2  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                     
230 Creating and administering plexes  
Taking plexes offline  
Note: You can also use the command vxassistmirrorvolumeto add a data  
plex as a mirror to an existing volume.  
Taking plexes offline  
Once a volume has been created and placed online (ENABLED), VxVM can  
temporarily disconnect plexes from the volume. This is useful, for example,  
when the hardware on which the plex resides needs repair or when a volume has  
been left unstartable and a source plex for the volume revive must be chosen  
manually.  
Resolving a disk or system failure includes taking a volume offline and attaching  
and detaching its plexes. The two commands used to accomplish disk failure  
resolution are vxmendand vxplex.  
To take a plex OFFLINEso that repair or maintenance can be performed on the  
physical disk containing subdisks of that plex, use the following command:  
# vxmend [-g diskgroup] off plex  
If a disk fails (for example, it has a head crash), you can use the vxmend  
command to take offline all plexes that have associated subdisks on the affected  
disk. For example, if plexes vol01-02and vol02-02in the disk group, mydg,  
had subdisks on a drive to be repaired, use the following command to take these  
plexes offline:  
# vxmend -g mydg off vol01-02 vol02-02  
This command places vol01-02and vol02-02in the OFFLINEstate, and they  
remain in that state until it is changed. The plexes are not automatically  
recovered on rebooting the system.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Creating and administering plexes 231  
Detaching plexes  
Detaching plexes  
To temporarily detach one data plex in a mirrored volume, use the following  
command:  
# vxplex [-g diskgroup] det plex  
For example, to temporarily detach a plex named vol01-02in the disk group,  
mydg, and place it in maintenance mode, use the following command:  
# vxplex -g mydg det vol01-02  
This command temporarily detaches the plex, but maintains the association  
between the plex and its volume. However, the plex is not used for I/O. A plex  
detached with the preceding command is recovered at system reboot. The plex  
state is set to STALE, so that if a vxvolstartcommand is run on the appropriate  
volume (for example, on system reboot), the contents of the plex is recovered  
and made ACTIVE.  
When the plex is ready to return as an active part of its volume, follow the  
procedures in the following section, “Reattaching plexes.”  
Reattaching plexes  
When a disk has been repaired or replaced and is again ready for use, the plexes  
must be put back online (plex state set to ACTIVE). To set the plexes to ACTIVE,  
use one of the following procedures depending on the state of the volume.  
■
If the volume is currently ENABLED, use the following command to reattach  
the plex:  
# vxplex [-g diskgroup] att volume plex ...  
For example, for a plex named vol01-02on a volume named vol01in the  
disk group, mydg, use the following command:  
# vxplex -g mydg att vol01 vol01-02  
As when returning an OFFLINEplex to ACTIVE, this command starts to  
recover the contents of the plex and, after the recovery is complete, sets the  
plex utility state to ACTIVE.  
■
If the volume is not in use (not ENABLED), use the following command to re-  
enable the plex for use:  
# vxmend [-g diskgroup] on plex  
For example, to re-enable a plex named vol01-02in the disk group, mydg,  
enter:  
# vxmend -g mydg on vol01-02  
In this case, the state of vol01-02is set to STALE. When the volume is next  
started, the data on the plex is revived from another plex, and incorporated  
into the volume with its state set to ACTIVE.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                 
232 Creating and administering plexes  
Moving plexes  
If the vxinfocommand shows that the volume is unstartable (see “Listing  
Unstartable Volumes” in the section “Recovery from Hardware Failure” in  
the Veritas Volume Manager Troubleshooting Guide), set one of the plexes  
to CLEANusing the following command:  
# vxmend [-g diskgroup] fix clean plex  
Start the volume using the following command:  
# vxvol [-g diskgroup] start volume  
Moving plexes  
Moving a plex copies the data content from the original plex onto a new plex. To  
move a plex, use the following command:  
# vxplex [-g diskgroup] mv original_plex new_plex  
For a move task to be successful, the following criteria must be met:  
■
■
■
The old plex must be an active part of an active (ENABLED) volume.  
The new plex must be at least the same size or larger than the old plex.  
The new plex must not be associated with another volume.  
The size of the plex has several implications:  
■
If the new plex is smaller or more sparse than the original plex, an  
incomplete copy is made of the data on the original plex. If an incomplete  
copy is desired, use the -oforceoption to vxplex.  
■
If the new plex is longer or less sparse than the original plex, the data that  
exists on the original plex is copied onto the new plex. Any area that is not  
on the original plex, but is represented on the new plex, is filled from other  
complete plexes associated with the same volume.  
■
If the new plex is longer than the volume itself, then the remaining area of  
the new plex above the size of the volume is not initialized and remains  
unused.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Creating and administering plexes 233  
Copying volumes to plexes  
Copying volumes to plexes  
This task copies the contents of a volume onto a specified plex. The volume to be  
copied must not be enabled. The plex cannot be associated with any other  
volume. To copy a plex, use the following command:  
# vxplex [-g diskgroup] cp volume new_plex  
After the copy task is complete, new_plex is not associated with the specified  
volume volume. The plex contains a complete copy of the volume data. The plex  
that is being copied should be the same size or larger than the volume. If the  
plex being copied is larger than the volume, an incomplete copy of the data  
results. For the same reason, new_plex should not be sparse.  
Dissociating and removing plexes  
When a plex is no longer needed, you can dissociate it from its volume and  
remove it as an object from VxVM. You might want to remove a plex for the  
following reasons:  
■
■
to provide free disk space  
to reduce the number of mirrors in a volume so you can increase the length  
of another mirror and its associated volume. When the plexes and subdisks  
are removed, the resulting space can be added to other volumes  
■
■
to remove a temporary mirror that was created to back up a volume and is no  
longer needed  
to change the layout of a plex  
Caution: To save the data on a plex to be removed, the configuration of that plex  
must be known. Parameters from that configuration (stripe unit size and  
subdisk ordering) are critical to the creation of a new plex to contain the same  
data. Before a plex is removed, you must record its configuration. See  
“Displaying plex information” on page 224” for more information.  
To dissociate a plex from the associated volume and remove it as an object from  
VxVM, use the following command:  
# vxplex [-g diskgroup] -o rm displex  
For example, to dissociate and remove a plex named vol01-02in the disk  
group, mydg, use the following command:  
# vxplex -g mydg -o rm dis vol01-02  
This command removes the plex vol01-02and all associated subdisks.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
               
234 Creating and administering plexes  
Changing plex attributes  
Alternatively, you can first dissociate the plex and subdisks, and then remove  
them with the following commands:  
# vxplex [-g diskgroup] dis plex  
# vxedit [-g diskgroup] -r rm plex  
When used together, these commands produce the same result as the vxplex -o  
rm discommand. The -roption to vxedit rmrecursively removes all objects  
from the specified object downward. In this way, a plex and its associated  
subdisks can be removed by a single vxeditcommand.  
Changing plex attributes  
Caution: Change plex attributes with extreme care.  
The vxeditcommand changes the attributes of plexes and other volume  
Manager objects. To change plex attributes, use the following command:  
# vxedit [-g diskgroup] set attribute=value ... plex  
Plex fields that can be changed using the vxeditcommand include:  
■ name  
■ putiln  
■ tutiln  
■ comment  
The putiln field attributes are maintained on reboot; tutiln fields are  
temporary and are not retained on reboot. VxVM sets the putil0and tutil0  
utility fields. Other Symantec products, such as the Veritas Enterprise  
Administrator (VEA), set the putil1and tutil1fields. The putil2and  
tutil2are available for you to use for site-specific purposes.  
The following example command sets the comment field, and also sets tutil2  
to indicate that the subdisk is in use:  
# vxedit -g mydg set comment=”plex comment” tutil2=”u” vol01-02  
To prevent a particular plex from being associated with a volume, set the  
putil0field to a non-null string, as shown in the following command:  
# vxedit -g mydg set putil0=”DO-NOT-USE” vol01-02  
See the vxedit(1M) manual page for more information about using the vxedit  
command to change the attribute fields of VxVM objects.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
               
Chapter  
7
Creating volumes  
This chapter describes how to create volumes in Veritas Volume Manager  
(VxVM). Volumes are logical devices that appear as physical disk partition  
devices to data management systems. Volumes enhance recovery from  
hardware failure, data availability, performance, and storage configuration.  
Note: You can also use the Veritas Intelligent Storage Provisioning (ISP) feature  
to create and administer application volumes. These volumes are very similar to  
the traditional VxVM volumes that are described in this chapter. However, there  
are significant differences between the functionality of the two types of volume  
that prevents them from being used interchangeably. Refer to the Veritas  
Storage Foundation Intelligent Storage Provisioning Administrator’s Guide for  
more information about creating and administering ISP application volumes.  
Volumes are created to take advantage of the VxVM concept of virtual disks. A  
file system can be placed on the volume to organize the disk space with files and  
directories. In addition, you can configure applications such as databases to  
organize data on volumes.  
Note: Disks and disk groups must be initialized and defined to VxVM before  
volumes can be created from them. See “Administering disks” on page 77 and  
“Creating and administering disk groups” on page 165 for more information.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
236 Creating volumes  
Types of volume layouts  
Types of volume layouts  
VxVM allows you to create volumes with the following layout types:  
Concatenated  
A volume whose subdisks are arranged both sequentially and  
contiguously within a plex. Concatenation allows a volume to be  
created from multiple regions of one or more disks if there is not  
enough space for an entire volume on a single region of a disk. For  
more information, see “Concatenation and spanning” on page 35.  
Striped  
A volume with data spread evenly across multiple disks. Stripes are  
equal-sized fragments that are allocated alternately and evenly to the  
subdisks of a single plex. There must be at least two subdisks in a  
striped plex, each of which must exist on a different disk. Throughput  
increases with the number of disks across which a plex is striped.  
Striping helps to balance I/O load in cases where high traffic areas  
exist on certain subdisks. For more information, see “Striping (RAID-  
0)” on page 38.  
Mirrored  
RAID-5  
A volume with multiple data plexes that duplicate the information  
contained in a volume. Although a volume can have a single data plex,  
at least two are required for true mirroring to provide redundancy of  
data. For the redundancy to be useful, each of these data plexes  
should contain disk space from different disks. For more information,  
see “Mirroring (RAID-1)” on page 42.  
A volume that uses striping to spread data and parity evenly across  
multiple disks in an array. Each stripe contains a parity stripe unit  
and data stripe units. Parity can be used to reconstruct data if one of  
the disks fails. In comparison to the performance of striped volumes,  
write throughput of RAID-5 volumes decreases since parity  
information needs to be updated each time data is modified. However,  
in comparison to mirroring, the use of parity to implement data  
redundancy reduces the amount of space required. For more  
information, see “RAID-5 (striping with parity)” on page 45.  
Mirrored-stripe  
A volume that is configured as a striped plex and another plex that  
mirrors the striped one. This requires at least two disks for striping  
and one or more other disks for mirroring (depending on whether the  
plex is simple or striped). The advantages of this layout are increased  
performance by spreading data across multiple disks and redundancy  
of data. “Striping plus mirroring (mirrored-stripe or RAID-0+1)” on  
page 42.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                 
Creating volumes 237  
Types of volume layouts  
Layered Volume  
A volume constructed from other volumes. Non-layered volumes are  
constructed by mapping their subdisks to VM disks. Layered volumes  
are constructed by mapping their subdisks to underlying volumes  
(known as storage volumes), and allow the creation of more complex  
forms of logical layout. Examples of layered volumes are striped-  
mirror and concatenated-mirror volumes.  
See “Layered volumes” on page 51.  
A striped-mirror volume is created by configuring several mirrored  
volumes as the columns of a striped volume. This layout offers the  
same benefits as a non-layered mirrored-stripe volume. In addition it  
provides faster recovery as the failure of single disk does not force an  
entire striped plex offline.  
See “Mirroring plus striping (striped-mirror, RAID-1+0 or RAID-10)”  
on page 43.  
A concatenated-mirror volume is created by concatenating several  
mirrored volumes. This provides faster recovery as the failure of a  
single disk does not force the entire mirror offline.  
Supported volume logs and maps  
Veritas Volume Manager supports the use of several types of logs and maps with  
volumes:  
■
FastResync Maps are used to perform quick and efficient resynchronization  
of mirrors (see “FastResync” on page 66 for details). These maps are  
supported either in memory (Non-Persistent FastResync), or on disk as part  
of a DCO volume (Persistent FastResync). Two types of DCO volume are  
supported:  
■
■
Version 0 DCO volumes only support Persistent FastResync for the  
traditional third-mirror break-off type of volume snapshot. See  
“Version 0 DCO volume layout” on page 69, and “Creating a volume  
with a version 0 DCO volume” on page 250 for more information.  
Version 20 DCO volumes, introduced in VxVM 4.0, support DRL logging  
(see below) and Persistent FastResync for full-sized and space-  
optimized instant volume snapshots. See “Version 20 DCO volume  
layout” on page 69, and “Creating a volume with a version 20 DCO  
volume” on page 252 for more information.  
See “Enabling FastResync on a volume” on page 292 for information on how  
to enable Persistent or Non-Persistent FastResync on a volume.  
■
Dirty region logs allow the fast recovery of mirrored volumes after a system  
crash (see “Dirty region logging” on page 60 for details). These logs are  
supported either as DRL log plexes, or as part of a version 20 DCO volume.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
238 Creating volumes  
Creating a volume  
Refer to the following sections for information on creating a volume on  
which DRL is enabled:  
■
“Creating a volume with dirty region logging enabled” on page 252 for  
creating a volume with DRL log plexes.  
■
“Creating a volume with a version 20 DCO volume” on page 252 for  
creating a volume with DRL configured within a version 20 DCO  
volume.  
■
RAID-5 logs are used to prevent corruption of data during recovery of RAID-  
5 volumes (see “RAID-5 logging” on page 50 for details). These logs are  
configured as plexes on disks other than those that are used for the columns  
of the RAID-5 volume.  
See “Creating a RAID-5 volume” on page 256 for information on creating a  
RAID-5 volume together with RAID-5 logs.  
Creating a volume  
You can create volumes using an advanced approach, an assisted approach, or  
the rule-based storage allocation approach that is provided by the Intelligent  
Storage Provisioning (ISP) feature. Each method uses different tools. You may  
switch between the advanced and the assisted approaches at will. For more  
information about ISP, see the Veritas Storage Foundation Intelligent Storage  
Provisioning Administrator’s Guide.  
Note: Most VxVM commands require superuser or equivalent privileges.  
Advanced approach  
The advanced approach consists of a number of commands that typically  
require you to specify detailed input. These commands use a “building block”  
approach that requires you to have a detailed knowledge of the underlying  
structure and components to manually perform the commands necessary to  
accomplish a certain task. Advanced operations are performed using several  
different VxVM commands.  
To create a volume using the advanced approach  
1
2
Create subdisks using vxmakesd; see “Creating subdisks” on page 215.  
Create plexes using vxmakeplex, and associate subdisks with them; see  
“Creating plexes” on page 223, “Associating subdisks with plexes” on  
page 218 and “Creating a volume using vxmake” on page 258.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Creating volumes 239  
Using vxassist  
3
4
Associate plexes with the volume using vxmakevol; see “Creating a volume  
using vxmake” on page 258.  
Initialize the volume using vxvolstartor vxvolinitzero; see “Initializing  
and starting a volume created using vxmake” on page 261.  
See “Creating a volume using a vxmake description file” on page 259 for an  
example of how you can combine steps 1 through 3 using a volume description  
file with vxmake.  
See “Creating a volume using vxmake” on page 258 for an example of how to  
perform steps 2 and 3 to create a RAID-5 volume.  
Assisted approach  
The assisted approach takes information about what you want to accomplish  
and then performs the necessary underlying tasks. This approach requires only  
minimal input from you, but also permits more detailed specifications.  
Assisted operations are performed primarily through the vxassistcommand or  
the Veritas Enterprise Administrator (VEA). vxassistand the VEA create the  
required plexes and subdisks using only the basic attributes of the desired  
volume as input. Additionally, they can modify existing volumes while  
automatically modifying any underlying or associated objects.  
Both vxassistand the VEA use default values for many volume attributes,  
unless you provide specific values. They do not require you to have a thorough  
understanding of low-level VxVM concepts, vxassistand the VEA do not  
conflict with other VxVM commands or preclude their use. Objects created by  
vxassistand the VEA are compatible and inter-operable with objects created  
by other VxVM commands and interfaces.  
For more information about the VEA, see the Veritas Enterprise Administrator  
User’s Guide and VEA online help.  
Using vxassist  
You can use the vxassistutility to create and modify volumes. Specify the basic  
requirements for volume creation or modification, and vxassistperforms the  
necessary tasks.  
The advantages of using vxassistrather than the advanced approach include:  
■
■
Most actions require that you enter only one command rather than several.  
You are required to specify only minimal information to vxassist. If  
necessary, you can specify additional parameters to modify or control its  
actions.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
240 Creating volumes  
Using vxassist  
■
Operations result in a set of configuration changes that either succeed or  
fail as a group, rather than individually. System crashes or other  
interruptions do not leave intermediate states that you have to clean up. If  
vxassistfinds an error or an exceptional condition, it exits after leaving  
the system in the same state as it was prior to the attempted operation.  
The vxassistutility helps you perform the following tasks:  
■
■
■
■
■
Creating volumes.  
Creating mirrors for existing volumes.  
Growing or shrinking existing volumes.  
Backing up volumes online.  
Reconfiguring a volume’s layout online.  
vxassistobtains most of the information it needs from sources other than  
your input. vxassistobtains information about the existing objects and their  
layouts from the objects themselves.  
For tasks requiring new disk space, vxassistseeks out available disk space and  
allocates it in the configuration that conforms to the layout specifications and  
that offers the best use of free space.  
The vxassistcommand takes this form:  
# vxassist [options] keyword volume [attributes...]  
where keyword selects the task to perform. The first argument after a vxassist  
keyword, volume, is a volume name, which is followed by a set of desired volume  
attributes. For example, the keyword makeallows you to create a new volume:  
# vxassist [options] make volume length [attributes]  
The length of the volume can be specified in sectors, kilobytes, megabytes, or  
gigabytes using a suffix character of s, k, m, or g. If no suffix is specified, the size  
is assumed to be in sectors. See the vxintro(1M) manual page for more  
information on specifying units.  
Additional attributes can be specified as appropriate, depending on the  
characteristics that you wish the volume to have. Examples are stripe unit  
width, number of columns in a RAID-5 or stripe volume, number of mirrors,  
number of logs, and log type.  
Note: By default, the vxassistcommand creates volumes in a default disk group  
according to the rules given in “Rules for determining the default disk group” on  
page 168. To use a different disk group, specify the -gdiskgroup option to  
vxassist.  
For details of available vxassistkeywords and attributes, refer to the  
vxassist(1M) manual page.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Creating volumes 241  
Using vxassist  
The section, “Creating a volume on any disk” on page 243 describes the simplest  
way to create a volume with default attributes. Later sections describe how to  
create volumes with specific attributes. For example, “Creating a volume on  
specific disks” on page 244 describes how to control how vxassistuses the  
available storage space.  
Setting default values for vxassist  
The default values that the vxassistcommand uses may be specified in the file  
/etc/default/vxassist. The defaults listed in this file take effect if you do  
not override them on the command line, or in an alternate defaults file that you  
specify using the -doption. A default value specified on the command line  
always takes precedence. vxassistalso has a set of built-in defaults that it uses  
if it cannot find a value defined elsewhere.  
Note: You must create the /etc/defaultdirectory and the vxassistdefault  
file if these do not already exist on your system.  
The format of entries in a defaults file is a list of attribute-value pairs separated  
by new lines. These attribute-value pairs are the same as those specified as  
options on the vxassistcommand line. Refer to the vxassist(1M) manual page  
for details.  
To display the default attributes held in the file /etc/default/vxassist, use  
the following form of the vxassistcommand:  
# vxassist help showattrs  
The following is a sample vxassistdefaults file:  
# By default:  
# create unmirrored, unstriped volumes  
# allow allocations to span drives  
# with RAID-5 create a log, with mirroring don’t create a log  
# align allocations on cylinder boundaries  
layout=nomirror,nostripe,span,nocontig,raid5log,noregionlog,  
diskalign  
# use the fsgen usage type, except when creating RAID-5 volumes  
usetype=fsgen  
# allow only root access to a volume  
mode=u=rw,g=,o=  
user=root  
group=root  
# when mirroring, create two mirrors  
nmirror=2  
# for regular striping, by default create between 2 and 8 stripe  
# columns  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
242 Creating volumes  
Discovering the maximum size of a volume  
max_nstripe=8  
min_nstripe=2  
# for RAID-5, by default create between 3 and 8 stripe columns  
max_nraid5stripe=8  
min_nraid5stripe=3  
# by default, create 1 log copy for both mirroring and RAID-5  
volumes  
nregionlog=1  
nraid5log=1  
# by default, limit mirroring log lengths to 32Kbytes  
max_regionloglen=32k  
# use 64K as the default stripe unit size for regular volumes  
stripe_stwid=64k  
# use 16K as the default stripe unit size for RAID-5 volumes  
raid5_stwid=16k  
Discovering the maximum size of a volume  
To find out how large a volume you can create within a disk group, use the  
following form of the vxassistcommand:  
# vxassist [-g diskgroup] maxsize layout=layout [attributes]  
For example, to discover the maximum size RAID-5 volume with 5 columns and  
2 logs that you can create within the disk group, dgrp, enter the following  
command:  
# vxassist -g dgrp maxsize layout=raid5 nlog=2  
You can use storage attributes if you want to restrict the disks that vxassist  
uses when creating volumes. See “Creating a volume on specific disks” on  
page 244 for more information.  
Note: The maximum size of a VxVM volume that you can create is 256TB.  
Disk group alignment constraints on volumes  
Certain constraints apply to the length of volumes and to the numeric values of  
size attributes that apply to volumes. If a volume is created in a disk group that  
is compatible with the Cross-platform Data Sharing (CDS) feature, the volume’s  
length and the values of volume attributes that define the sizes of objects such  
as logs or stripe units, must be an integer multiple of the alignment value of 8  
blocks (8 kilobytes). If the disk group is not compatible with the CDS feature, the  
volume’s length and attribute size values must be multiples of 1 block  
(1kilobyte).  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Creating volumes 243  
Creating a volume on any disk  
To discover the value in blocks of the alignment that is set on a disk group, use  
this command:  
# vxprint -g diskgroup -G -F %align  
By default, vxassistautomatically rounds up the volume size and attribute size  
values to a multiple of the alignment value. (This is equivalent to specifying the  
attribute dgalign_checking=roundas an additional argument to the  
vxassistcommand.)  
If you specify the attribute dgalign_checking=strictto vxassist, the  
command fails with an error if you specify a volume length or attribute size  
value that is not a multiple of the alignment value for the disk group.  
Creating a volume on any disk  
By default, the vxassistmakecommand creates a concatenated volume that  
uses one or more sections of disk space. On a fragmented disk, this allows you to  
put together a volume larger than any individual section of free disk space  
available.  
Note: To change the default layout, edit the definition of the layoutattribute  
defined in the /etc/default/vxassistfile.  
If there is not enough space on a single disk, vxassistcreates a spanned  
volume. A spanned volume is a concatenated volume with sections of disk space  
spread across more than one disk. A spanned volume can be larger than any disk  
on a system, since it takes space from more than one disk.  
To create a concatenated, default volume, use the following form of the  
vxassistcommand:  
# vxassist [-b] [-g diskgroup] make volume length  
Note: Specify the -boption if you want to make the volume immediately  
available for use. See “Initializing and starting a volume” on page 260 for  
details.  
For example, to create the concatenated volume voldefaultwith a length of  
10 gigabytes in the default disk group:  
# vxassist -b make voldefault 10g  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
244 Creating volumes  
Creating a volume on specific disks  
Creating a volume on specific disks  
VxVM automatically selects the disks on which each volume resides, unless you  
specify otherwise. If you want a volume to be created on specific disks, you must  
designate those disks to VxVM. More than one disk can be specified.  
To create a volume on a specific disk or disks, use the following command:  
# vxassist [-b] [-g diskgroup] make volume length \  
[layout=layout] diskname ...  
For example, to create the volume volspecwith length 5 gigabytes on disks  
mydg03and mydg04, use the following command:  
# vxassist -b -g mydg make volspec 5g mydg03 mydg04  
The vxassistcommand allows you to specify storage attributes. These give you  
control over the devices, including disks, controllers and targets, which  
vxassistuses to configure a volume. For example, you can specifically exclude  
disk mydg05:  
# vxassist -b -g mydg make volspec 5g !mydg05  
or exclude all disks that are on controller c2:  
# vxassist -b -g mydg make volspec 5g !ctlr:c2  
or include only disks on controller c1except for target t5:  
# vxassist -b -g mydg make volspec 5g ctlr:c1 !target:c1t5  
If you want a volume to be created using only disks from a specific disk group,  
use the -goption to vxassist, for example:  
# vxassist -g bigone -b make volmega 20g bigone10 bigone11  
or alternatively, use the diskgroupattribute:  
# vxassist -b make volmega 20g diskgroup=bigone bigone10 \  
bigone11  
Note: Any storage attributes that you specify for use must belong to the disk  
group. Otherwise, vxassistwill not use them to create a volume.  
You can also use storage attributes to control how vxassistuses available  
storage, for example, when calculating the maximum size of a volume, when  
growing a volume or when removing mirrors or logs from a volume. The  
following example excludes disks dgrp07and dgrp08when calculating the  
maximum size of RAID-5 volume that vxassistcan create using the disks in the  
disk group dg:  
# vxassist -b -g dgrp maxsize layout=raid5 nlog=2 !dgrp07 \  
!dgrp08  
See the vxassist(1M) manual page for more information about using storage  
attributes. It is also possible to control how volumes are laid out on the specified  
storage as described in the next section “Specifying ordered allocation of  
storage to volumes.”  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Creating volumes 245  
Creating a volume on specific disks  
Specifying ordered allocation of storage to volumes  
Ordered allocation gives you complete control of space allocation. It requires  
that the number of disks that you specify to the vxassistcommand must match  
the number of disks that are required to create a volume. The order in which you  
specify the disks to vxassistis also significant.  
If you specify the -oorderedoption to vxassistwhen creating a volume, any  
storage that you also specify is allocated in the following order:  
1
2
3
Concatenate disks.  
Form columns.  
Form mirrors.  
For example, the following command creates a mirrored-stripe volume with 3  
columns and 2 mirrors on 6 disks in the disk group, mydg:  
# vxassist -b -g mydg -o ordered make mirstrvol 10g \  
layout=mirror-stripe ncol=3 \  
mydg01 mydg02 mydg03 mydg04 mydg05 mydg06  
This command places columns 1, 2 and 3 of the first mirror on disks mydg01,  
mydg02and mydg03respectively, and columns 1, 2 and 3 of the second mirror  
on disks mydg04, mydg05and mydg06respectively. This arrangement is  
illustrated in Figure 7-1.  
Figure 7-1  
Example of using ordered allocation to create a mirrored-stripe  
volume  
Column 1  
Column 2  
Column 3  
Striped plex  
Mirror  
mydg01-01  
mydg02-01  
mydg03-01  
Column 1  
Column 2  
Column 3  
Striped plex  
mydg04-01  
mydg05-01  
mydg06-01  
Mirrored-stripe  
volume  
For layered volumes, vxassistapplies the same rules to allocate storage as for  
non-layered volumes. For example, the following command creates a striped-  
mirror volume with 2 columns:  
# vxassist -b -g mydg -o ordered make strmirvol 10g \  
layout=stripe-mirror ncol=2 mydg01 mydg02 mydg03 mydg04  
This command mirrors column 1 across disks mydg01and mydg03, and column  
2 across disks mydg02and mydg04, as illustrated in Figure 7-2.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
246 Creating volumes  
Creating a volume on specific disks  
Figure 7-2  
Example of using ordered allocation to create a striped-mirror  
volume  
Underlying mirrored volumes  
Column 1  
mydg01-01  
Column 2  
mydg02-01  
Mirror  
Column 1  
Column 2  
mydg03-01  
mydg04-01  
Striped plex  
Striped-mirror volume  
Additionally, you can use the col_switchattribute to specify how to  
concatenate space on the disks into columns. For example, the following  
command creates a mirrored-stripe volume with 2 columns:  
# vxassist -b -g mydg -o ordered make strmir2vol 10g \  
layout=mirror-stripe ncol=2 col_switch=3g,2g \  
mydg01 mydg02 mydg03 mydg04 mydg05 mydg06 mydg07 mydg08  
This command allocates 3 gigabytes from mydg01and 2 gigabytes from mydg02  
to column 1, and 3 gigabytes from mydg03and 2 gigabytes from mydg04to  
column 2. The mirrors of these columns are then similarly formed from disks  
mydg05through mydg08. This arrangement is illustrated in Figure 7-3.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Creating volumes 247  
Creating a volume on specific disks  
Figure 7-3  
Example of using concatenated disk space to create a mirrored-  
stripe volume  
Column 1  
Column 2  
mydg01-01  
mydg02-01  
mydg03-01  
mydg04-01  
Striped plex  
Mirror  
Column 1  
Column 2  
mydg05-01  
mydg07-01  
Striped plex  
mydg06-01  
mydg08-01  
Mirrored-stripe volume  
Other storage specification classes for controllers, enclosures, targets and trays  
can be used with ordered allocation. For example, the following command  
creates a 3-column mirrored-stripe volume between specified controllers:  
# vxassist -b -g mydg -o ordered make mirstr2vol 80g \  
layout=mirror-stripe ncol=3 \  
ctlr:c1 ctlr:c2 ctlr:c3 ctlr:c4 ctlr:c5 ctlr:c6  
This command allocates space for column 1 from disks on controllers c1, for  
column 2 from disks on controller c2, and so on as illustrated in Figure 7-4.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
248 Creating volumes  
Creating a volume on specific disks  
Figure 7-4  
Example of storage allocation used to create a mirrored-stripe  
volume across controllers  
c1  
c2  
c3  
Controllers  
Column 1  
Column 2  
Column 3  
Striped plex  
Mirror  
Column 1  
Column 2  
Column 3  
Striped plex  
Mirrored-stripe  
volume  
c4  
c5  
c6  
Controllers  
For other ways in which you can control how vxassistlays out mirrored  
volumes across controllers, see “Mirroring across targets, controllers or  
enclosures” on page 255.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Creating volumes 249  
Creating a mirrored volume  
Creating a mirrored volume  
Note: You need a full license to use this feature.  
A mirrored volume provides data redundancy by containing more than one copy  
of its data. Each copy (or mirror) is stored on different disks from the original  
copy of the volume and from other mirrors. Mirroring a volume ensures that its  
data is not lost if a disk in one of its component mirrors fails.  
Note: A mirrored volume requires space to be available on at least as many disks  
in the disk group as the number of mirrors in the volume.  
To create a new mirrored volume, use the following command:  
# vxassist [-b] [-g diskgroup] make volume length \  
layout=mirror [nmirror=number] [init=active]  
For example, to create the mirrored volume, volmir, in the disk group, mydg,  
use the following command:  
# vxassist -b -g mydg make volmir 5g layout=mirror  
To create a volume with 3 instead of the default of 2 mirrors, modify the  
command to read:  
# vxassist -b -g mydg make volmir 5g layout=mirror nmirror=3  
Creating a mirrored-concatenated volume  
Note: You need a full license to use this feature.  
A mirrored-concatenated volume mirrors several concatenated plexes. To create  
a concatenated-mirror volume, use the following command:  
# vxassist [-b] [-g diskgroup] make volume length \  
layout=mirror-concat [nmirror=number]  
Alternatively, first create a concatenated volume, and then mirror it as  
described in “Adding a mirror to a volume” on page 271.  
Creating a concatenated-mirror volume  
Note: You need a full license to use this feature.  
A concatenated-mirror volume is an example of a layered volume which  
concatenates several underlying mirror volumes. To create a concatenated-  
mirror volume, use the following command:  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                         
250 Creating volumes  
Creating a volume with a version 0 DCO volume  
# vxassist [-b] [-g diskgroup] make volume length \  
layout=concat-mirror [nmirror=number]  
Creating a volume with a version 0 DCO volume  
If a data change object (DCO) and DCO volume are associated with a volume, this  
allows Persistent FastResync to be used with the volume. (See “How persistent  
FastResync works with snapshots” on page 70 for details of how Persistent  
FastResync performs fast resynchronization of snapshot mirrors when they are  
returned to their original volume.)  
Note: The procedure described in this section creates a volume with a data  
change object (DCO) and DCO volume that has a version 0 layout as introduced  
in VxVM 3.2. The version 0 layout supports traditional (third-mirror) snapshots,  
but not full-sized instant snapshots, space-optimized instant snapshots nor DRL  
configured within the DCO volume. See “Version 0 DCO volume layout” on  
page 69 and “Version 20 DCO volume layout” on page 69 for a description of the  
differences between the old and new DCO volume layouts.  
For details of how to configure a volume with a version 20 DCO and DCO volume,  
see “Creating a volume with a version 20 DCO volume” on page 252. This is the  
preferred and recommended method.  
See “Determining the DCO version number” on page 277 for details of how to  
determine the version number of a volume’s DCO.  
To perform fast resynchronization of mirrors after a system crash or reboot,  
you must also enable dirty region logging (DRL) on a mirrored volume. To add a  
DCO object and DCO volume to a volume on which DRL logging is enabled, follow  
the procedure described in “Adding a version 0 DCO and DCO volume” on  
page 356.  
TM  
Note: You need a Veritas FlashSnap or FastResync license to use the Persistent  
FastResync feature. Even if you do not have a license, you can configure a DCO  
object and DCO volume so that snap objects are associated with the original and  
snapshot volumes. For more information about snap objects, see “How  
persistent FastResync works with snapshots” on page 70.  
To create a volume with an attached version 0 DCO object and volume  
1
Ensure that the disk group has been upgraded to version 90. Use the  
following command to check the version of a disk group:  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Creating volumes 251  
Creating a volume with a version 0 DCO volume  
# vxdg list diskgroup  
To upgrade a disk group to version 90, use the following command:  
# vxdg -T 90 upgrade diskgroup  
For more information, see “Upgrading a disk group” on page 208.  
2
Use the following command to create the volume (you may need to specify  
additional attributes to create a volume with the desired characteristics):  
# vxassist [-g diskgroup] make volume length layout=layout \  
logtype=dco [ndcomirror=number] [dcolen=size] \  
[fastresync=on] [other attributes]  
For non-layered volumes, the default number of plexes in the mirrored DCO  
volume is equal to the lesser of the number of plexes in the data volume or  
2. For layered volumes, the default number of DCO plexes is always 2. If  
required, use the ndcomirrorattribute to specify a different number. It is  
recommended that you configure as many DCO plexes as there are data  
plexes in the volume. For example, specify ndcomirror=3when creating a 3-  
way mirrored volume.  
The default size of each plex is 132 blocks unless you use the dcolen  
attribute to specify a different size. If specified, the size of the plex must be  
a multiple of 33 blocks from 33 up to a maximum of 2112 blocks.  
By default, FastResync is not enabled on newly created volumes. Specify the  
fastresync=onattribute if you want to enable FastResync on the volume. If  
a DCO object and DCO volume are associated with the volume, Persistent  
FastResync is enabled; otherwise, Non-Persistent FastResync is enabled.  
3
To enable DRL or sequential DRL logging on the newly created volume, use  
the following command:  
# vxvol [-g diskgroup] set logtype=drl|drlseq volume  
For more information, see the vxassist(1M) and vxvol(1M) manual pages.  
If you use ordered allocation when creating a mirrored volume on specified  
storage, you can use the optional logdiskattribute to specify on which disks  
dedicated log plexes should be created. Use the following form of the vxassist  
command to specify the disks from which space for the logs is to be allocated:  
# vxassist [-g diskgroup] -o ordered make volume length \  
layout=mirror logtype=log_type logdisk=disk[,disk,...] \  
storage_attributes  
If you do not specify the logdiskattribute, vxassistlocates the logs in the data  
plexes of the volume.  
For more information about ordered allocation, see “Specifying ordered  
allocation of storage to volumes” on page 245 and the vxassist(1M) manual  
page.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
252 Creating volumes  
Creating a volume with a version 20 DCO volume  
Creating a volume with a version 20 DCO volume  
To create a volume with an attached version 20 DCO object and volume  
1
Ensure that the disk group has been upgraded to the latest version. Use the  
following command to check the version of a disk group:  
# vxdg list diskgroup  
To upgrade a disk group to the most recent version, use the following  
command:  
# vxdg upgrade diskgroup  
For more information, see “Upgrading a disk group” on page 208.  
2
Use the following command to create the volume (you may need to specify  
additional attributes to create a volume with the desired characteristics):  
# vxassist [-g diskgroup] make volume length layout=layout \  
logtype=dco dcoversion=20 [drl=on|sequential|off] \  
[ndcomirror=number] [fastresync=on] [other attributes]  
Set the value of the drlattribute to onif dirty region logging (DRL) is to be  
used with the volume (this is the default setting). For a volume that will be  
written to sequentially, such as a database log volume, set the value to  
sequentialto enable sequential DRL. The DRL logs are created in the DCO  
volume. The redundancy of the logs is determined by the number of mirrors  
that you specify using the ndcomirrorattribute.  
By default, Persistent FastResync is not enabled on newly created volumes.  
Specify the fastresync=onattribute if you want to enable Persistent  
FastResync on the volume.  
For more information, see the vxassist(1M) manual page.  
Note: See “Determining the DCO version number” on page 277 for details of how  
to determine the version number of a volume’s DCO.  
Creating a volume with dirty region logging enabled  
Note: The procedure in this section is applicable to volumes that are created in  
disk groups with a version number of less than 110. To enable DRL or sequential  
DRL on a volume that is created within a disk group with a version number of  
110 or greater, follow the procedure described in “Creating a volume with a  
version 20 DCO volume” on page 252, which creates the DRL logs within the  
plexes of a version 20 DCO volume.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                   
Creating volumes 253  
Creating a striped volume  
Dirty region logging (DRL), if enabled, speeds recovery of mirrored volumes  
after a system crash. To enable DRL on a volume that is created within a disk  
group with a version number between 20 and 100, specify the logtype=drl  
attribute to the vxassistmakecommand as shown in this example usage:  
# vxassist [-g diskgroup] make volume length layout=layout \  
logtype=drl [nlog=n] [loglen=size] [other attributes]  
The nlogattribute can be used to specify the number of log plexes to add. By  
default, one log plex is added. The loglenattribute specifies the size of the log,  
where each bit represents one region in the volume. For example, the size of the  
log would need to be 20K for a 10GB volume with a region size of 64 kilobytes.  
For example, to create a mirrored 10GB volume, vol02, with two log plexes in  
the disk group, mydg, use the following command:  
# vxassist -g mydg make vol02 10g layout=mirror logtype=drl \  
nlog=2 nmirror=2  
Sequential DRL limits the number of dirty regions for volumes that are written  
to sequentially, such as database replay logs. To enable sequential DRL on a  
volume that is created within a disk group with a version number between 70  
and 100, specify the logtype=drlseqattribute to the vxassistmakecommand.  
# vxassist [-g diskgroup] make volume length layout=layout \  
logtype=drlseq [nlog=n] [other attributes]  
Note: If you also want to allow the use of Persistent FastResync with the volume,  
use the procedure described in “Creating a volume with a version 0 DCO volume”  
on page 250.  
Creating a striped volume  
A striped volume contains at least one plex that consists of two or more subdisks  
located on two or more physical disks. For more information on striping, see  
“Striping (RAID-0)” on page 38.  
Note: A striped volume requires space to be available on at least as many disks in  
the disk group as the number of columns in the volume.  
To create a striped volume, use the following command:  
# vxassist [-b] [-g diskgroup] makevolume length layout=stripe  
For example, to create the 10-gigabyte striped volume volzebra, in the disk  
group, mydg, use the following command:  
# vxassist -b -g mydg make volzebra 10g layout=stripe  
This creates a striped volume with the default stripe unit size (64 kilobytes) and  
the default number of stripes (2).  
Download from Www.Somanuals.com. All Manuals Search And Download.  
               
254 Creating volumes  
Creating a striped volume  
You can specify the disks on which the volumes are to be created by including  
the disk names on the command line. For example, to create a 30-gigabyte  
striped volume on three specific disks, mydg03, mydg04, and mydg05, use the  
following command:  
# vxassist -b -g mydg make stripevol 30g layout=stripe \  
mydg03 mydg04 mydg05  
To change the number of columns or the stripe width, use the ncolumnand  
stripeunitmodifiers with vxassist. For example, the following command  
creates a striped volume with 5 columns and a 32-kilobyte stripe size:  
# vxassist -b -g mydg make stripevol 30g layout=stripe \  
stripeunit=32k ncol=5  
Creating a mirrored-stripe volume  
A mirrored-stripe volume mirrors several striped data plexes.  
Note: A mirrored-stripe volume requires space to be available on at least as  
many disks in the disk group as the number of mirrors multiplied by the number  
of columns in the volume.  
To create a striped-mirror volume, use the following command:  
# vxassist [-b] [-g diskgroup] make volume length \  
layout=mirror-stripe [nmirror=number_mirrors] \  
[ncol=number_of_columns] [stripewidth=size]  
Alternatively, first create a striped volume, and then mirror it as described in  
“Adding a mirror to a volume” on page 271. In this case, the additional data  
plexes may be either striped or concatenated.  
Creating a striped-mirror volume  
A striped-mirror volume is an example of a layered volume which stripes several  
underlying mirror volumes.  
Note: A striped-mirror volume requires space to be available on at least as many  
disks in the disk group as the number of columns multiplied by the number of  
stripes in the volume.  
To create a striped-mirror volume, use the following command:  
# vxassist [-b] [-g diskgroup] make volume length \  
layout=stripe-mirror [nmirror=number_mirrors] \  
[ncol=number_of_columns] [stripewidth=size]  
By default, VxVM attempts to create the underlying volumes by mirroring  
subdisks rather than columns if the size of each column is greater than the value  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                 
Creating volumes 255  
Mirroring across targets, controllers or enclosures  
for the attribute stripe-mirror-col-split-trigger-ptthat is defined in  
the vxassistdefaults file.  
If there are multiple subdisks per column, you can choose to mirror each subdisk  
individually instead of each column. To mirror at the subdisk level, specify the  
layout as stripe-mirror-sdrather than stripe-mirror. To mirror at the  
column level, specify the layout as stripe-mirror-colrather than stripe-  
mirror.  
Mirroring across targets, controllers or enclosures  
To create a volume whose mirrored data plexes lie on different controllers (also  
known as disk duplexing) or in different enclosures, use the vxassistcommand  
as described in this section.  
In the following command, the attribute mirror=targetspecifies that volumes  
should be mirrored between identical target IDs on different controllers.  
# vxassist [-b] [-g diskgroup] make volume length \  
layout=layout mirror=target [attributes]  
The attribute mirror=ctlrspecifies that disks in one mirror should not be on  
the same controller as disks in other mirrors within the same volume:  
# vxassist [-b] [-g diskgroup] make volume length \  
layout=layout mirror=ctlr [attributes]  
Note: Both paths of an active/passive array are not considered to be on different  
controllers when mirroring across controllers.  
The following command creates a mirrored volume with two data plexes in the  
disk group, mydg:  
# vxassist -b -g mydg make volspec 10g layout=mirror nmirror=2 \  
mirror=ctlr ctlr:c2 ctlr:c3  
The disks in one data plex are all attached to controller c2, and the disks in the  
other data plex are all attached to controller c3. This arrangement ensures  
continued availability of the volume should either controller fail.  
The attribute mirror=enclrspecifies that disks in one mirror should not be in  
the same enclosure as disks in other mirrors within the same volume.  
The following command creates a mirrored volume with two data plexes:  
# vxassist -b make -g mydg volspec 10g layout=mirror nmirror=2 \  
mirror=enclr enclr:enc1 enclr:enc2  
The disks in one data plex are all taken from enclosure enc1, and the disks in  
the other data plex are all taken from enclosure enc2. This arrangement  
ensures continued availability of the volume should either enclosure become  
unavailable.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
256 Creating volumes  
Creating a RAID-5 volume  
See “Specifying ordered allocation of storage to volumes” on page 245 for a  
description of other ways in which you can control how volumes are laid out on  
the specified storage.  
Creating a RAID-5 volume  
Note: VxVM supports this feature for private disk groups, but not for shareable  
disk groups in a cluster environment.  
A RAID-5 volume requires space to be available on at least as many disks in the  
disk group as the number of columns in the volume. Additional disks may be  
required for any RAID-5 logs that are created.  
You need a full license to use this feature.  
You can create RAID-5 volumes by using either the vxassistcommand  
(recommended) or the vxmakecommand. Both approaches are described below.  
A RAID-5 volume contains a RAID-5 data plex that consists of three or more  
subdisks located on three or more physical disks. Only one RAID-5 data plex can  
exist per volume. A RAID-5 volume can also contain one or more RAID-5 log  
plexes, which are used to log information about data and parity being written to  
the volume. For more information on RAID-5 volumes, see “RAID-5 (striping  
with parity)” on page 45.  
Caution: Do not create a RAID-5 volume with more than 8 columns because the  
volume will be unrecoverable in the event of the failure of more than one disk.  
To create a RAID-5 volume, use the following command:  
# vxassist [-b] [-g diskgroup] make volume length layout=raid5 \  
[ncol=number_of_columns] [stripewidth=size] [nlog=number] \  
[loglen=log_length]  
For example, to create the RAID-5 volume volraidtogether with 2 RAID-5 logs  
in the disk group, mydg, use the following command:  
# vxassist -b -g mydg make volraid 10g layout=raid5 nlog=2  
This creates a RAID-5 volume with the default stripe unit size on the default  
number of disks. It also creates two RAID-5 logs rather than the default of one  
log.  
Note: If you require RAID-5 logs, you must use the logdiskattribute to specify  
the disks to be used for the log plexes.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
Creating volumes 257  
Creating tagged volumes  
RAID-5 logs can be concatenated or striped plexes, and each RAID-5 log  
associated with a RAID-5 volume has a complete copy of the logging information  
for the volume. To support concurrent access to the RAID-5 array, the log  
should be several times the stripe size of the RAID-5 plex.  
It is suggested that you configure a minimum of two RAID-5 log plexes for each  
RAID-5 volume. These log plexes should be located on different disks. Having  
two RAID-5 log plexes for each RAID-5 volume protects against the loss of  
logging information due to the failure of a single disk.  
If you use ordered allocation when creating a RAID-5 volume on specified  
storage, you must use the logdiskattribute to specify on which disks the RAID-  
5 log plexes should be created. Use the following form of the vxassistcommand  
to specify the disks from which space for the logs is to be allocated:  
# vxassist [-b] [-g diskgroup] -o ordered make volume length \  
layout=raid5 [ncol=number_columns] [nlog=number] \  
[loglen=log_length] logdisk=disk[,disk,...] storage_attributes  
For example, the following command creates a 3-column RAID-5 volume with  
the default stripe unit size on disks mydg04, mydg05and mydg06. It also creates  
two RAID-5 logs on disks mydg07and mydg08.  
# vxassist -b -g mydg -o ordered make volraid 10g layout=raid5 \  
ncol=3 nlog=2 logdisk=mydg07,mydg08 mydg04 mydg05 mydg06  
Note: The number of logs must equal the number of disks specified to logdisk.  
For more information about ordered allocation, see “Specifying ordered  
allocation of storage to volumes” on page 245 and the vxassist(1M) manual  
page.  
If you need to add more logs to a RAID-5 volume at a later date, follow the  
procedure described in “Adding a RAID-5 log” on page 283.  
Creating tagged volumes  
Volume tags are used to implement the Dynamic Storage Tiering feature of the  
Storage Foundation software. For more information about this feature, see the  
Veritas File System Administrator’s Guide.  
You can use the tagattribute with the vxassistmakecommand to set a named  
tag and optional tag value on a volume, for example:  
# vxassist -b -g mydg make volmir 5g layout=mirror tag=mirvol=5g  
To list the tags that are associated with a volume, use this command:  
# vxassist [-g diskgroup] listtag volume  
To list the volumes that have a specified tag name, use this command:  
# vxassist [-g diskgroup] list tag=tagname volume  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
258 Creating volumes  
Creating a volume using vxmake  
Tag names and tag values are case-sensitive character strings of up to 256  
characters. Tag names can consist of letters (A through Z and a through z),  
numbers (0 through 9), dashes (-), underscores (_) or periods (.) from the ASCII  
character set. A tag name must start with either a letter or an underscore. Tag  
values can consist of any character from the ASCII character set with a decimal  
value from 32 through 127. If a tag value includes any spaces, use the vxassist  
settagcommand to set the tag on the newly created volume.  
Dotted tag hierarchies are understood by the listoperation. For example, the  
listing for tag=a.bincludes all volumes that have tag names that start with a.b.  
The tag names site, udidand vdidare reserved and should not be used. To  
avoid possible clashes with future product features, it is recommended that tag  
names do not start with any of the following strings: asl, be, isp, nbu, sf,  
symcor vx.  
See “Setting tags on volumes” on page 288.  
Creating a volume using vxmake  
As an alternative to using vxassist, you can create a volume using the vxmake  
command to arrange existing subdisks into plexes, and then to form these  
plexes into a volume. Subdisks can be created using the method described in  
“Creating subdisks” on page 215. The example given in this section is to create a  
RAID-5 volume using vxmake.  
Creating a RAID-5 plex for a RAID-5 volume is similar to creating striped plexes,  
except that the layoutattribute is set to raid5. Subdisks can be implicitly  
associated in the same way as with striped plexes. For example, to create a four-  
column RAID-5 plex with a stripe unit size of 32 sectors, use the following  
command:  
# vxmake -g mydg plex raidplex layout=raid5 stwidth=32 \  
sd=mydg00-01,mydg01-00,mydg02-00,mydg03-00  
Note that because four subdisks are specified, but the number of columns is not  
specified, the vxmakecommand assumes a four-column RAID-5 plex and places  
one subdisk in each column. Striped plexes are created using the same method  
except that the layout is specified as stripe. If the subdisks are to be created  
and added later, use the following command to create the plex:  
# vxmake -g mydg plex raidplex layout=raid5 ncolumn=4 stwidth=32  
Note: If no subdisks are specified, the ncolumnattribute must be specified.  
Subdisks can be added to the plex later using the vxsdassoccommand (see  
“Associating subdisks with plexes” on page 218).  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Creating volumes 259  
Creating a volume using vxmake  
If each column in a RAID-5 plex is to be created from multiple subdisks which  
may span several physical disks, you can specify to which column each subdisk  
should be added. For example, to create a three-column RAID-5 plex using six  
subdisks, use the following form of the vxmakecommand:  
# vxmake -g mydg plex raidplex layout=raid5 stwidth=32 \  
sd=mydg00-00:0,mydg01-00:1,mydg02-00:2,mydg03-00:0, \  
mydg04-00:1,mydg05-00:2  
This command stacks subdisks mydg00-00and mydg03-00consecutively in  
column 0, subdisks mydg01-00and mydg04-00consecutively in column 1, and  
subdisks mydg02-00and mydg05-00in column 2. Offsets can also be specified  
to create sparse RAID-5 plexes, as for striped plexes.  
Log plexes may be created as default concatenated plexes by not specifying a  
layout, for example:  
# vxmake -g mydg plex raidlog1 sd=mydg06-00  
# vxmake -g mydg plex raidlog2 sd=mydg07-00  
The following command creates a RAID-5 volume, and associates the prepared  
RAID-5 plex and RAID-5 log plexes with it:  
# vxmake -g mydg -Uraid5 vol raidvol \  
plex=raidplex,raidlog1,raidlog2  
Note: Each RAID-5 volume has one RAID-5 plex where the data and parity are  
stored. Any other plexes associated with the volume are used as RAID-5 log  
plexes to log information about data and parity being written to the volume.  
After creating a volume using vxmake, you must initialize it before it can be used.  
The procedure is described in “Initializing and starting a volume” on page 260.  
Creating a volume using a vxmake description file  
You can use the vxmakecommand to add a new volume, plex or subdisk to the set  
of objects managed by VxVM. vxmakeadds a record for each new object to the  
VxVM configuration database. You can create records either by specifying  
parameters to vxmakeon the command line, or by using a file which contains  
plain-text descriptions of the objects. The file can also contain commands for  
performing a list of tasks. Use the following form of the command to have  
vxmakeread the file from the standard input:  
# vxmake [-g diskgroup] < description_file  
Alternatively, you can specify the file to vxmakeusing the -doption:  
# vxmake [-g diskgroup] -d description_file  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
260 Creating volumes  
Initializing and starting a volume  
The following sample description file defines a volume, db, with two plexes, db-  
01and db-02:  
#rty #name  
#options  
sd  
sd  
sd  
sd  
sd  
mydg03-01 disk=mydg03 offset=0 len=10000  
mydg03-02 disk=mydg03 offset=25000 len=10480  
mydg04-01 disk=mydg04 offset=0 len=8000  
mydg04-02 disk=mydg04 offset=15000 len=8000  
mydg04-03 disk=mydg04 offset=30000 len=4480  
plex db-01  
layout=STRIPE ncolumn=2 stwidth=16k  
sd=mydg03-01:0/0,mydg03-02:0/10000,mydg04-01:1/0,  
mydg04-02:1/8000,mydg04-03:1/16000  
sd  
ramd1-01  
disk=ramd1 len=640  
comment=”Hot spot for dbvol  
sd=ramd1-01:40320  
usetype=gen plex=db-01,db-02  
readpol=prefer prefname=db-02  
comment=”Uses mem1 for hot spot in last 5m  
plex db-02  
vol db  
Note: The subdisk definition for plex, db-01, must be specified on a single line.  
It is shown here split across two lines because of space constraints.  
The first plex, db-01, is striped and has five subdisks on two physical disks,  
mydg03and mydg04. The second plex, db-02, is the preferred plex in the  
mirror, and has one subdisk, ramd1-01, on a volatile memory disk.  
For detailed information about how to use vxmake, refer to the vxmake(1M)  
manual page.  
After creating a volume using vxmake, you must initialize it before it can be used.  
The procedure is described in “Initializing and starting a volume created using  
vxmake” on page 261.  
Initializing and starting a volume  
If you create a volume using the vxassistcommand, vxassistinitializes and  
starts the volume automatically unless you specify the attribute init=none.  
When creating a volume, you can make it immediately available for use by  
specifying the -boption to the vxassistcommand, as shown here:  
# vxassist -b [-g diskgroup] make volume length layout=mirror  
The -boption makes VxVM carry out any required initialization as a  
background task. It also greatly speeds up the creation of striped volumes by  
initializing the columns in parallel.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Creating volumes 261  
Initializing and starting a volume  
As an alternative to the -boption, you can specify the init=activeattribute to  
make a new volume immediately available for use. In this example, init=active  
is specified to prevent VxVM from synchronizing the empty data plexes of a new  
mirrored volume:  
# vxassist [-g diskgroup] make volume length layout=mirror \  
init=active  
Caution: There is a very small risk of errors occurring when the init=active  
attribute is used. Although written blocks are guaranteed to be consistent, read  
errors can arise in the unlikely event that fsckattempts to verify uninitialized  
space in the file system, or if a file remains uninitialized following a system  
crash. If in doubt, use the -boption to vxassistinstead.  
This command writes zeroes to the entire length of the volume and to any log  
plexes. It then makes the volume active. You can also zero out a volume by  
specifying the attribute init=zeroto vxassist, as shown in this example:  
# vxassist [-g diskgroup] make volume length layout=raid5 \  
init=zero  
Note: You cannot use the -boption to make this operation a background task.  
Initializing and starting a volume created using vxmake  
A volume may be initialized by running the vxvolcommand if the volume was  
created by the vxmakecommand and has not yet been initialized, or if the  
volume has been set to an uninitialized state.  
To initialize and start a volume, use the following command:  
# vxvol [-g diskgroup] start volume  
The following command can be used to enable a volume without initializing it:  
# vxvol [-g diskgroup] init enable volume  
This allows you to restore data on the volume from a backup before using the  
following command to make the volume fully active:  
# vxvol [-g diskgroup] init active volume  
If you want to zero out the contents of an entire volume, use this command to  
initialize it:  
# vxvol [-g diskgroup] init zero volume  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
262 Creating volumes  
Accessing a volume  
Accessing a volume  
As soon as a volume has been created and initialized, it is available for use as a  
virtual disk partition by the operating system for the creation of a file system, or  
by application programs such as relational databases and other data  
management software.  
Creating a volume in a disk group sets up block and character (raw) device files  
that can be used to access the volume:  
/dev/vx/dsk/diskgroup/volume  
/dev/vx/rdsk/diskgroup/volume  
block device file for volume  
character device file for volume  
The pathnames include a directory named for the disk group. Use the  
appropriate device node to create, mount and repair file systems, and to lay out  
databases that require raw partitions.  
Note: As the rootdgdisk group no longer has special significance, VxVM only  
creates volume device nodes for this disk group in the /dev/vx/dsk/rootdg  
and /dev/vx/rdsk/rootdgdirectories. VxVM does not create device nodes in  
the /dev/vx/dskor /dev/vx/rdskdirectories for the rootdgdisk group.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Chapter  
8
Administering volumes  
This chapter describes how to perform common maintenance tasks on volumes  
in Veritas Volume Manager (VxVM). This includes displaying volume  
information, monitoring tasks, adding and removing logs, resizing volumes,  
removing mirrors, removing volumes, and changing the layout of volumes  
without taking them offline.  
Note: You can also use the Veritas Intelligent Storage Provisioning (ISP) feature  
to create and administer application volumes. These volumes are very similar to  
the traditional VxVM volumes that are described in this chapter. However, there  
are significant differences between the functionality of the two types of volumes  
that prevents them from being used interchangeably. Refer to the Veritas  
Storage Foundation Intelligent Storage Provisioning Administrator’s Guide for  
more information about creating and administering ISP application volumes.  
Most VxVM commands require superuser or equivalent privileges.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
264 Administering volumes  
Displaying volume information  
Displaying volume information  
You can use the vxprintcommand to display information about how a volume  
is configured.  
To display the volume, plex, and subdisk record information for all volumes in  
the system, use the following command:  
# vxprint -hvt  
The vxprintcommand can also be applied to a single disk group:  
# vxprint -g mydg -hvt  
This is example output from this command:  
V NAME  
PL NAME  
SD NAME  
SV NAME  
SC NAME  
DC NAME  
SP NAME  
RVG/VSET/COKSTATE  
STATE  
STATE  
DISKOFFS LENGTH  
LENGTH  
LENGTH  
READPOL  
LAYOUT  
[COL/]OFF DEVICE  
[COL/]OFF AM/NM  
[COL/]OFF DEVICE  
PREFPLEX UTYPE  
NCOL/WID MODE  
VOLUME  
PLEX  
KSTATE  
DISK  
MODE  
MODE  
MODE  
PLEX  
PLEX  
VOLNAME NVOLLAYR LENGTH  
CACHE DISKOFFS LENGTH  
PARENTVOL LOGVOL  
SNAPVOL  
DCO  
v pubs  
pl pubs-01 pubs  
sd mydg11-01pubs-01  
-
ENABLED ACTIVE  
ENABLED ACTIVE  
22880  
22880  
22880  
SELECT  
CONCAT  
0
-
-
fsgen  
RW  
ENA  
mydg11  
0
c1t0d0  
c1t1d0  
v voldef  
pl voldef-01voldef  
sd mydg12-02voldef-0  
-
ENABLED ACTIVE  
ENABLED ACTIVE  
20480  
20480  
20480  
SELECT  
CONCAT  
0
-
-
fsgen  
RW  
ENA  
mydg12  
0
Here vis a volume, plis a plex, and sdis a subdisk. The top few lines indicate  
the headers that match each type of output line that follows. Each volume is  
listed along with its associated plexes and subdisks.  
Note: The headings for sub-volumes (SV), storage caches (SC), data change  
objects (DCO) and snappoints (SP) can be ignored here. No such objects are  
associated with these volumes.  
To display volume-related information for a specific volume, use the following  
command:  
# vxprint [-g diskgroup] -t volume  
For example, to display information about the volume, voldef, in the disk  
group, mydg, use the following command:  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Administering volumes 265  
Displaying volume information  
# vxprint -g mydg -t voldef  
This is example output from this command:  
V NAME  
RVG/VSET/COKSTATE  
STATE  
LENGTH  
20480  
READPOL  
SELECT  
PREFPLEX UTYPE  
v voldef  
-
ENABLED ACTIVE  
-
fsgen  
Note: If you enable enclosure-based naming, and use the vxprintcommand to  
display the structure of a volume, it shows enclosure-based disk device names  
(disk access names) rather than c#t#d# names. See “Discovering the association  
between enclosure and OS based disk names” on page 94 for information on how  
to obtain the true device names.  
The following section describes the meaning of the various volume states that  
may be displayed.  
Volume states  
The following volume states may be displayed by VxVM commands such as  
vxprint:  
ACTIVE volume state  
The volume has been started (kernel state is currently ENABLED) or was in use  
(kernel state was ENABLED) when the machine was rebooted. If the volume is  
currently ENABLED, the state of its plexes at any moment is not certain (since  
the volume is in use).  
If the volume is currently DISABLED, this means that the plexes cannot be  
guaranteed to be consistent, but are made consistent when the volume is  
started.  
For a RAID-5 volume, if the volume is currently DISABLED, parity cannot be  
guaranteed to be synchronized.  
CLEAN volume state  
The volume is not started (kernel state is DISABLED) and its plexes are  
synchronized. For a RAID-5 volume, its plex stripes are consistent and its parity  
is good.  
EMPTY volume state  
The volume contents are not initialized. The kernel state is always DISABLED  
when the volume is EMPTY.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
266 Administering volumes  
Displaying volume information  
INVALID volume state  
The contents of an instant snapshot volume no longer represent a true point-in-  
time image of the original volume.  
NEEDSYNC volume state  
The volume requires a resynchronization operation the next time it is started.  
For a RAID-5 volume, a parity resynchronization operation is required.  
REPLAY volume state  
The volume is in a transient state as part of a log replay. A log replay occurs  
when it becomes necessary to use logged parity and data. This state is only  
applied to RAID-5 volumes.  
SYNC volume state  
The volume is either in read-writeback recovery mode (kernel state is currently  
ENABLED) or was in read-writeback mode when the machine was rebooted  
(kernel state is DISABLED). With read-writeback recovery, plex consistency is  
recovered by reading data from blocks of one plex and writing the data to all  
other writable plexes. If the volume is ENABLED, this means that the plexes are  
being resynchronized through the read-writeback recovery. If the volume is  
DISABLED, it means that the plexes were being resynchronized through read-  
writeback when the machine rebooted and therefore still need to be  
synchronized.  
For a RAID-5 volume, the volume is either undergoing a parity  
resynchronization (kernel state is currently ENABLED) or was having its parity  
resynchronized when the machine was rebooted (kernel state is DISABLED).  
Note: The interpretation of these flags during volume startup is modified by the  
persistent state log for the volume (for example, the DIRTY/CLEAN flag). If the  
clean flag is set, an ACTIVE volume was not written to by any processes or was  
not even open at the time of the reboot; therefore, it can be considered CLEAN.  
The clean flag is always set in any case where the volume is marked CLEAN.  
Volume kernel states  
The volume kernel state indicates the accessibility of the volume. The volume  
kernel state allows a volume to have an offline (DISABLED), maintenance  
(DETACHED), or online (ENABLED) mode of operation.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Administering volumes 267  
Monitoring and controlling tasks  
Note: No user intervention is required to set these states; they are maintained  
internally. On a system that is operating properly, all volumes are ENABLED.  
The following volume kernel states are defined:  
DETACHED volume kernel state  
Maintenance is being performed on the volume. The volume cannot be read  
from or written to, but certain plex operations and ioctlfunction calls are  
accepted.  
DISABLED volume kernel state  
The volume is offline and cannot be accessed.  
ENABLED volume kernel state  
The volume is online and can be read from or written to.  
Monitoring and controlling tasks  
Note: VxVM supports this feature for private disk groups, but not for shareable  
disk groups in a cluster environment.  
The VxVM task monitor tracks the progress of system recovery by monitoring  
task creation, maintenance, and completion. The task monitor allows you to  
monitor task progress and to modify characteristics of tasks, such as pausing  
and recovery rate (for example, to reduce the impact on system performance).  
Specifying task tags  
Every task is given a unique task identifier. This is a numeric identifier for the  
task that can be specified to the vxtaskutility to specifically identify a single  
task. Several VxVM utilities also provide a -toption to specify an alphanumeric  
tag of up to 16 characters in length. This allows you to group several tasks by  
associating them with the same tag.  
The vxassist, vxevac, vxplex, vxmirror, vxrecover, vxrelayout, vxresize,  
vxsd, and vxvolutilities allow you to specify a tag using the -toption. For  
example, to execute a vxrecovercommand and track all the resulting tasks as a  
group with the task tag myrecovery, use the following command:  
# vxrecover -g mydg -t myrecovery -b mydg05  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                 
268 Administering volumes  
Monitoring and controlling tasks  
Any tasks started by the utilities invoked by vxrecoveralso inherit its task ID  
and task tag, so establishing a parent-child task relationship.  
For more information about the utilities that support task tagging, see their  
respective manual pages.  
Managing tasks with vxtask  
Note: New tasks take time to be set up, and so may not be immediately available  
for use after a command is invoked. Any script that operates on tasks may need  
to poll for the existence of a new task.  
You can use the vxtaskcommand to administer operations on VxVM tasks that  
are running on the system. Operations include listing tasks, modifying the state  
of a task (pausing, resuming, aborting) and modifying the rate of progress of a  
task. For detailed information about how to use vxtask, refer to the vxtask(1M)  
manual page.  
VxVM tasks represent long-term operations in progress on the system. Every  
task gives information on the time the operation started, the size and progress  
of the operation, and the state and rate of progress of the operation. The  
administrator can change the state of a task, giving coarse-grained control over  
the progress of the operation. For those operations that support it, the rate of  
progress of the task can be changed, giving more fine-grained control over the  
task.  
vxtask operations  
The vxtaskcommand supports the following operations:  
abort  
Causes the specified task to cease operation. In most cases, the  
operations “back out” as if an I/O error occurred, reversing what  
has been done so far to the largest extent possible.  
list  
Lists tasks running on the system in one-line summaries. The -l  
option prints tasks in long format. The -hoption prints tasks  
hierarchically, with child tasks following the parent tasks. By  
default, all tasks running on the system are printed. If a taskid  
argument is supplied, the output is limited to those tasks whose  
taskidor task tag match taskid. The remaining arguments are  
used to filter tasks and limit the tasks actually listed.  
monitor  
Prints information continuously about a task or group of tasks as  
task information changes. This allows you to track the progression  
of tasks. Specifying -lcauses a long listing to be printed. By  
default, short one-line listings are printed. In addition to printing  
task information when a task state changes, output is also  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Administering volumes 269  
Monitoring and controlling tasks  
generated when the task completes. When this occurs, the state of  
the task is printed as EXITED.  
pause  
Puts a running task in the paused state, causing it to suspend  
operation.  
resume  
set  
Causes a paused task to continue operation.  
Changes modifiable parameters of a task. Currently, there is only  
one modifiable parameter, slow[=iodelay], which can be used to  
reduce the impact that copy operations have on system  
performance. If slowis specified, this introduces a delay between  
such operations with a default value for iodelay of 250  
milliseconds. The larger the value of iodelay that is specified, the  
slower is the progress of the task and the fewer system resources  
that it consumes in a given time. (The slowattribute is also  
accepted by the vxplex, vxvoland vxrecovercommands.)  
Using the vxtask command  
To list all tasks currently running on the system, use the following command:  
# vxtask list  
To print tasks hierarchically, with child tasks following the parent tasks, specify  
the -hoption, as follows:  
# vxtask -h list  
To trace all tasks in the disk group, foodg, that are currently paused, as well as  
any tasks with the tag sysstart, use the following command:  
# vxtask -g foodg -p -i sysstart list  
Use the vxtask -p listcommand lists all paused tasks, and use vxtaskresume  
to continue execution (the task may be specified by its ID or by its tag):  
# vxtask -p list  
# vxtask resume 167  
To monitor all tasks with the tag myoperation, use the following command:  
# vxtask monitor myoperation  
To cause all tasks tagged with recovallto exit, use the following command:  
# vxtask abort recovall  
This command causes VxVM to attempt to reverse the progress of the operation  
so far. For an example of how to use vxtaskto monitor and modify the progress  
of the Online Relayout feature, see “Controlling the progress of a relayout” on  
page 299.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
               
270 Administering volumes  
Stopping a volume  
Stopping a volume  
Stopping a volume renders it unavailable to the user, and changes the volume  
kernel state from ENABLED or DETACHED to DISABLED. If the volume cannot  
be disabled, it remains in its current state. To stop a volume, use the following  
command:  
# vxvol [-g diskgroup] [-f] stop volume ...  
To stop all volumes in a specified disk group, use the following command:  
# vxvol [-g diskgroup] [-f] stopall  
Caution: If you use the -foption to forcibly disable a volume that is currently  
open to an application, the volume remains open, but its contents are  
inaccessible. I/O operations on the volume fail, and this may cause data loss. It is  
not possible to deport a disk group until all of its volumes are closed.  
If you need to prevent a closed volume from being opened, it is recommended  
that you use the vxvolmaintcommand, as described in the following section.  
Putting a volume in maintenance mode  
If all mirrors of a volume become STALE, you can place the volume in  
maintenance mode. Then you can view the plexes while the volume is DETACHED  
and determine which plex to use for reviving the others. To place a volume in  
maintenance mode, use the following command:  
# vxvol [-g diskgroup] maint volume  
To assist in choosing the revival source plex, use vxprintto list the stopped  
volume and its plexes.  
To take a plex (in this example, vol01-02in the disk group, mydg) offline, use  
the following command:  
# vxmend -g mydg off vol01-02  
The vxmendoncommand can change the state of an OFFLINEplex of a  
DISABLEDvolume to STALE. For example, to put a plex named vol01-02in the  
STALE state, use the following command:  
# vxmend -g mydg on vol01-02  
Running the vxvolstartcommand on the volume then revives the plex as  
described in the next section.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                   
Administering volumes 271  
Starting a volume  
Starting a volume  
Starting a volume makes it available for use, and changes the volume state from  
DISABLED or DETACHED to ENABLED. To start a DISABLED or DETACHED  
volume, use the following command:  
# vxvol [-g diskgroup] start volume ...  
If a volume cannot be enabled, it remains in its current state.  
To start all DISABLED or DETACHED volumes in a disk group, enter:  
# vxvol -g diskgroup startall  
Alternatively, to start a DISABLED volume, use the following command:  
# vxrecover -g diskgroup -s volume ...  
To start all DISABLED volumes, enter:  
# vxrecover -s  
To prevent any recovery operations from being performed on the volumes,  
additionally specify the -noption to vxrecover.  
Adding a mirror to a volume  
A mirror can be added to an existing volume with the vxassistcommand, as  
follows:  
# vxassist [-b] [-g diskgroup] mirror volume  
Note: If specified, the -boption makes synchronizing the new mirror a  
background task.  
For example, to create a mirror of the volume voltestin the disk group, mydg,  
use the following command:  
# vxassist -b -g mydg mirror voltest  
Another way to mirror an existing volume is by first creating a plex, and then  
attaching it to a volume, using the following commands:  
# vxmake [-g diskgroup] plex plex sd=subdisk ...  
# vxplex [-g diskgroup] att volumeplex  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                     
272 Administering volumes  
Adding a mirror to a volume  
Mirroring all volumes  
To mirror all volumes in a disk group to available disk space, use the following  
command:  
# /etc/vx/bin/vxmirror -g diskgroup -a  
To configure VxVM to create mirrored volumes by default, use the following  
command:  
# /etc/vx/bin/vxmirror -d yes  
If you make this change, you can still make unmirrored volumes by specifying  
nmirror=1as an attribute to the vxassistcommand. For example, to create an  
unmirrored 20-gigabyte volume named nomirrorin the disk group, mydg, use  
the following command:  
# vxassist -g mydg make nomirror 20g nmirror=1  
Mirroring volumes on a VM disk  
Mirroring volumes on a VM disk gives you one or more copies of your volumes in  
another disk location. By creating mirror copies of your volumes, you protect  
your system against loss of data in case of a disk failure.  
Note: This task only mirrors concatenated volumes. Volumes that are already  
mirrored or that contain subdisks that reside on multiple disks are ignored.  
To mirror volumes on a disk  
1
2
3
Make sure that the target disk has an equal or greater amount of space as  
the originating disk.  
Select menu item 5 (Mirror volumes on a disk)from the  
vxdiskadmmain menu.  
At the following prompt, enter the disk name of the disk that you wish to  
mirror:  
Mirror volumes on a disk  
Menu: VolumeManager/Disk/Mirror  
This operation can be used to mirror volumes on a disk. These  
volumes can be mirrored onto another disk or onto any  
available disk space. Volumes will not be mirrored if they are  
already mirrored. Also, volumes that are comprised of more  
than one subdisk will not be mirrored.  
Enter disk name [<disk>,list,q,?] mydg02  
4
At the following prompt, enter the target disk name (this disk must be the  
same size or larger than the originating disk):  
Download from Www.Somanuals.com. All Manuals Search And Download.  
               
Administering volumes 273  
Removing a mirror  
You can choose to mirror volumes from disk mydg02 onto any  
available disk space, or you can choose to mirror onto a  
specific disk. To mirror to a specific disk, select the name of  
that disk. To mirror to any available disk space, select  
"any".  
Enter destination disk [<disk>,list,q,?] (default: any) mydg01  
5
At the following prompt, press Return to make the mirror:  
The requested operation is to mirror all volumes on disk  
mydg02 in disk group mydg onto available disk space on disk  
mydg01.  
VxVM NOTICE V-5-2-229 This operation can take a long time to  
complete.  
Continue with operation? [y,n,q,?] (default: y)  
The vxdiskadmprogram displays the status of the mirroring operation, as  
follows:  
VxVM vxmirror INFO V-5-2-22 Mirror volume voltest-bk00 ...  
VxVM INFO V-5-2-674 Mirroring of disk mydg01 is complete.  
6
At the following prompt, indicate whether you want to mirror volumes on  
another disk (y) or return to the vxdiskadmmain menu (n):  
Mirror volumes on another disk? [y,n,q,?] (default: n)  
Removing a mirror  
When a mirror is no longer needed, you can remove it to free up disk space.  
Note: The last valid plex associated with a volume cannot be removed.  
To remove a mirror from a volume, use the following command:  
# vxassist [-g diskgroup] remove mirror volume  
Additionally, you can use storage attributes to specify the storage to be  
removed. For example, to remove a mirror on disk mydg01from volume vol01,  
enter:  
# vxassist -g mydg remove mirror vol01 !mydg01  
For more information about storage attributes, see “Creating a volume on  
specific disks” on page 244.  
Alternatively, use the following command to dissociate and remove a mirror  
from a volume:  
# vxplex [-g diskgroup] -o rm dis plex  
For example, to dissociate and remove a mirror named vol01-02from the disk  
group, mydg, use the following command:  
# vxplex -g mydg -o rm dis vol01-02  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
274 Administering volumes  
Adding logs and maps to volumes  
This command removes the mirror vol01-02and all associated subdisks. This  
is equivalent to entering the following separate commands:  
# vxplex -g mydg dis vol01-02  
# vxedit -g mydg -r rm vol01-02  
Adding logs and maps to volumes  
In Veritas Volume Manager, several types of volume logs and maps are  
supported:  
■
FastResync Maps are used to perform quick and efficient resynchronization  
of mirrors (see “FastResync” on page 66 for details). These maps are  
supported either in memory (Non-Persistent FastResync), or on disk as part  
of a DCO volume (Persistent FastResync). Two types of DCO volumes are  
supported:  
■
Version 0 DCO volumes only support Persistent FastResync for the  
traditional third-mirror break-off type of volume snapshot. See  
“Version 0 DCO volume layout” on page 69, and “Adding a version 0  
DCO and DCO volume” on page 356 for more information.  
■
Version 20 DCO volumes, introduced in VxVM 4.0, support DRL logging  
(see below) and Persistent FastResync for full-sized and space-  
optimized instant volume snapshots. See “Version 20 DCO volume  
layout” on page 69, and “Preparing a volume for DRL and instant  
snapshots” on page 275 for more information.  
See “Enabling FastResync on a volume” on page 292 for information on how  
to enable Persistent or Non-Persistent FastResync on a volume.  
■
Dirty Region Logs allow the fast recovery of mirrored volumes after a system  
crash (see “Dirty region logging” on page 60 for details). These logs are  
supported either as DRL log plexes, or as part of a version 20 DCO volume.  
Refer to the following sections for information on enabling DRL on a  
volume:  
■
“Adding traditional DRL logging to a mirrored volume” on page 281  
describes how to add DRL log plexes to a volume.  
■
“Preparing a volume for DRL and instant snapshots” on page 275  
describes how to configure DRL for a volume using a version 20 DCO  
volume.  
■
RAID-5 logs are used to prevent corruption of data during recovery of RAID-  
5 volumes (see “RAID-5 logging” on page 50 for details). These logs are  
configured as plexes on disks other than those that are used for the columns  
of the RAID-5 volume.  
See “Adding a RAID-5 log” on page 283 for information on adding RAID-5  
logs to a RAID-5 volume.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Administering volumes 275  
Preparing a volume for DRL and instant snapshots  
Preparing a volume for DRL and instant snapshots  
Note: This procedure describes how to add a version 20 data change object (DCO)  
and DCO volume to a volume that you previously created in a disk group with a  
version number of 110 or greater. If you are creating a new volume in a disk  
group with a version number of 110 or greater, you can specify the co-creation  
of a DCO and DCO volume and enable DRL as described in “Creating a volume  
with a version 20 DCO volume” on page 252. If the volume was created in a  
release prior to VxVM 4.0, use the procedure in “Upgrading existing volumes to  
use version 20 DCOs” on page 279.  
TM  
You need a full VxVM license and a Veritas FlashSnap or FastResync license to  
use the DRL and FastResync features. Even if you do not have a license, you can  
configure a DCO object and DCO volume so that snap objects are associated with  
the original and snapshot volumes. For more information about snap objects,  
see “How persistent FastResync works with snapshots” on page 70. See  
“Determining the DCO version number” on page 277 for details of how to  
determine the version number of a volume’s DCO.  
Use the following command to add a version 20 DCO and DCO volume to a  
volume:  
# vxsnap [-g diskgroup] prepare volume [ndcomirs=number] \  
[regionsize=size] [drl=on|sequential|off] \  
[storage_attribute ...]  
The ndcomirsattribute specifies the number of DCO plexes that are created in  
the DCO volume. It is recommended that you configure as many DCO plexes as  
there are data and snapshot plexes in the volume. The DCO plexes are used to set  
up a DCO volume for any snapshot volume that you subsequently create from  
the snapshot plexes. For example, specify ndcomirs=5for a volume with 3 data  
plexes and 2 snapshot plexes.  
The value of the regionsizeattribute specifies the size of the tracked regions  
in the volume. A write to a region is tracked by setting a bit in the change map.  
The default value is 64k(64KB). A smaller value requires more disk space for the  
change maps, but the finer granularity provides faster resynchronization.  
To enable DRL logging on the volume, specify drl=on(this is the default setting).  
If sequential DRL is required, specify drl=sequential. If DRL is not required,  
specify drl=off.  
You can also specify storage attributes to define the disks that can and/or  
cannot be used for the plexes of the DCO volume. See “Specifying storage for  
version 20 DCO plexes” on page 276 for details.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
276 Administering volumes  
Preparing a volume for DRL and instant snapshots  
Note: The vxsnappreparecommand automatically enables Persistent  
FastResync on the volume. Persistent FastResync is also set automatically on  
any snapshots that are generated from a volume on which this feature is  
enabled.  
If the volume is a RAID-5 volume, it is converted to a layered volume that can be  
used with instant snapshots and Persistent FastResync. See “Using a DCO and  
DCO volume with a RAID-5 volume” on page 277 for details.  
By default, a version 20 DCO volume contains 32 per-volume maps. If you  
require more maps than this, you can use the vxsnapaddmapcommand to add  
more maps. See the vxsnap(1M) manual page for details of this command.  
Specifying storage for version 20 DCO plexes  
If the disks that contain volumes and their snapshots are to be moved into  
different disk groups, you must ensure that the disks that contain their DCO  
plexes can accompany them. You can use storage attributes to specify which  
disks to use for the DCO plexes. (If you do not want to use dirty region logging  
(DRL) with a volume, you can specify the same disks as those on which the  
volume is configured, assuming that space is available on the disks). For  
example, to add a DCO object and mirrored DCO volume with plexes on disk05  
and disk06to the volume, myvol, use the following command:  
# vxsnap -g mydg prepare myvol ndcomirs=2 alloc=disk05,disk06  
To view the details of the DCO object and DCO volume that are associated with a  
volume, use the vxprintcommand. The following is example vxprint-vh  
output for the volume named vol1(the TUTIL0 and PUTIL0 columns are  
omitted for clarity):  
TY  
v
NAME  
vol1  
ASSOC  
fsgen  
vol1  
vol1-01  
vol1  
vol1-02  
vol1  
gen  
KSTATE  
LENGTH PLOFFS STATE  
...  
ENABLED 1024  
ENABLED 1024  
ENABLED 1024  
ENABLED 1024  
ENABLED 1024  
-
-
0
-
0
-
-
-
0
-
0
ACTIVE  
ACTIVE  
-
ACTIVE  
-
-
ACTIVE  
ACTIVE  
-
ACTIVE  
-
pl  
sd  
pl  
sd  
dc  
v
pl  
sd  
pl  
sd  
vol1-01  
disk01-01  
foo-02  
disk02-01  
vol1_dco  
vol1_dcl  
-
-
ENABLED 132  
ENABLED 132  
vol1_dcl-01 vol1_dcl  
disk03-01  
vol1_dcl-02 vol1_dcl  
disk04-01 vol1_dcl-02ENABLED 132  
vol1_dcl-01ENABLED 132  
ENABLED 132  
In this output, the DCO object is shown as vol1_dco, and the DCO volume as  
vol1_dclwith 2 plexes, vol1_dcl-01and vol1_dcl-02.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Administering volumes 277  
Preparing a volume for DRL and instant snapshots  
If required, you can use the vxassistmovecommand to relocate DCO plexes to  
different disks. For example, the following command moves the plexes of the  
DCO volume, vol1_dcl, for volume vol1from disk03and disk04to disk07  
and disk08:  
# vxassist -g mydg move vol1_dcl !disk03 !disk04 disk07 disk08  
For more information, see “Moving DCO volumes between disk groups” on  
page 200, and the vxassist(1M) and vxsnap(1M) manual pages.  
Using a DCO and DCO volume with a RAID-5 volume  
The procedure in the previous section can be used to add a DCO and DCO volume  
to a RAID-5 volume. This allows you to use Persistent FastResync on the volume  
for fast resynchronization of snapshots on returning them to their original  
volume. However, the procedure has the side effect of converting the RAID-5  
volume into a special type of layered volume. You can create space-optimized  
instant snapshots of such a volume, and you can add mirrors that may be broken  
off as full-sized instant snapshots. You cannot relayout or resize such a volume  
unless you convert it back to a pure RAID-5 volume.  
To convert a volume back to a RAID-5 volume, remove any snapshot plexes from  
the volume, and dissociate the DCO and DCO volume from the layered volume  
using the procedure described in “Removing support for DRL and instant  
snapshots from a volume” on page 279. You can then perform relayout and  
resize operations on the resulting non-layered RAID-5 volume.  
To allow Persistent FastResync to be used with the RAID-5 volume again, re-  
associate the DCO and DCO volume as described in “Preparing a volume for DRL  
and instant snapshots” on page 275.  
Note: Dissociating a DCO and DCO volume disables FastResync on the volume. A  
full resynchronization of any remaining snapshots is required when they are  
snapped back.  
Determining the DCO version number  
The instant snapshot and DRL-enabled DCO features require that a version 20  
DCO be associated with a volume, rather than an earlier version 0 DCO.  
To find out the version number of a DCO that is associated with a volume  
1
Use the vxprintcommand on the volume to discover the name of its DCO:  
# DCONAME=‘vxprint [-g diskgroup] -F%dco_name volume‘  
2
Use the vxprintcommand on the DCO to determine its version number:  
# vxprint [-g diskgroup] -F%version $DCONAME  
Download from Www.Somanuals.com. All Manuals Search And Download.  
               
278 Administering volumes  
Preparing a volume for DRL and instant snapshots  
Determining if DRL is enabled on a volume  
To determine if DRL (configured using a version 20 DCO volume) is enabled on  
a volume  
1
Use the vxprintcommand on the volume to discover the name of its DCO:  
# DCONAME=‘vxprint [-g diskgroup] -F%dco_name volume‘  
2
To determine if DRL is enabled on the volume, use the following command  
with the volume’s DCO:  
# vxprint [-g diskgroup] -F%drl $DCONAME  
DRL is enabled if this command displays on.  
3
If DRL is enabled, use this command with the DCO to determine if sequential  
DRL is enabled:  
# vxprint [-g diskgroup] -F%sequentialdrl $DCONAME  
Sequential DRL is enabled if this command displays on.  
Alternatively, you can use this command with the volume:  
# vxprint [-g diskgroup] -F%log_type volume  
This displays the logging type as REGIONfor DRL, DRLSEQfor sequential DRL,  
or NONEif DRL is not enabled.  
Note: If the number of active mirrors in the volume is less than 2, DRL logging is  
not performed even if DRL is enabled on the volume. To find out if DRL logging  
is active, see “Determining if DRL logging is active on a volume” on page 278.  
Determining if DRL logging is active on a volume  
To determine if DRL logging is active on a mirrored volume  
1
Use the following vxprintcommands to discover the name of the volume’s  
DCO volume:  
# DCONAME=‘vxprint [-g diskgroup] -F%dco_name volume‘  
# DCOVOL=‘vxprint [-g diskgroup] -F%parent_vol $DCONAME‘  
2
Use the vxprintcommand on the DCO volume to find out if DRL logging is  
active:  
# vxprint [-g diskgroup] -F%drllogging $DCOVOL  
This command returns onif DRL logging is enabled.  
Disabling and re-enabling DRL  
To disable DRL (configured using a version 20 DCO volume) on a volume  
# vxvol [-g diskgroup] set drl=off volume  
Download from Www.Somanuals.com. All Manuals Search And Download.  
               
Administering volumes 279  
Upgrading existing volumes to use version 20 DCOs  
To re-enable DRL on a volume, enter this command:  
# vxvol [-g diskgroup] set drl=on volume  
To re-enable sequential DRL on a volume, enter:  
# vxvol [-g diskgroup] set drl=sequential volume  
You can use these commands to change the DRL policy on a volume by first  
disabling and then re-enabling DRL as required. DRL is automatically disabled if  
a data change map (DCM, used with Veritas Volume Replicator) is attached to a  
volume.  
Removing support for DRL and instant snapshots from a volume  
To remove support for DRL and instant snapshot operation from a volume, use  
the following command to remove the DCO and DCO volume that are associated  
with the volume:  
# vxsnap [-g diskgroup] unprepare volume  
This command also has the effect of disabling FastResync tracking on the  
volume.  
Note: This command fails if the volume is part of a snapshot hierarchy.  
Upgrading existing volumes to use version 20 DCOs  
The procedure described in this section describes how to upgrade a volume  
created before VxVM 4.0 so that it can take advantage of new features such as  
instant snapshots, and DRL logs that are configured within the DCO volume.  
This requires upgrading the version of the disk groups, removing any existing  
snapshots and version 0 DCOs that are associated with volumes in the disk  
groups, and finally configuring the volumes with version 20 DCOs.  
Note: The plexes of the DCO volume require persistent storage space on disk to  
be available. To make room for the DCO plexes, you may need to add extra disks  
to the disk group, or reconfigure existing volumes to free up space in the disk  
group. Another way to add disk space is to use the disk group move feature to  
bring in spare disks from a different disk group. For more information, see  
“Reorganizing the contents of disk groups” on page 195.  
To upgrade an existing disk group and the volumes that it contains  
1
Upgrade the disk group that contains the volume to the latest version before  
performing the remainder of the procedure described in this section. Use  
the following command to check the version of a disk group:  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
280 Administering volumes  
Upgrading existing volumes to use version 20 DCOs  
# vxdg list diskgroup  
To upgrade a disk group to the latest version, use the following command:  
# vxdg upgrade diskgroup  
For more information, see “Upgrading a disk group” on page 208.  
2
Use the following command to discover which volumes in the disk group  
have version 0 DCOs associated with them:  
# vxprint [-g diskgroup] -F “%name” -e “v_hasdcolog”  
Note: This command assumes that the volumes can only have version 0  
DCOs as the disk group has just been upgraded. See “Determining the DCO  
version number” on page 277 for a description of how to find out the DCO  
version number of a volume in any disk group.  
Repeat the following steps to upgrade each volume within the disk group as  
required.  
3
4
If the volume to be upgraded has a traditional DRL plex or subdisk (that is,  
the DRL logs are not held in a version 20 DCO volume), use the following  
command to remove this:  
# vxassist [-g diskgroup] remove log volume [nlog=n]  
Use the optional attribute nlog=nto specify the number, n, of logs to be  
removed. By default, the vxassistcommand removes one log.  
For a volume that has one or more associated snapshot volumes, use the  
following command to reattach and resynchronize each snapshot:  
# vxassist [-g diskgroup] snapback snapvol  
If FastResync was enabled on the volume before the snapshot was taken,  
the data in the snapshot plexes is quickly resynchronized from the original  
volume. If FastResync was not enabled, a full resynchronization is  
performed.  
5
6
Use the following command to turn off FastResync for the volume:  
# vxvol [-g diskgroup] set fastresync=off volume  
Use the following command to dissociate a version 0 DCO object, DCO  
volume and snap objects from the volume:  
# vxassist [-g diskgroup] remove log volume logtype=dco  
7
Use the following command on the volume to upgrade it:  
# vxsnap [-g diskgroup] prepare volume [ndcomirs=number] \  
[regionsize=size] [drl=on|sequential|off] \  
[storage_attribute ...]  
The ndcomirsattribute specifies the number of DCO plexes that are  
created in the DCO volume. It is recommended that you configure as many  
DCO plexes as there are data and snapshot plexes in the volume. The DCO  
plexes are used to set up a DCO volume for any snapshot volume that you  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Administering volumes 281  
Adding traditional DRL logging to a mirrored volume  
subsequently create from the snapshot plexes. For example, specify  
ndcomirs=5for a volume with 3 data plexes and 2 snapshot plexes.  
The value of the regionsizeattribute specifies the size of the tracked  
regions in the volume. A write to a region is tracked by setting a bit in the  
change map. The default value is 64k(64KB). A smaller value requires more  
disk space for the change maps, but the finer granularity provides faster  
resynchronization.  
To enable DRL logging on the volume, specify drl=on(this is the default  
setting). If sequential DRL is required, specify drl=sequential.If DRL is not  
required, specify drl=off.  
You can also specify vxassist-style storage attributes to define the disks  
that can or cannot be used for the plexes of the DCO volume.  
Note: The vxsnappreparecommand automatically enables FastResync on the  
volume and on any snapshots that are generated from it.  
If the volume is a RAID-5 volume, it is converted to a layered volume that can be  
used with snapshots and FastResync.  
Adding traditional DRL logging to a mirrored volume  
Note: The procedure described in this section creates a DRL log that is  
configured within a dedicated DRL plex. The version 20 DCO volume layout  
includes space for a DRL log. The new DCO volume layout also supports  
traditional (third-mirror), instant (copy-on-write), and instant space-optimized  
snapshots. However, a version 20 DCO volume cannot be used in conjunction  
with a separate DRL plex. For full details, see “Preparing a volume for DRL and  
instant snapshots” on page 275.  
To put dirty region logging (DRL) into effect for a mirrored volume, a log subdisk  
must be added to that volume. Only one log subdisk can exist per plex.  
To add DRL logs to an existing volume, use the following command:  
# vxassist [-b] [-g diskgroup] addlog volume logtype=drl \  
[nlog=n] [loglen=size]  
Note: If specified, the -boption makes adding the new logs a background task.  
The nlogattribute can be used to specify the number of log plexes to add. By  
default, one log plex is added. The loglenattribute specifies the size of the log,  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
282 Administering volumes  
Adding traditional DRL logging to a mirrored volume  
where each bit represents one region in the volume. For example, the size of the  
log would need to be 20K for a 10GB volume with a region size of 64 kilobytes.  
For example, to add a single log plex for the volume vol03, in the disk group,  
mydg, use the following command:  
# vxassist -g mydg addlog vol03 logtype=drl  
When the vxassistcommand is used to add a log subdisk to a volume, by  
default a log plex is also created to contain the log subdisk unless you include  
the keyword nologin the layout specification.  
For a volume that will be written to sequentially, such as a database log volume,  
include the logtype=drlseqattribute to specify that sequential DRL is to be  
used:  
# vxassist -g mydg addlog volume logtype=drlseq [nlog=n]  
Once created, the plex containing a log subdisk can be treated as a regular plex.  
Data subdisks can be added to the log plex. The log plex and log subdisk can be  
removed using the procedure described in “Removing a traditional DRL log” on  
page 282.  
Removing a traditional DRL log  
Note: The procedure described in this section removes a DRL log that is  
configured within a dedicated DRL plex. The version 20 DCO volume layout  
includes space for a DRL log.  
To remove a DRL log, use the vxassistcommand as follows:  
# vxassist [-g diskgroup] remove log volume logtype=drl [nlog=n]  
By default, the vxassistcommand removes one log. Use the optional attribute  
nlog=nto specify the number of logs that are to remain after the operation  
completes.  
You can use storage attributes to specify the storage from which a log is to be  
removed. For example, to remove a log on disk mydg10from volume vol01,  
enter:  
# vxassist -g mydg remove log vol01 !mydg10 logtype=drl  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Administering volumes 283  
Adding a RAID-5 log  
Adding a RAID-5 log  
Note: You need a full license to use this feature.  
Only one RAID-5 plex can exist per RAID-5 volume. Any additional plexes  
become RAID-5 log plexes, which are used to log information about data and  
parity being written to the volume. When a RAID-5 volume is created using the  
vxassistcommand, a log plex is created for that volume by default.  
To add a RAID-5 log to an existing volume, use the following command:  
# vxassist [-b] [-g diskgroup] addlog volume [loglen=length]  
Note: If specified, the -boption makes adding the new log a background task.  
You can specify the log length used when adding the first log to a volume. Any  
logs that you add subsequently are configured with the same length as the  
existing log.  
For example, to create a log for the RAID-5 volume volraid, in the disk group,  
mydg, use the following command:  
# vxassist -g mydg addlog volraid  
Adding a RAID-5 log using vxplex  
As an alternative to using vxassist, you can add a RAID-5 log using the vxplex  
command. For example, to attach a RAID-5 log plex, r5log, to a RAID-5 volume,  
r5vol, in the disk group, mydg, use the following command:  
# vxplex -g mydg att r5vol r5log  
The attach operation can only proceed if the size of the new log is large enough  
to hold all of the data on the stripe. If the RAID-5 volume already contains logs,  
the new log length is the minimum of each individual log length. This is because  
the new log is a mirror of the old logs.  
If the RAID-5 volume is not enabled, the new log is marked as BADLOG and is  
enabled when the volume is started. However, the contents of the log are  
ignored.  
If the RAID-5 volume is enabled and has other enabled RAID-5 logs, the new  
log’s contents are synchronized with the other logs.  
If the RAID-5 volume currently has no enabled logs, the new log is zeroed before  
it is enabled.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
284 Administering volumes  
Resizing a volume  
Removing a RAID-5 log  
To identify the plex of the RAID-5 log, use the following command:  
# vxprint [-g diskgroup] -ht volume  
where volume is the name of the RAID-5 volume. For a RAID-5 log, the output  
lists a plex with a STATEfield entry of LOG.  
To dissociate and remove a RAID-5 log and any associated subdisks from an  
existing volume, use the following command:  
# vxplex [-g diskgroup] -o rm dis plex  
For example, to dissociate and remove the log plex volraid-02from volraid  
in the disk group, mydg, use the following command:  
# vxplex -g mydg -o rm dis volraid-02  
You can also remove a RAID-5 log with the vxassistcommand, as follows:  
# vxassist [-g diskgroup] remove log volume [nlog=n]  
By default, the vxassistcommand removes one log. Use the optional attribute  
nlog=nto specify the number of logs that are to remain after the operation  
completes.  
Note: When removing the log leaves the volume with less than two valid logs, a  
warning is printed and the operation is not allowed to continue. The operation  
may be forced by additionally specifying the -foption to vxplexor vxassist.  
Resizing a volume  
Resizing a volume changes the volume size. For example, you might need to  
increase the length of a volume if it is no longer large enough for the amount of  
data to be stored on it. To resize a volume, use one of the commands: vxresize  
(preferred), vxassist, or vxvol. Alternatively, you can use the graphical Veritas  
Enterprise Administrator (VEA) to resize volumes.  
If a volume is increased in size, the vxassistcommand automatically locates  
available disk space. The vxresizecommand requires that you specify the  
names of the disks to be used to increase the size of a volume. The vxvol  
command requires that you have previously ensured that there is sufficient  
space available in the plexes of the volume to increase its size. The vxassistand  
vxresizecommands automatically free unused space for use by the disk group.  
For the vxvolcommand, you must do this yourself. To find out by how much you  
can grow a volume, use the following command:  
# vxassist [-g diskgroup] maxgrow volume  
When you resize a volume, you can specify the length of a new volume in  
sectors, kilobytes, megabytes, or gigabytes. The unit of measure is added as a  
suffix to the length (s, m, k, or g). If no unit is specified, sectors are assumed. The  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                   
Administering volumes 285  
Resizing a volume  
vxassistcommand also allows you to specify an increment by which to change  
the volume’s size.  
Caution: If you use vxassistor vxvolto resize a volume, do not shrink it below  
the size of the file system which is located on it. If you do not shrink the file  
system first, you risk unrecoverable data loss. If you have a VxFSfile system,  
shrink the file system first, and then shrink the volume. Other file systems may  
require you to back up your data so that you can later recreate the file system  
and restore its data.  
Resizing volumes using vxresize  
Use the vxresizecommand to resize a volume containing a file system.  
Although other commands can be used to resize volumes containing file  
systems, the vxresizecommand offers the advantage of automatically resizing  
certain types of file system as well as the volume.  
See the following table for details of what operations are permitted and whether  
the file system must first be unmounted to resize the file system:  
Table 8-1  
Permitted resizing operations on file systems  
Online JFS (Full- Base JFS (Lite-  
HFS  
VxFS)  
VxFS)  
Grow and shrink  
Grow only  
Not allowed  
Grow only  
Not allowed  
Grow only  
Mounted File System  
Unmounted File System  
For example, the following command resizes the 1-gigabyte volume, homevol,  
in the disk group, mydg, that contains a VxFS file system to 10 gigabytes using  
the spare disks mydg10and mydg11:  
# vxresize -g mydg -b -F vxfs -t homevolresize homevol 10g \  
mydg10 mydg11  
The -boption specifies that this operation runs in the background. Its progress  
can be monitored by specifying the task tag homevolresizeto the vxtask  
command.  
Note the following restrictions for using vxresize:  
■
■
vxresize works with VxFS, JFS (derived from VxFS) and HFS file systems  
only.  
In some situations, when resizing large volumes, vxresizemay take a long  
time to complete.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
286 Administering volumes  
Resizing a volume  
■
■
Resizing a volume with a usage type other than FSGEN or RAID5 can result  
in loss of data. If such an operation is required, use the -foption to forcibly  
resize such a volume.  
You cannot resize a volume that contains plexes with different layout types.  
Attempting to do so results in the following error message:  
VxVM vxresize ERROR V-5-1-2536 Volume volume has different  
organization in each mirror  
To resize such a volume successfully, you must first reconfigure it so that  
each data plex has the same layout.  
For more information about the vxresizecommand, see the vxresize(1M)  
manual page.  
Resizing volumes using vxassist  
The following modifiers are used with the vxassistcommand to resize a  
volume:  
growto  
Increase volume to a specified length.  
Increase volume by a specified amount.  
Reduce volume to a specified length.  
Reduce volume by a specified amount.  
growby  
shrinkto  
shrinkby  
Extending to a given length  
To extend a volume to a specific length, use the following command:  
# vxassist [-b] [-g diskgroup] growto volumelength  
Note: If specified, the -boption makes growing the volume a background task.  
For example, to extend volcatto 2000 sectors, use the following command:  
# vxassist -g mydg growto volcat 2000  
Note: If you previously performed a relayout on the volume, additionally specify  
the attribute layout=nodiskalignto the growtocommand if you want the  
subdisks to be grown using contiguous disk space.  
Extending by a given length  
To extend a volume by a specific length, use the following command:  
# vxassist [-b] [-g diskgroup] growby volumelength  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Administering volumes 287  
Resizing a volume  
Note: If specified, the -boption makes growing the volume a background task.  
For example, to extend volcatby 100 sectors, use the following command:  
# vxassist -g mydg growby volcat 100  
Note: If you previously performed a relayout on the volume, additionally specify  
the attribute layout=nodiskalignto the growbycommand if you want the  
subdisks to be grown using contiguous disk space.  
Shrinking to a given length  
To shrink a volume to a specific length, use the following command:  
# vxassist [-g diskgroup] shrinkto volumelength  
For example, to shrink volcatto 1300 sectors, use the following command:  
# vxassist -g mydg shrinkto volcat 1300  
Caution: Do not shrink the volume below the current size of the file system or  
database using the volume. The vxassistshrinktocommand can be safely used  
on empty volumes.  
Shrinking by a given length  
To shrink a volume by a specific length, use the following command:  
# vxassist [-g diskgroup] shrinkby volumelength  
For example, to shrink volcatby 300 sectors, use the following command:  
# vxassist -g mydg shrinkby volcat 300  
Caution: Do not shrink the volume below the current size of the file system or  
database using the volume. The vxassistshrinkbycommand can be safely used  
on empty volumes.  
Resizing volumes using vxvol  
To change the length of a volume using the vxvolsetcommand, use the  
following command:  
# vxvol [-g diskgroup] set len=length volume  
For example, to change the length of the volume, vol01, in the disk group,  
mydg, to 100000 sectors, use the following command:  
# vxvol -g mydg set len=100000 vol01  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
288 Administering volumes  
Setting tags on volumes  
Note: The vxvolsetlencommand cannot increase the size of a volume unless  
the needed space is available in the plexes of the volume. When the size of a  
volume is reduced using the vxvolsetlencommand, the freed space is not  
released into the disk group’s free space pool.  
If a volume is active and its length is being reduced, the operation must be  
forced using the -oforce option to vxvol. This prevents accidental removal of  
space from applications using the volume.  
The length of logs can also be changed using the following command:  
# vxvol [-g diskgroup] set loglen=length log_volume  
Note: Sparse log plexes are not valid. They must map the entire length of the log.  
If increasing the log length would make any of the logs invalid, the operation is  
not allowed. Also, if the volume is not active and is dirty (for example, if it has  
not been shut down cleanly), the log length cannot be changed. This avoids the  
loss of any of the log contents (if the log length is decreased), or the introduction  
of random data into the logs (if the log length is being increased).  
Setting tags on volumes  
Volume tags are used to implement the Dynamic Storage Tiering feature of the  
Storage Foundation software. For more information about this feature, see the  
Veritas File System Administrator’s Guide.  
You can use the following forms of the vxassistcommand to set a named tag  
and optional tag value on a volume, to replace a tag, and to remove a tag from a  
volume:  
# vxassist [-g diskgroup] settag volume tagname[=tagvalue]  
# vxassist [-g diskgroup] replacetag volume oldtag newtag  
# vxassist [-g diskgroup] removetag volume tagname  
To list the tags that are associated with a volume, use this command:  
# vxassist [-g diskgroup] listtag volume  
To list the volumes that have a specified tag name, use this command:  
# vxassist [-g diskgroup] list tag=tagname volume  
Tag names and tag values are case-sensitive character strings of up to 256  
characters. Tag names can consist of letters (A through Z and a through z),  
numbers (0 through 9), dashes (-), underscores (_) or periods (.) from the ASCII  
character set. A tag name must start with either a letter or an underscore. Tag  
values can consist of any character from the ASCII character set with a decimal  
value from 32 through 127. If a tag value includes any spaces, quote the  
specification to protect it from the shell, as shown here:  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                 
Administering volumes 289  
Changing the read policy for mirrored volumes  
# vxassist -g mydg settag myvol "dbvol=table space 1"  
Dotted tag hierarchies are understood by the listoperation. For example, the  
listing for tag=a.bincludes all volumes that have tag names that start with a.b.  
The tag names site, udidand vdidare reserved and should not be used. To  
avoid possible clashes with future product features, it is recommended that tag  
names do not start with any of the following strings: asl, be, isp, nbu, sf,  
symcor vx.  
Changing the read policy for mirrored volumes  
VxVM offers the choice of the following read policies on the data plexes in a  
mirrored volume:  
round  
Reads each plex in turn in “round-robin” fashion for each  
nonsequential I/O detected. Sequential access causes only one  
plex to be accessed. This takes advantage of the drive or  
controller read-ahead caching policies.  
prefer  
select  
Reads first from a plex that has been named as the preferred  
plex.  
Chooses a default policy based on plex associations to the  
volume. If the volume has an enabled striped plex, the select  
option defaults to preferring that plex; otherwise, it defaults  
to round-robin.  
siteread  
split  
Reads preferentially from plexes at the locally defined site.  
This is the default policy for volumes in disk groups where site  
consistency has been enabled.  
See “Administering sites and remote mirrors” on page 431.  
Divides read requests and distributes them across all the  
available plexes.  
Note: You cannot set the read policy on a RAID-5 volume. RAID-5 plexes have  
their own read policy (RAID).  
To set the read policy to round, use the following command:  
# vxvol [-g diskgroup] rdpol round volume  
For example, to set the read policy for the volume, vol01, in disk group, mydg,  
to round-robin, use the following command:  
# vxvol -g mydg rdpol round vol01  
To set the read policy to prefer, use the following command:  
# vxvol [-g diskgroup] rdpol prefer volume preferred_plex  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                   
290 Administering volumes  
Removing a volume  
For example, to set the policy for vol01to read preferentially from the plex  
vol01-02, use the following command:  
# vxvol -g mydg rdpol prefer vol01 vol01-02  
To set the read policy to select, use the following command:  
# vxvol [-g diskgroup] rdpol select volume  
For more information about how read policies affect performance, see “Volume  
read policies” on page 466.  
Removing a volume  
Once a volume is no longer necessary (it is inactive and its contents have been  
archived, for example), it is possible to remove the volume and free up the disk  
space for other uses.  
To stop all activity on a volume before removing it  
1
2
3
Remove all references to the volume by application programs, including  
shells, that are running on the system.  
If the volume is mounted as a file system, unmount it with this command:  
# umount /dev/vx/dsk/diskgroup/volume  
If the volume is listed in the /etc/fstabfile, remove its entry by editing  
this file. Refer to your operating system documentation for more  
information about the format of this file and how you can modify it.  
4
Stop all activity by VxVM on the volume with the command:  
# vxvol [-g diskgroup] stop volume  
After following these steps, remove the volume with the vxassistcommand:  
# vxassist [-g diskgroup] remove volume volume  
Alternatively, you can use the vxeditcommand to remove a volume:  
# vxedit [-g diskgroup] [-r] [-f] rm volume  
The -roption to vxeditindicates recursive removal. This removes all plexes  
associated with the volume and all subdisks associated with those plexes. The -f  
option to vxeditforces removal. This is necessary if the volume is still enabled.  
Moving volumes from a VM disk  
Before you disable or remove a disk, you can move the data from that disk to  
other disks on the system that have sufficient space.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                     
Administering volumes 291  
Moving volumes from a VM disk  
To move volumes from a disk  
1
Select menu item 6 (Move volumes from a disk)from the  
vxdiskadmmain menu.  
2
At the following prompt, enter the disk name of the disk whose volumes you  
wish to move, as follows:  
Move volumes from a disk  
Menu: VolumeManager/Disk/Evacuate  
Use this menu operation to move any volumes that are using a  
disk onto other disks. Use this menu immediately prior to  
removing a disk, either permanently or for replacement. You  
can specify a list of disks to move volumes onto, or you can  
move the volumes to any available disk space in the same disk  
group.  
NOTE: Simply moving volumes off of a disk, without also  
removing the disk, does not prevent volumes from being moved  
onto the disk by future operations. For example, using two  
consecutive move operations may move volumes from the  
second disk to the first.  
Enter disk name [<disk>,list,q,?] mydg01  
You can now optionally specify a list of disks to which the volume(s) should  
be moved:  
VxVM INFO V-5-2-516 You can now specify a list of disks to move  
onto. Specify a list of disk media names (e.g., mydg01) all on  
one line separated by blanks. If you do not enter any disk  
media names, then the volumes will be moved to any available  
space in the disk group.  
At the prompt, press Return to move the volumes onto available space in  
the disk group, or specify the disks in the disk group that should be used:  
Enter disks [<disk ...>,list]  
VxVM NOTICE V-5-2-283 Requested operation is to move all  
volumes from disk mydg01 in group mydg.  
NOTE: This operation can take a long time to complete.  
Continue with operation? [y,n,q,?] (default: y)  
As the volumes are moved from the disk, the vxdiskadmprogram displays  
the status of the operation:  
VxVM vxevac INFO V-5-2-24  
Move volume voltest ...  
When the volumes have all been moved, the vxdiskadmprogram displays  
the following success message:  
VxVM INFO V-5-2-188 Evacuation of disk mydg02 is complete.  
3
At the following prompt, indicate whether you want to move volumes from  
another disk (y) or return to the vxdiskadmmain menu (n):  
Move volumes from another disk? [y,n,q,?] (default: n)  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
292 Administering volumes  
Enabling FastResync on a volume  
Enabling FastResync on a volume  
Note: The recommended method for enabling FastResync on a volume with a  
version 20 DCO is to use the vxsnappreparecommand as described in  
“Preparing a volume for DRL and instant snapshots” on page 275.  
TM  
You need a Veritas FlashSnap or FastResync license to use this feature.  
FastResync performs quick and efficient resynchronization of stale mirrors. It  
also increases the efficiency of the VxVM snapshot mechanism when used with  
operations such as backup and decision support. See “Administering volume  
snapshots” on page 303 and “FastResync” on page 66 for more information.  
There are two possible versions of FastResync that can be enabled on a volume:  
■
Persistent FastResync holds copies of the FastResync maps on disk. These  
can be used for the speedy recovery of mirrored volumes if a system is  
rebooted. This form of FastResync requires that either a version 0 or a  
version 20 data change object (DCO) and DCO volume first be associated  
with the volume.  
See “Adding a version 0 DCO and DCO volume” on page 356, and “Preparing  
a volume for DRL and instant snapshots” on page 275 for more information  
on version 0 and version 20 DCO volumes respectively.  
If the existing volume was created in a previous release of VxVM, and it has  
any attached snapshot plexes or it is associated with any snapshot volumes,  
follow the procedure given in “Upgrading existing volumes to use version  
20 DCOs” on page 279.  
■
Non-Persistent FastResync holds the FastResync maps in memory. These do  
not survive on a system that is rebooted.  
By default, FastResync is not enabled on newly created volumes. Specify the  
fastresync=onattribute to the vxassistmakecommand if you want to enable  
FastResync on a volume that you are creating.  
Note: It is not possible to configure both Persistent and Non-Persistent  
FastResync on a volume. Persistent FastResync is used if a DCO is associated  
with the volume. Otherwise, Non-Persistent FastResync is used.  
To turn FastResync on for an existing volume, specify fastresync=onto the  
vxvolcommand as shown here:  
# vxvol [-g diskgroup] set fastresync=on volume  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Administering volumes 293  
Enabling FastResync on a volume  
Note: To use FastResync with a snapshot, FastResync must be enabled before  
the snapshot is taken, and must remain enabled until after the snapback is  
completed.  
Checking whether FastResync is enabled on a volume  
To check whether FastResync is enabled on a volume, use the following  
command:  
# vxprint [-g diskgroup] -F%fastresync volume  
This command returns onif FastResync is enabled; otherwise, it returns off.  
If FastResync is enabled, to check whether it is Non-Persistent or Persistent  
FastResync, use the following command:  
# vxprint [-g diskgroup] -F%hasdcolog volume  
This command returns onif Persistent FastResync is enabled; otherwise, it  
returns off.  
To list all volumes on which Non-Persistent FastResync is enabled, use the  
following command:  
# vxprint [-g diskgroup] -F “%name” \  
-e “v_fastresync=on && !v_hasdcolog”  
To list all volumes on which Persistent FastResync is enabled, use the following  
command:  
# vxprint [-g diskgroup] -F “%name” -e “v_fastresync=on \  
&& v_hasdcolog”  
Disabling FastResync  
Use the vxvolcommand to turn off Persistent or Non-Persistent FastResync for  
an existing volume, as shown here:  
# vxvol [-g diskgroup] set fastresync=off volume  
Turning FastResync off releases all tracking maps for the specified volume. All  
subsequent reattaches will not use the FastResync facility, but perform a full  
resynchronization of the volume. This occurs even if FastResync is later turned  
on.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
               
294 Administering volumes  
Performing online relayout  
Performing online relayout  
Note: You need a full license to use this feature.  
You can use the vxassistrelayoutcommand to reconfigure the layout of a  
volume without taking it offline. The general form of this command is:  
# vxassist [-b] [-g diskgroup] relayout volume [layout=layout] \  
[relayout_options]  
Note: If specified, the -boption makes relayout of the volume a background task.  
The following are valid destination layout configurations as determined by the  
tables in “Permitted relayout transformations” on page 295:  
concat-mirror—concatenated-mirror  
concator span, nostripe, nomirror—concatenated  
raid5—RAID-5 (not supported for shared disk groups)  
stripe—striped  
stripe-mirror—striped-mirror  
For example, the following command changes a concatenated volume, vol02, in  
disk group, mydg, to a striped volume with the default number of columns, 2,  
and default stripe unit size, 64 kilobytes:  
# vxassist -g mydg relayout vol02 layout=stripe  
On occasions, it may be necessary to perform a relayout on a plex rather than on  
a volume. See “Specifying a plex for relayout” on page 298 for more information.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Administering volumes 295  
Performing online relayout  
Permitted relayout transformations  
The tables below give details of the relayout operations that are possible for  
each type of source storage layout.  
Table 8-2  
Supported relayout transformations for concatenated volumes  
Relayout to  
concat  
From concat  
No.  
No. Add a mirror, and then use vxassistconvertinstead.  
concat-mirror  
mirror-concat  
mirror-stripe  
No. Add a mirror instead.  
No. Use vxassistconvertafter relayout to striped-mirror volume  
instead.  
Yes. The stripe width and number of columns may be defined.  
Yes. The stripe width and number of columns may be defined.  
Yes. The stripe width and number of columns may be defined.  
raid5  
stripe  
stripe-mirror  
Table 8-3  
Supported relayout transformations for concatenated-mirror  
volumes  
Relayout to  
concat  
From concat-mirror  
No. Use vxassist convert, and then remove unwanted mirrors  
from the resulting mirrored-concatenated volume instead.  
No.  
concat-mirror  
mirror-concat  
mirror-stripe  
No. Use vxassistconvertinstead.  
No. Use vxassistconvertafter relayout to striped-mirror volume  
instead.  
Yes.  
raid5  
Yes. This removes a mirror and adds striping. The stripe width and  
number of columns may be defined.  
stripe  
Yes. The stripe width and number of columns may be defined.  
stripe-mirror  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
296 Administering volumes  
Performing online relayout  
Table 8-4  
Supported relayout transformations for RAID-5 volumes  
Relayout to  
concat  
From raid5  
Yes.  
Yes.  
concat-mirror  
mirror-concat  
No. Use vxassistconvertafter relayout to concatenated-mirror  
volume instead.  
No. Use vxassistconvertafter relayout to striped-mirror volume  
mirror-stripe  
instead.  
Yes. The stripe width and number of columns may be changed.  
Yes. The stripe width and number of columns may be changed.  
Yes. The stripe width and number of columns may be changed.  
raid5  
stripe  
stripe-mirror  
Table 8-5  
Supported relayout transformations for mirrored-concatenated  
volumes  
Relayout to  
concat  
From mirror-concat  
No. Remove unwanted mirrors instead.  
No. Use vxassistconvertinstead.  
No.  
concat-mirror  
mirror-concat  
mirror-stripe  
No. Use vxassistconvertafter relayout to striped-mirror volume  
instead.  
Yes. The stripe width and number of columns may be defined.  
Choose a plex in the existing mirrored volume on which to perform  
the relayout. The other plexes are removed at the end of the relayout  
operation.  
raid5  
Yes.  
Yes.  
stripe  
stripe-mirror  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Administering volumes 297  
Performing online relayout  
Table 8-6  
Supported relayout transformations for mirrored-stripe volumes  
Relayout to  
concat  
From mirror-stripe  
Yes.  
Yes.  
concat-mirror  
mirror-concat  
No. Use vxassistconvertafter relayout to concatenated-mirror  
volume instead.  
No. Use vxassistconvertafter relayout to striped-mirror volume  
mirror-stripe  
instead.  
Yes. The stripe width and number of columns may be changed.  
Yes. The stripe width or number of columns must be changed.  
raid5  
stripe  
Yes. The stripe width or number of columns must be changed.  
stripe-mirror  
Otherwise, use vxassistconvert.  
Table 8-7  
Supported relayout transformations for unmirrored stripe and  
layered striped-mirror volumes  
Relayout to  
concat  
From stripe or stripe-mirror  
Yes.  
Yes.  
concat-mirror  
mirror-concat  
No. Use vxassistconvertafter relayout to concatenated-mirror  
volume instead.  
No. Use vxassistconvertafter relayout to striped-mirror volume  
mirror-stripe  
instead.  
Yes. The stripe width and number of columns may be changed.  
Yes. The stripe width or number of columns must be changed.  
Yes. The stripe width or number of columns must be changed.  
raid5  
stripe  
stripe-mirror  
Download from Www.Somanuals.com. All Manuals Search And Download.  
298 Administering volumes  
Performing online relayout  
Specifying a non-default layout  
You can specify one or more relayout options to change the default layout  
configuration. Examples of these options are:  
ncol=number  
Specifies the number of columns.  
ncol=+number  
ncol=-number  
stripeunit=size  
Specifies the number of columns to add.  
Specifies the number of colums to remove.  
Specifies the stripe width.  
See the vxassist(1M) manual page for more information about relayout  
options.  
The following are some examples of using vxassist to change the stripe width  
and number of columns for a striped volume in the disk group dbaseg:  
# vxassist -g dbaseg relayout vol03 stripeunit=64k ncol=6  
# vxassist -g dbaseg relayout vol03 ncol=+2  
# vxassist -g dbaseg relayout vol03 stripeunit=128k  
The next example changes a concatenated volume to a RAID-5 volume with four  
columns:  
# vxassist -g fsgrp relayout vol04 layout=raid5 ncol=4  
Specifying a plex for relayout  
Any layout can be changed to RAID-5 if there are sufficient disks and space in  
the disk group. If you convert a mirrored volume to RAID-5, you must specify  
which plex is to be converted. All other plexes are removed when the conversion  
has finished, releasing their space for other purposes. If you convert a mirrored  
volume to a layout other than RAID-5, the unconverted plexes are not removed.  
You can specify the plex to be converted by naming it in place of a volume:  
# vxassist [-g diskgroup] relayout plex [layout=layout] \  
[relayout_options]  
Tagging a relayout operation  
If you want to control the progress of a relayout operation, for example to pause  
or reverse it, use the -toption to vxassistto specify a task tag for the  
operation. For example, this relayout is performed as a background task and has  
the tag myconv:  
# vxassist -b -g fsgrp -t myconv relayout vol04 layout=raid5 \  
ncol=4  
See the following sections, “Viewing the status of a relayout” on page 299 and  
“Controlling the progress of a relayout” on page 299, for more information  
about tracking and controlling the progress of relayout.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                     
Administering volumes 299  
Performing online relayout  
Viewing the status of a relayout  
Online relayout operations take some time to perform. You can use the  
vxrelayoutcommand to obtain information about the status of a relayout  
operation. For example, the command:  
# vxrelayout -g mydg status vol04  
might display output similar to this:  
STRIPED, columns=5, stwidth=128--> STRIPED, columns=6,  
stwidth=128  
Relayout running, 68.58% completed.  
In this example, the reconfiguration of a striped volume from 5 to 6 columns is  
in progress, and is just over two-thirds complete.  
See the vxrelayout(1M) manual page for more information about this  
command.  
If you specified a task tag to vxassistwhen you started the relayout, you can  
use this tag with the vxtaskcommand to monitor the progress of the relayout.  
For example, to monitor the task tagged as myconv, enter:  
# vxtask monitor myconv  
Controlling the progress of a relayout  
You can use the vxtaskcommand to stop (pause) the relayout temporarily, or to  
cancel it altogether (abort). If you specified a task tag to vxassistwhen you  
started the relayout, you can use this tag to specify the task to vxtask. For  
example, to pause the relayout operation tagged as myconv, enter:  
# vxtask pause myconv  
To resume the operation, use the vxtaskcommand:  
# vxtask resume myconv  
For relayout operations that have not been stopped using the vxtaskpause  
command (for example, the vxtaskabortcommand was used to stop the task,  
the transformation process died, or there was an I/O failure), resume the  
relayout by specifying the startkeyword to vxrelayout, as shown here:  
# vxrelayout -g mydg -o bg start vol04  
Note: If you use the vxrelayoutstartcommand to restart a relayout that you  
previously suspended using the vxtaskpausecommand, a new untagged task is  
created to complete the operation. You cannot then use the original task tag to  
control the relayout.  
The -o bgoption restarts the relayout in the background. You can also specify  
the slowand iosizeoption modifiers to control the speed of the relayout and  
the size of each region that is copied. For example, the following command  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                           
300 Administering volumes  
Converting between layered and non-layered volumes  
inserts a delay of 1000 milliseconds (1 second) between copying each 10-  
megabyte region:  
# vxrelayout -g mydg -o bg,slow=1000,iosize=10m start vol04  
The default delay and region size values are 250 milliseconds and 1 megabyte  
respectively.  
To reverse the direction of relayout operation that is currently stopped, specify  
the reversekeyword to vxrelayoutas shown in this example:  
# vxrelayout -g mydg -o bg reverse vol04  
This undoes changes made to the volume so far, and returns it to its original  
layout.  
If you cancel a relayout using vxtaskabort, the direction of the conversion is  
also reversed, and the volume is returned to its original configuration.  
See the vxrelayout(1M) and vxtask(1M) manual pages for more information  
about these commands. See “Managing tasks with vxtask” on page 268 for more  
information about controlling tasks in VxVM.  
Converting between layered and non-layered  
volumes  
The vxassistconvertcommand transforms volume layouts between layered  
and non-layered forms:  
# vxassist [-b] [-g diskgroup] convert volume [layout=layout] \  
[convert_options]  
Note: If specified, the -boption makes conversion of the volume a background  
task.  
The following conversion layouts are supported:  
stripe-mirror  
mirror-stripe  
concat-mirror  
mirror-concat  
mirrored-stripe to striped-mirror  
striped-mirror to mirrored-stripe  
mirrored-concatenated to concatenated-mirror  
concatenated-mirror to mirrored-concatenated  
Volume conversion can be used before or after performing online relayout to  
achieve a larger number of transformations than would otherwise be possible.  
During relayout process, a volume may also be converted into a layout that is  
intermediate to the one that is desired. For example, to convert a volume from a  
4-column mirrored-stripe to a 5-column mirrored-stripe, first use vxassist  
relayoutto convert the volume to a 5-column striped-mirror as shown here:  
# vxassist -g mydg relayout vol1 ncol=5  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                     
Administering volumes 301  
Converting between layered and non-layered volumes  
When the relayout has completed, use the vxassist convertcommand to  
change the resulting layered striped-mirror volume to a non-layered mirrored-  
stripe:  
# vxassist -g mydg convert vol1 layout=mirror-stripe  
Note: If the system crashes during relayout or conversion, the process continues  
when the system is rebooted. However, if the crash occurred during the first  
stage of a two-stage relayout and convert operation, only the first stage will be  
completed. You must run vxassistconvertmanually to complete the  
operation.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
302 Administering volumes  
Converting between layered and non-layered volumes  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter  
9
Administering volume  
snapshots  
Veritas Volume Manager (VxVM) provides the capability for taking an image of a  
volume at a given point in time. Such an image is referred to as a volume  
snapshot. You can also take a snapshot of a volume set as described in “Creating  
instant snapshots of volume sets” on page 334.  
Volume snapshots allow you to make backup copies of your volumes online with  
minimal interruption to users. You can then use the backup copies to restore  
data that has been lost due to disk failure, software errors or human mistakes, or  
to create replica volumes for the purposes of report generation, application  
development, or testing.  
For an introduction to the volume snapshot feature, see “Volume snapshots” on  
page 63. More detailed descriptions of each type of volume snapshot and the  
operations that you can perform on them may be found in the following  
sections:  
■
■
■
■
■
■
■
■
“Traditional third-mirror break-off snapshots” on page 305  
“Full-sized instant snapshots” on page 307  
“Space-optimized instant snapshots” on page 309  
“Emulation of third-mirror break-off snapshots” on page 310  
“Linked break-off snapshot volumes” on page 311  
“Cascaded snapshots” on page 312  
“Creating multiple snapshots” on page 317  
“Restoring the original volume from a snapshot” on page 317  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
304 Administering volume snapshots  
Note: A volume snapshot represents the data that exists in a volume at a given  
point in time. As such, VxVM does not have any knowledge of data that is cached  
by the overlying file system, or by applications such as databases that have files  
open in the file system. If the fsgenvolume usage type is set on a volume that  
contains a Veritas File System (VxFS), intent logging of the file system metadata  
ensures the internal consistency of the file system that is backed up. For other  
file system types, depending on the intent logging capabilities of the file system,  
there may potentially be inconsistencies between data in memory and in the  
snapshot image.  
For databases, a suitable mechanism must additionally be used to ensure the  
integrity of tablespace data when the volume snapshot is taken. The facility to  
temporarily suspend file system I/O is provided by most modern database  
software. For ordinary files in a file system, which may be open to a wide variety  
of different applications, there may be no way to ensure the complete integrity  
of the file data other than by shutting down the applications and temporarily  
unmounting the file system. In many cases, it may only be important to ensure  
the integrity of file data that is not in active use at the time that you take the  
snapshot.  
Methods of creating volume snapshots are described in the following sections:  
■
“Creating instant snapshots” on page 319 describes how to use the vxsnap  
command to create and administer full-sized and space-optimized instant  
snapshots.  
■
“Creating traditional third-mirror break-off snapshots” on page 348  
describes how to use the vxassistcommand to create and administer  
traditional third-mirror snapshots.  
For details of how to use volume snapshots to implement off-host online backup,  
see “Configuring off-host processing” on page 369.  
Note: Snapshot creation using the vxsnapcommand is the preferred mechanism  
for implementing online and off-host point-in-time copy solutions in VxVM.  
Support for traditional third-mirror snapshots that are created using the  
vxassistcommand may be removed in a future release.  
Most VxVM commands require superuser or equivalent privileges.  
Full details of how to recover from failures of instant snapshot commands may  
be found in the “Recovery from failure of instant snapshot operations’’ chapter  
of the Veritas Volume Manager Troubleshooting Guide.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Administering volume snapshots 305  
Traditional third-mirror break-off snapshots  
Traditional third-mirror break-off snapshots  
The traditional third-mirror break-off volume snapshot model that is supported  
by the vxassistcommand is shown in Figure 9-1. This also shows the  
transitions that are supported by the snapbackand snapclearcommands to  
vxassist.  
Figure 9-1  
Third-mirror snapshot creation and usage  
vxassist snapstart  
Start  
vxassist snapshot  
Original  
volume  
Original  
volume  
Backup  
Cycle  
Snapshot  
mirror  
Snapshot  
volume  
Refresh on  
snapback  
vxassist snapback  
Back up to disk,  
tape or other  
media, or use to  
create replica  
database or file  
system.  
vxassist snapclear  
Independent  
volume  
The vxassistsnapstartcommand creates a mirror to be used for the snapshot,  
and attaches it to the volume as a snapshot mirror. (The vxassistsnapabort  
command can be used to cancel this operation and remove the snapshot mirror.)  
Note: As is usual when creating a mirror, the process of copying the volume’s  
contents to the new snapshot plexes can take some time to complete. For  
methods of making snapshot plexes immediately available, see “Full-sized  
instant snapshots” on page 307 and “Space-optimized instant snapshots” on  
page 309.  
When the attachment is complete, the vxassistsnapshotcommand is used to  
create a new snapshot volume by taking one or more snapshot mirrors to use as  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
306 Administering volume snapshots  
Traditional third-mirror break-off snapshots  
its data plexes. The snapshot volume contains a copy of the original volume’s  
data at the time that you took the snapshot. If more than one snapshot mirror is  
used, the snapshot volume is itself mirrored.  
The command, vxassistsnapback,can be used to return snapshot plexes to the  
original volume from which they were snapped, and to resynchronize the data in  
the snapshot mirrors from the data in the original volume. This enables you to  
refresh the data in a snapshot after each time that you use it to make a backup.  
You can also use a variation of the same command to restore the contents of the  
original volume from a snapshot that you took at an earlier point in time. See  
“Restoring the original volume from a snapshot” on page 317 for more  
information.  
As described in “FastResync” on page 66, you can use the FastResync feature of  
VxVM to minimize the time needed to resynchronize the data in the snapshot  
mirror. If FastResync is not enabled, a full resynchronization of the data is  
required.  
Finally, you can use the vxassistsnapclearcommand to break the association  
between the original volume and the snapshot volume. The snapshot volume  
then has an existence that is independent of the original volume. This is useful  
for applications that do not require the snapshot to be resynchronized with the  
original volume.  
Note: The use of the vxassistcommand to administer traditional (third-mirror  
break-off) snapshots is not supported for volumes that are prepared for instant  
snapshot creation. Instead, the vxsnapcommand may be used as described in  
the following section.  
See “Creating traditional third-mirror break-off snapshots” on page 348 for a  
description of the procedures for creating and using this type of snapshot.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Administering volume snapshots 307  
Full-sized instant snapshots  
Full-sized instant snapshots  
Full-sized instant snapshots are a variation on the third-mirror volume  
snapshot model that make a snapshot volume available for access as soon as the  
snapshot plexes have been created. The full-sized instant volume snapshot  
model is illustrated in Figure 9-2.  
Figure 9-2  
Full-sized instant snapshot creation and usage in a backup cycle  
Start  
vxsnap prepare  
vxsnap make  
vxsnap refresh  
Backup  
Cycle  
Original  
volume  
Snapshot  
volume  
vxsnap reattach  
vxsnap dis  
or  
Back up to disk, tape or  
other media. The snapshot  
volume can also be used to  
create a replica database or  
file system when  
vxsnap split  
Independent  
volume  
synchronization is complete.  
To create an instant snapshot, you use the vxsnapmakecommand. This  
command can either be applied to a suitably prepared empty volume that is to be  
used as the snapshot volume, or it can be used to break off one or more  
synchronized plexes from the original volume (which is similar to the way that  
the vxassistcommand creates its snapshots).  
Unlike a third-mirror break-off snapshot created using the vxassistcommand,  
you can make a backup of a full-sized instant snapshot, instantly refresh its  
contents from the original volume, or attach its plexes to the original volume,  
without needing to completely synchronize the snapshot plexes from the  
original volume.  
VxVM uses a copy-on-write mechanism to ensure that the snapshot volume  
preserves the contents of the original volume at the time that the snapshot is  
taken. Any time that the original contents of the volume are about to be  
overwritten, the original data in the volume is preserved on the snapshot  
volume before the write proceeds. As time goes by, and the contents of the  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
308 Administering volume snapshots  
Full-sized instant snapshots  
volume are updated, its original contents are gradually relocated to the  
snapshot volume.  
If desired, you can additionally select to perform either a background (non-  
blocking) or foreground (blocking) synchronization of the snapshot volume.  
This is useful if you intend to move the snapshot volume into a separate disk  
group for off-host processing, or you want to use the vxsnapdisor vxsnap  
splitcommands to turn the snapshot volume into an independent volume.  
The vxsnaprefreshcommand allows you to update the data in a snapshot each  
time that you make a backup.  
The command, vxsnapreattach,can be used to attach snapshot plexes to the  
original volume, and to resynchronize the data in these plexes from the original  
volume. Alternatively, you can use the vxsnaprestorecommand to restore the  
contents of the original volume from a snapshot that you took at an earlier point  
in time. You can also choose whether or not to keep the snapshot volume after  
restoration of the original volume is complete. See “Restoring the original  
volume from a snapshot” on page 317 for more information.  
By default, the FastResync feature of VxVM is used to minimize the time needed  
to resynchronize the data in the snapshot mirror. If FastResync is not enabled, a  
full resynchronization of the data is required. For details, see “FastResync” on  
page 66.  
See “Creating and managing full-sized instant snapshots” on page 327 for  
details of the procedures for creating and using this type of snapshot.  
For information about how to prepare an empty volume for use by full-sized  
instant snapshots, see “Creating a volume for use as a full-sized instant or  
linked break-off snapshot” on page 323.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Administering volume snapshots 309  
Space-optimized instant snapshots  
Space-optimized instant snapshots  
Volume snapshots, such as those described in “Traditional third-mirror break-  
off snapshots” on page 305 and “Full-sized instant snapshots” on page 307,  
require the creation of a complete copy of the original volume, and use as much  
storage space as the original volume.  
Instead of requiring a complete copy of the original volume’s storage space,  
space-optimized instant snapshots use a storage cache. The size of this cache  
may be configured when the snapshot is created.  
Note: A storage cache may be named and shared among several volumes in the  
same disk group. If so, the size of the cache that is declared must be the same for  
each volume’s space-optimized snapshot. You may find it convenient to  
configure a single storage cache in a disk group that can be shared by all the  
volumes in that disk group. See “Creating a shared cache object” on page 322 for  
details.  
When the original volume is written to, VxVM preserves the original data  
contents in the cache before the write is committed. As the storage cache can be  
configured to require much less storage than the original volume, it is referred  
to as being space-optimized. If the cache becomes too full, you can configure  
VxVM to grow the size of the cache automatically using any available free space  
in the disk group.  
The instant space-optimized snapshot model is illustrated in Figure 9-3.  
Figure 9-3  
Space-optimized instant snapshot creation and usage in a backup  
cycle  
Start  
vxsnap prepare  
vxsnap make  
vxsnap refresh  
Backup  
Cycle  
Original  
volume  
Snapshot  
volume  
Back up to disk, tape or  
other media.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
310 Administering volume snapshots  
Emulation of third-mirror break-off snapshots  
As for instant snapshots, space-optimized snapshots use a copy-on-write  
mechanism to make them immediately available for use when they are first  
created, or when their data is refreshed. Unlike instant snapshots, however, you  
cannot enable synchronization on space-optimized snapshots, reattach them to  
their original volume, or turn them into independent volumes.  
See “Creating and managing space-optimized instant snapshots” on page 324  
for details of the procedures for creating and using this type of snapshot.  
For information about how to set up a cache for use by space-optimized instant  
snapshots, see “Creating a shared cache object” on page 322.  
Emulation of third-mirror break-off snapshots  
Third-mirror break-off snapshots are suitable for write-intensive volumes (such  
as for database redo logs) where the copy-on-write mechanism of space-  
optimized or full-sized instant snapshots might degrade the performance of the  
volume.  
If you use the vxsnappreparecommand to enable a volume for use with instant  
and space-optimized snapshots, you cannot use the vxassistsnapshot  
commands to administer snapshots that you create for the volume. If you  
require snapshots that behave as third-mirror break-off snapshots (that is, they  
must be fully synchronized before they can be used), there are three ways to  
achieve this:  
■
Use the vxsnapaddmircommand to create and attach one or more snapshot  
mirrors to the volume. When the plexes have been synchronized and are in  
the SNAPDONEstate, the vxsnapmakecommand can then be used with the  
nmirrorattribute to create the snapshot volume. This technique is similar  
to using the vxassistsnapstartand vxassistsnapshotcommands that  
are described in “Traditional third-mirror break-off snapshots” on  
page 305.  
■
Use the vxsnapmakecommand with the plexattribute to use one or more  
existing plexes of a volume as snapshot plexes. The volume must have a  
sufficient number of available plexes that are in the ACTIVEstate.  
Note: The volume must be a non-layered volume with a mirroror mirror-  
stripelayout, or a RAID-5 volume that you have converted to a special  
layered volume (see “Using a DCO and DCO volume with a RAID-5 volume”  
on page 277) and then mirrored.  
The plexes in a volume with a stripe-mirrorlayout are mirrored at the  
sub-volume level, and cannot be broken off.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Administering volume snapshots 311  
Linked break-off snapshot volumes  
■
Use the vxsnapmakecommand with the sync=yesand type=fullattributes  
specified to create the snapshot volume, and then use the vxsnapsyncwait  
command to wait for synchronization of the snapshot volume to complete.  
See “Creating and managing third-mirror break-off snapshots” on page 329 for  
details of the procedures for creating and using this type of snapshot.  
For information about how to add snapshot mirrors to a volume, see “Adding  
snapshot mirrors to a volume” on page 336.  
Linked break-off snapshot volumes  
A variant of the third-mirror break-off snapshot type are linked break-off  
snapshot volumes, which use the vxsnapaddmircommand to link a specially  
prepared volume with the data volume. The volume that is used for the snapshot  
is prepared in the same way as for full-sized instant snapshots. However, unlike  
full-sized instant snapshots, this volume can be set up in a different disk group  
from the data volume. This makes linked break-off snapshots especially suitable  
for off-host processing applications where you may want to create the snapshot  
on storage with different characteristics from that used for the data volumes. As  
for third-mirror break-off snapshots, you must wait for the contents of the  
snapshot volume to be synchronized with the data volume before you can use  
the vxsnapmakecommand to take the snapshot.  
When a link is created between a volume and the mirror that will become the  
snapshot, separate link objects (similar to snap objects) are associated with the  
volume and with its mirror. The link object for the original volume points to the  
mirror volume, and the link object for the mirror volume points to the original  
volume. All I/O is directed to both the original volume and its mirror, and a  
synchronization of the mirror from the data in the original volume is started.  
You can use the vxprintcommand to display the state of link objects, which  
appear as type ln. Link objects can have the following states:  
ACTIVE  
The mirror volume has been fully synchronized from the  
original volume. The vxsnapmakecommand can be run to  
create a snapshot volume.  
ATTACHING  
Synchronization of the mirror volume is in progress. The  
vxsnapmakecommand cannot be used to create a snapshot  
volume until the state changes to ACTIVE. The vxsnap  
snapwaitcommand can be used to wait for the  
synchronization to complete.  
BROKEN  
The mirror volume has been detached from the original  
volume because of an I/O error or an unsuccessful attempt to  
grow the mirror volume. The vxrecovercommand can be used  
Download from Www.Somanuals.com. All Manuals Search And Download.  
               
312 Administering volume snapshots  
Cascaded snapshots  
to recover the mirror volume in the same way as for a  
DISABLEDvolume. See “Starting a volume” on page 271.  
If you resize (that is, grow or shrink) a volume, all its ACTIVElinked mirror  
volumes are also resized at the same time. The volume and its mirrors can be in  
the same disk group or in different disk groups. If the operation is successful,  
the volume and its mirrors will have the same size.  
If a volume has been grown, a resynchronization of the grown regions in its  
linked mirror volumes is started, and the links remain in the ATTACHINGstate  
until resynchronization is complete. The vxsnapsnapwaitcommand can be  
used to wait for the state to become ACTIVE.  
When you use the vxsnapmakecommand to create the snapshot volume, this  
removes the link, and establishes a snapshot relationship between the snapshot  
volume and the original volume.  
The vxsnapreattachoperation re-establishes the link relationship between the  
two volumes, and starts a resynchronization of the mirror volume.  
See “Creating and managing linked break-off snapshot volumes” on page 331  
for details of the procedures for creating and using this type of snapshot.  
For information about how to prepare an empty volume for use by full-sized  
instant snapshots, see “Creating a volume for use as a full-sized instant or  
linked break-off snapshot” on page 323.  
Cascaded snapshots  
A snapshot hierarchy known as a snapshot cascade can improve write  
performance for some applications. Instead of having several independent  
snapshots of the volume, it is more efficient to make the older snapshots into  
children of the latest snapshot as shown in Figure 9-4.  
Figure 9-4  
Snapshot cascade  
Most recent  
snapshot  
Oldest  
snapshot  
Original  
volume  
V
Snapshot  
volume  
Sn  
Snapshot  
volume  
Sn-1  
Snapshot  
volume  
S1  
. . .  
A snapshot may be added to a cascade by specifying the infrontofattribute to  
the vxsnapmakecommand when the second and subsequent snapshots in the  
cascade are created. Changes to blocks in the original volume are only written to  
the most recently created snapshot volume in the cascade. If an attempt is made  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Administering volume snapshots 313  
Cascaded snapshots  
to read data from an older snapshot that does not exist in that snapshot, it is  
obtained by searching recursively up the hierarchy of more recent snapshots.  
A snapshot cascade is most likely to be used for regular online backup of a  
volume where space-optimized snapshots are written to disk but not to tape.  
A snapshot cascade improves write performance over the alternative of several  
independent snapshots, and also requires less disk space if the snapshots are  
space-optimized. Only the latest snapshot needs to be updated when the original  
volume is updated. If and when required, the older snapshots can obtain the  
changed data from the most recent snapshot.  
The following points determine whether it is appropriate for an application to  
use a snapshot cascade:  
■
Deletion of a snapshot in the cascade takes time to copy the snapshot’s data  
to the next snapshot in the cascade.  
■
The reliability of a snapshot in the cascade depends on all the newer  
snapshots in the chain. Thus the oldest snapshot in the cascade is the most  
vulnerable.  
■
Reading from a snapshot in the cascade may require data to be fetched from  
one or more other snapshots in the cascade.  
For these reasons, it is recommended that you do not attempt to use a snapshot  
cascade with applications that need to remove or split snapshots from the  
cascade. In such cases, it may be more appropriate to create a snapshot of a  
snapshot as described in the following section.  
See “Adding a snapshot to a cascaded snapshot hierarchy” on page 337 for an  
example of the use of the infrontofattribute.  
Note: Only unsynchronized full-sized or space-optimized instant snapshots are  
usually cascaded. It is of little utility to create cascaded snapshots if the  
infrontofsnapshot volume is fully synchronized (as, for example, with break-  
off type snapshots).  
Creating a snapshot of a snapshot  
For some applications, it may be desirable to create a snapshot of an existing  
snapshot as illustrated in Figure 9-5.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
314 Administering volume snapshots  
Cascaded snapshots  
Figure 9-5  
Creating a snapshot of a snapshot  
vxsnap make source=V  
vxsnap make source=S1  
Original  
volume  
V
Snapshot  
volume  
S1  
Snapshot  
volume  
S2  
Even though the arrangement of the snapshots in this figure appears similar to  
the snapshot hierarchy shown in “Snapshot cascade” on page 312, the  
relationship between the snapshots is not recursive. When reading from the  
snapshot S2, data is obtained directly from the original volume, V, if it does not  
exist in S2itself.  
Such an arrangement may be useful if the snapshot volume, S1, is critical to the  
operation. For example, S1could be used as a stable copy of the original volume,  
V. The additional snapshot volume, S2, can be used to restore the original  
volume if that volume becomes corrupted. For a database, you might need to  
replay a redo log on S2before you could use it to restore V. These steps are  
illustrated in Figure 9-6.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Administering volume snapshots 315  
Cascaded snapshots  
Figure 9-6  
Using a snapshot of a snapshot to restore a database  
1. Create instant snapshot S1 of volume V  
vxsnap make source=V  
Original  
volume  
V
Snapshot  
volume of V:  
S1  
2. Create instant snapshot S2 of S1  
vxsnap make source=S1  
Original  
volume  
V
Snapshot  
volume of V:  
S1  
Snapshot  
volume of S1:  
S2  
3. After contents of V have gone bad, apply the database redo logs to S2  
Apply redo logs  
Original  
volume  
V
Snapshot  
volume of V:  
S1  
Snapshot  
volume of S1:  
S2  
4. Restore contents of V instantly from snapshot S2 and keep S1 as a stable copy  
vxsnap restore V source=S2  
Original  
volume  
V
Snapshot  
volume of V:  
S1  
Snapshot  
volume of S1:  
S2  
If you have configured snapshots in this way, you may wish to make one or more  
of the snapshots into independent volumes. There are two vxsnapcommands  
that you can use to do this:  
■
vxsnapdisdissociates a snapshot volume and turns it into an independent  
volume. The volume to be dissociated must have been fully synchronized  
from its parent. If a snapshot volume has a child snapshot volume, the child  
must also have been fully synchronized. If the command succeeds, the child  
snapshot becomes a snapshot of the original volume. Figure 9-7 illustrates  
the effect of applying this command to snapshots with and without  
dependent snapshots.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
316 Administering volume snapshots  
Cascaded snapshots  
Figure 9-7  
Dissociating a snapshot volume  
vxsnap dis is applied to snapshot S2, which has no snapshots of its own  
Original  
volume  
V
Snapshot  
volume of V:  
S1  
Snapshot  
volume of S1:  
S2  
vxsnap dis S2  
Original  
volume  
V
Snapshot  
volume of V:  
S1  
Volume  
S2  
S1 remains owned by V  
S2 is independent  
vxsnap dis is applied to snapshot S1, which has one snapshot S2  
Original  
volume  
V
Snapshot  
volume of V:  
S1  
Snapshot  
volume of S1:  
S2  
vxsnap dis S1  
Original  
volume  
V
Snapshot  
volume of V:  
S2  
Volume  
S1  
S1 is independent  
S2 is adopted by V  
■
vxsnapsplitdissociates a snapshot and its dependent snapshots from its  
parent volume. The snapshot volume that is to be split must have been fully  
synchronized from its parent volume. This operation is illustrated in  
Figure 9-8.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Administering volume snapshots 317  
Creating multiple snapshots  
Figure 9-8  
Splitting snapshots  
Original  
volume  
V
Snapshot  
volume of V:  
S1  
Snapshot  
volume of S1:  
S2  
vxsnap split S1  
Original  
volume  
V
Snapshot  
volume of S1:  
S2  
Volume  
S1  
S1 is independent  
S2 continues to be a  
snapshot of S1  
Creating multiple snapshots  
To make it easier to create snapshots of several volumes at the same time, both  
the vxsnapmakeand vxassistsnapshotcommands accept more than one  
volume name as their argument.  
For traditional snapshots, you can create snapshots of all the volumes in a single  
disk group by specifying the option -o allvolsto the vxassistsnapshot  
command.  
By default, each replica volume is named SNAPnumber-volume, where number is  
a unique serial number, and volume is the name of the volume for which a  
snapshot is being taken. This default can be overridden by using the option -o  
name=pattern, as described on the vxsnap(1M) and vxassist(1M) manual pages.  
It is also possible to take several snapshots of the same volume. A new  
FastResync change map is produced for each snapshot taken to minimize the  
resynchronization time for each snapshot.  
Restoring the original volume from a snapshot  
For traditional snapshots, the snapshot plex is resynchronized from the data in  
the original volume during a vxassistsnapbackoperation. Alternatively, you  
can choose the snapshot plex as the preferred copy of the data when performing  
a snapback as illustrated in Figure 9-9. Specifying the option -o  
resyncfromreplicato vxassistresynchronizes the original volume from  
the data in the snapshot.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
318 Administering volume snapshots  
Restoring the original volume from a snapshot  
Figure 9-9  
Resynchronizing an original volume from a snapshot  
Refresh on  
snapback  
snapshot  
Original  
volume  
Snapshot  
mirror  
Snapshot  
volume  
-o resyncfromreplica snapback  
Note: The original volume must not be in use during a snapbackoperation that  
specifies the option -oresyncfromreplicato resynchronize the volume from a  
snapshot. Stop any application, such as a database, and unmount any file  
systems that are configured to use the volume.  
For instant snapshots, the vxsnaprestorecommand may be used to restore the  
contents of the original volume from an instant snapshot or from a volume  
derived from an instant snapshot. The volume that is used to restore the original  
volume can either be a true backup of the contents of the original volume at  
some point in time, or it may have been modified in some way (for example, by  
applying a database log replay or by running a file system checking utility such  
as fsck) to create a synthetic replica. All synchronization of the contents of this  
backup or synthetic replica volume must have been completed before the  
original volume can be restored from it. The original volume is immediately  
available for use while its contents are being restored.  
Note: You can perform either a destructive or non-destructive restoration of an  
original volume from an instant snapshot. Only non-destructive restoration is  
possible from a space-optimized snapshot. In this case, the snapshot remains in  
existence after the restoration is complete.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Administering volume snapshots 319  
Creating instant snapshots  
Creating instant snapshots  
Note: You need a full license to use this feature.  
VxVM allows you to make instant snapshots of volumes by using the vxsnap  
command.  
Note: The information in this section also applies to RAID-5 volumes that have  
been converted to a special layered volume layout by the addition of a DCO and  
DCO volume. See “Using a DCO and DCO volume with a RAID-5 volume” on  
page 277 for details.  
A plex in a full-sized instant snapshot requires as much space as the original  
volume. If you instead make a space-optimized instant snapshot of a volume,  
this only requires enough storage to record the original contents of the parent  
volume as they are changed during the life of the snapshot.  
The recommended approach to performing volume backup from the command  
line, or from a script, is to use the vxsnapcommand. The vxsnapprepareand  
maketasks allow you to back up volumes online with minimal disruption to  
users.  
The vxsnappreparestep creates a DCO and DCO volume and associates this  
with the volume. It also enables Persistent FastResync on the volume.  
The vxsnapmakestep creates an instant snapshot that is immediately available  
for making a backup. After the snapshot has been taken, read requests for data  
in the original volume are satisfied by reading either from a non-updated region  
of the original volume, or from the copy of the original contents of an updated  
region that have been recorded by the snapshot.  
Note: Synchronization of a full-sized instant snapshot from the original volume  
is enabled by default. If you specify the syncing=noattribute to vxsnapmake,  
this disables synchronization, and the contents of the instant snapshot are  
unlikely ever to become fully synchronized with the contents of the original  
volume at the point in time that the snapshot was taken. If you wish to move an  
instant snapshot volume to another disk group for export to another machine  
for off-host processing, or to turn it into an independent volume, you must  
ensure that the snapshot volume has been completely synchronized.  
You can immediately retake a full-sized or space-optimized instant snapshot at  
any time by using the vxsnaprefreshcommand. If a fully synchronized instant  
snapshot is required, you must wait for the new resynchronization to complete.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
320 Administering volume snapshots  
Creating instant snapshots  
You can create instant snapshots of volume sets by replacing volume names  
with volume set names in the vxsnapcommand. For more information, see  
“Creating instant snapshots of volume sets” on page 334.  
Note: When using the vxsnapprepareor vxassistmakecommands to make a  
volume ready for instant snapshot operations, if the specified region size  
exceeds half the value of the tunable voliomem_maxpool_sz(see  
“voliomem_maxpool_sz” on page 483), the operation succeeds but gives a  
warning such as the following (for a system where voliomem_maxpool_szis set  
to 12MB):  
VxVM vxassist WARNING V-5-1-0 Specified regionsize is  
larger than the limit on the system  
(voliomem_maxpool_sz/2=6144k).  
If this message is displayed, vxsnapmake, refreshand restoreoperations on  
such volumes fail as they might potentially hang the system. Such volumes can  
be used only for break-off snapshot operations using the reattachand make  
operations.  
To make the volumes usable for instant snapshot operations, use vxsnap  
unprepareon the volume, and then use vxsnapprepareto re-prepare the  
volume with a region size that is less than half the size of  
voliomem_maxpool_sz(in this example, 1MB):  
#vxsnap -g mydg -f unprepare vol1  
#vxsnap -g mydg prepare vol1 regionsize=1M  
To create and manage a snapshot of a volume with the vxsnapcommand, follow  
the procedure in “Preparing to create instant and break-off snapshots” on  
page 321, and then use one of the procedures described in the following  
sections:  
■
■
■
■
“Creating and managing space-optimized instant snapshots” on page 324  
“Creating and managing full-sized instant snapshots” on page 327  
“Creating and managing third-mirror break-off snapshots” on page 329  
“Creating and managing linked break-off snapshot volumes” on page 331  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Administering volume snapshots 321  
Creating instant snapshots  
Preparing to create instant and break-off snapshots  
To prepare a volume for the creation of instant and break-off snapshots  
1
Use the following commands to see if the volume is associated with a  
version 20 data change object (DCO) and DCO volume that allow instant  
snapshots and Persistent FastResync to be used with the volume, and to  
check that FastResync is enabled on the volume:  
# vxprint -g volumedg -F%instant volume  
# vxprint -g volumedg -F%fastresync volume  
If both commands return a value of on, the volume can be used for instant  
snapshot operations, and you should skip to step 3. Otherwise continue  
with step 2.  
2
To prepare a volume for instant snapshots, use the following command:  
# vxsnap [-g diskgroup] prepare volume [regionsize=size] \  
[ndcomirs=number] [alloc=storage_attributes]  
Note: It is only necessary to run the vxsnappreparecommand on a volume  
if it does not already have a version 20 DCO volume (for example, if you  
have run the vxsnapunpreparecommand on the volume). See “Creating a  
volume with a version 20 DCO volume” on page 252, “Preparing a volume  
for DRL and instant snapshots” on page 275 and “Removing support for  
DRL and instant snapshots from a volume” on page 279 for more  
information.  
For example, to prepare the volume, myvol, in the disk group, mydg, use  
the following command:  
# vxsnap -g mydg prepare myvol regionsize=128k ndcomirs=2 \  
alloc=mydg10,mydg11  
This example creates a DCO object and redundant DCO volume with two  
plexes located on disks mydg10and mydg11, and associates them with  
myvol. The region size is also increased to 128KB from the default size of  
64KB. The region size must be a power of 2, and be greater than or equal to  
16KB. A smaller value requires more disk space for the change maps, but  
the finer granularity provides faster resynchronization.  
3
If you need to create several space-optimized instant snapshots for the  
volumes in a disk group, you may find it more convenient to create a single  
shared cache object in the disk group rather than a separate cache object for  
each snapshot. Follow the procedure in “Creating a shared cache object” on  
page 322 to create the cache object.  
For full-sized instant snapshots and linked break-off snapshots, you must  
prepare a volume that is to be used as the snapshot volume. This volume  
must be the same size as the data volume for which the snapshot is being  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
322 Administering volume snapshots  
Creating instant snapshots  
created, and it must also have the same region size. See “Creating a volume  
for use as a full-sized instant or linked break-off snapshot” on page 323 for  
details.  
Creating a shared cache object  
To create a shared cache object  
1
Decide on the following characteristics that you want to allocate to the cache  
volume that underlies the cache object:  
■
The size of the cache volume should be sufficient to record changes to  
the parent volumes during the interval between snapshot refreshes. A  
suggested value is 10% of the total size of the parent volumes for a  
refresh interval of 24 hours.  
■
■
If redundancy is a desired characteristic of the cache volume, it should  
be mirrored. This increases the space that is required for the cache  
volume in proportion to the number of mirrors that it has.  
If the cache volume is mirrored, space is required on at least as many  
disks as it has mirrors. These disks should not be shared with the disks  
used for the parent volumes. The disks should also be chosen to avoid  
impacting I/O performance for critical volumes, or hindering disk  
group split and join operations.  
2
Having decided on its characteristics, use the vxassistcommand to create  
the volume that is to be used for the cache volume. The following example  
creates a mirrored cache volume, cachevol, with size 1GB in the disk  
group, mydg, on the disks mydg16and mydg17:  
# vxassist -g mydg make cachevol 1g layout=mirror \  
init=active mydg16 mydg17  
The attribute init=activeis specified to make the cache volume  
immediately available for use.  
3
Use the vxmakecachecommand to create a cache object on top of the cache  
volume that you created in the previous step:  
# vxmake [-g diskgroup] cache cache_object \  
cachevolname=volume [regionsize=size] [autogrow=on] \  
[highwatermark=hwmk] [autogrowby=agbvalue] \  
[maxautogrow=maxagbvalue]]  
If the region size, regionsize, is specified, it must be a power of 2, and be  
greater than or equal to 16KB (16k). If not specified, the region size of the  
cache is set to 64KB.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                   
Administering volume snapshots 323  
Creating instant snapshots  
Note: All space-optimized snapshots that share the cache must have a  
region size that is equal to or an integer multiple of the region size set on  
the cache. Snapshot creation also fails if the original volume’s region size is  
smaller than the cache’s region size.  
If the region size of a space-optimized snapshot differs from the region size  
of the cache, this can degrade the system’s performance compared to the  
case where the region sizes are the same.  
If the cache is to be allowed to grow in size as required, specify  
autogrow=on. By default, the ability to automatically grow the cache is  
turned off.  
In the following example, the cache object, cobjmydg, is created over the  
cache volume, cachevol, the region size of the cache is set to 32KB, and  
the autogrowfeature is enabled:  
# vxmake -g mydg cache cobjmydg cachevolname=cachevol \  
regionsize=32k autogrow=on  
4
Having created the cache object, use the following command to enable it:  
# vxcache [-g diskgroup] start cache_object  
For example to start the cache object, cobjmydg:  
# vxcache -g mydg start cobjmydg  
For details of how to remove a cache, see “Removing a cache” on page 347.  
Creating a volume for use as a full-sized instant or linked  
break-off snapshot  
To create an empty volume for use by a full-sized instant snapshot or a linked  
break-off snapshot  
1
Use the vxprintcommand on the original volume to find the required size  
for the snapshot volume.  
# LEN=‘vxprint [-g diskgroup] -F%len volume‘  
Note: The command shown in this and subsequent steps assumes that you  
are using a Bourne-type shell such as sh, kshor bash. You may need to  
modify the command for other shells such as cshor tcsh.  
2
3
Use the vxprintcommand on the original volume to discover the name of  
its DCO:  
# DCONAME=‘vxprint [-g diskgroup] -F%dco_name volume‘  
Use the vxprintcommand on the DCO to discover its region size (in blocks):  
# RSZ=‘vxprint [-g diskgroup] -F%regionsz $DCONAME‘  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
324 Administering volume snapshots  
Creating instant snapshots  
4
Use the vxassistcommand to create a volume, snapvol, of the required size  
and redundancy, together with a version 20 DCO volume with the correct  
region size:  
# vxassist [-g diskgroup] make snapvol $LEN \  
[layout=mirror nmirror=number] logtype=dco drl=off \  
dcoversion=20 [ndcomirror=number] regionsz=$RSZ \  
init=active [storage_attributes]  
Specify the same number of DCO mirrors (ndcomirror) as the number of  
mirrors in the volume (nmirror). The init=activeattribute is used to  
make the volume available immediately. You can use storage attributes to  
specify which disks should be used for the volume.  
As an alternative to creating the snapshot volume and its DCO volume in a  
single step, you can first create the volume, and then prepare it for instant  
snapshot operations as shown here:  
# vxassist [-g diskgroup] make snapvol $LEN \  
[layout=mirror nmirror=number] init=active \  
[storage_attributes]  
# vxsnap [-g diskgroup] prepare snapvol [ndcomirs=number] \  
regionsize=$RSZ [storage_attributes]  
Creating and managing space-optimized instant snapshots  
Note: Space-optimized instant snapshots are not suitable for write-intensive  
volumes (such as for database redo logs) because the copy-on-write mechanism  
may degrade the performance of the volume.  
If you intend to split the volume and snapshot into separate disk groups (for  
example, to perform off-host processing), you must use a fully synchronized  
full-sized instant, third-mirror break-off or linked break-off snapshot (which do  
not require a cache object). You cannot use a space-optimized instant snapshot  
for this purpose.  
Creation of space-optimized snapshots that use a shared cache fails if the region  
size specified for the volume is smaller than the region size set on the cache.  
If the region size of a space-optimized snapshot differs from the region size of  
the cache, this can degrade the system’s performance compared to the case  
where the region sizes are the same.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Administering volume snapshots 325  
Creating instant snapshots  
For space-optimized instant snapshots that share a cache object, the specified  
region size must be greater than or equal to the region size specified for the  
cache object. See “Creating a shared cache object” on page 322 for details.  
The attributes for a snapshot are specified as a tuple to the vxsnapmake  
command. This command accepts multiple tuples. One tuple is required for each  
snapshot that is being created. Each element of a tuple is separated from the  
next by a slash character (/). Tuples are separated by white space.  
To create and manage a space-optimized instant snapshot  
1
Use the vxsnapmakecommand to create a space-optimized instant  
snapshot. This snapshot can be created by using an existing cache object in  
the disk group, or a new cache object can be created for its use.  
◆
To create a space-optimized instant snapshot, snapvol, that uses a named  
shared cache object:  
# vxsnap [-g diskgroup] make source=vol/newvol=snapvol\  
/[cache=cacheobject] [alloc=storage_attributes]  
For example, to create the space-optimized instant snapshot, snap3myvol,  
of the volume, myvol, in the disk group, mydg, on the disk mydg14, and  
which uses the shared cache object, cobjmydg, use the following command:  
# vxsnap -g mydg make source=myvol/newvol=snap3myvol\  
/cache=cobjmydg alloc=mydg14  
For details of how to create a shared cache object, see “Creating a shared  
cache object” on page 322.  
◆
To create a space-optimized instant snapshot, snapvol, and also create a  
cache object for it to use:  
# vxsnap [-g diskgroup] make source=vol/newvol=snapvol\  
[/cachesize=size][/autogrow=yes][/ncachemirror=number]\  
[alloc=storage_attributes]  
The cachesizeattribute determines the size of the cache relative to the  
size of the volume. The autogrowattribute determines whether VxVM will  
automatically enlarge the cache if it is in danger of overflowing. By default,  
the cache is not grown.  
Note: If autogrowis enabled, but the cache cannot be grown, VxVM  
disables the oldest and largest snapshot that is using the same cache, and  
releases its cache space for use.  
The ncachemirrorattribute specifies the number of mirrors to create in the  
cache volume. For backup purposes, the default value of 1 should be  
sufficient.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
326 Administering volume snapshots  
Creating instant snapshots  
For example, to create the space-optimized instant snapshot, snap4myvol,  
of the volume, myvol, in the disk group, mydg, on the disk mydg15, and  
which uses a newly allocated cache object that is 1GB in size, but which can  
automatically grow in size, use the following command:  
# vxsnap -g mydg make source=myvol/new=snap4myvol\  
/cachesize=1g/autogrow=yes alloc=mydg15  
Note: If a cache is created implicitly by specifying cachesize, and  
ncachemirroris specified to be greater than 1, a DCO is attached to the  
cache volume to enable dirty region logging (DRL). DRL allows fast recovery  
of the cache backing store after a system crash. The DCO is allocated on the  
same disks as those that are occupied by the DCO of the source volume. This  
is done to allow the cache and the source volume to remain in the same disk  
group for disk group move, split and join operations.  
2
3
Use fsck(or some utility appropriate for the application running on the  
volume) to clean the temporary volume’s contents. For example, you can use  
this command with a VxFS file system:  
# fsck -F vxfs /dev/vx/rdsk/diskgroup/snapshot  
If you require a backup of the data in the snapshot, use an appropriate  
utility or operating system command to copy the contents of the snapshot to  
tape, or to some other backup medium.  
4
You now have the following choices of what to do with a space-optimized  
instant snapshot:  
■
Refresh the contents of the snapshot. This creates a new point-in-time  
image of the original volume ready for another backup. If  
synchronization was already in progress on the snapshot, this  
operation may result in large portions of the snapshot having to be  
resynchronized. See “Refreshing an instant snapshot” on page 337 for  
details.  
■
Restore the contents of the original volume from the snapshot volume.  
The space-optimized instant snapshot remains intact at the end of the  
operation. See “Restoring a volume from an instant snapshot” on  
page 340 for details.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Administering volume snapshots 327  
Creating instant snapshots  
Creating and managing full-sized instant snapshots  
Note: Full-sized instant snapshots are not suitable for write-intensive volumes  
(such as for database redo logs) because the copy-on-write mechanism may  
degrade the performance of the volume.  
For full-sized instant snapshots, you must prepare a volume that is to be used as  
the snapshot volume. This must be the same size as the volume for which the  
snapshot is being created, and it must also have the same region size. See  
“Creating a volume for use as a full-sized instant or linked break-off snapshot”  
on page 323 for details.  
The attributes for a snapshot are specified as a tuple to the vxsnapmake  
command. This command accepts multiple tuples. One tuple is required for each  
snapshot that is being created. Each element of a tuple is separated from the  
next by a slash character (/). Tuples are separated by white space.  
To create and manage a full-sized instant snapshot  
1
To create a full-sized instant snapshot, use the following form of the vxsnap  
makecommand:  
# vxsnap [-g diskgroup] make source=volume/snapvol=snapvol\  
[/snapdg=snapdiskgroup] [/syncing=off]  
The command specifies the volume, snapvol, that you prepared earlier.  
For example, to use the prepared volume, snap1myvol, as the snapshot for  
the volume, myvol, in the disk group, mydg, use the following command:  
# vxsnap -g mydg make source=myvol/snapvol=snap1myvol  
For full-sized instant snapshots that are created from an empty volume,  
background synchronization is enabled by default (equivalent to specifying  
the syncing=onattribute). If you want to move a snapshot into a separate  
disk group, or to turn it into an independent volume, you must wait for its  
contents to be synchronized with those of its parent volume.  
You can use the vxsnapsyncwaitcommand to wait for the synchronization  
of the snapshot volume to be completed, as shown here:  
# vxsnap [-g diskgroup] syncwait snapvol  
For example, you would use the following command to wait for  
synchronization to finish on the snapshot volume, snap2myvol:  
# vxsnap -g mydg syncwait snap2myvol  
This command exits (with a return code of zero) when synchronization of  
the snapshot volume is complete. The snapshot volume may then be moved  
to another disk group or turned into an independent volume.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
328 Administering volume snapshots  
Creating instant snapshots  
If required, you can use the following command to test if the  
synchronization of a volume is complete:  
# vxprint [-g diskgroup] -F%incomplete snapvol  
This command returns the value offif synchronization of the volume,  
snapvol, is complete; otherwise, it returns the value on.  
You can also use the vxsnapprintcommand to check on the progress of  
synchronization as described in “Displaying instant snapshot information”  
on page 342.  
See “Controlling instant snapshot synchronization” on page 344 for more  
information.  
If you do not want to move the snapshot into a separate disk group, or to  
turn it into an independent volume, specify the syncing=offattribute. This  
avoids creating unnecessary system overhead. For example, to turn off  
synchronization when creating the snapshot of the volume, myvol, you  
would use the following form of the vxsnapmakecommand:  
# vxsnap -g mydg make source=myvol/snapvol=snap1myvol\  
/syncing=off  
2
3
Use fsck(or some utility appropriate for the application running on the  
volume) to clean the temporary volume’s contents. For example, you can use  
this command with a VxFS file system:  
# fsck -F vxfs /dev/vx/rdsk/diskgroup/snapshot  
If you require a backup of the data in the snapshot, use an appropriate  
utility or operating system command to copy the contents of the snapshot to  
tape, or to some other backup medium.  
4
You now have the following choices of what to do with a full-sized instant  
snapshot:  
■
Refresh the contents of the snapshot. This creates a new point-in-time  
image of the original volume ready for another backup. If  
synchronization was already in progress on the snapshot, this  
operation may result in large portions of the snapshot having to be  
resynchronized. See “Refreshing an instant snapshot” on page 337 for  
details.  
■
■
Reattach some or all of the plexes of the snapshot volume with the  
original volume. See “Reattaching an instant snapshot” on page 338 for  
details.  
Restore the contents of the original volume from the snapshot volume.  
You can choose whether none, a subset, or all of the plexes of the  
snapshot volume are returned to the original volume as a result of the  
operation. See “Restoring a volume from an instant snapshot” on  
page 340 for details.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Administering volume snapshots 329  
Creating instant snapshots  
■
■
Dissociate the snapshot volume entirely from the original volume. This  
may be useful if you want to use the copy for other purposes such as  
testing or report generation. If desired, you can delete the dissociated  
volume. See “Dissociating an instant snapshot” on page 340 for details.  
If the snapshot is part of a snapshot hierarchy, you can also choose to  
split this hierarchy from its parent volumes. See “Splitting an instant  
snapshot hierarchy” on page 341 for details.  
Creating and managing third-mirror break-off snapshots  
Note: Break-off snapshots are suitable for write-intensive volumes, such as  
database redo logs.  
To turn one or more existing plexes in a volume into a break-off instant  
snapshot volume, the volume must be a non-layered volume with a mirroror  
mirror-stripelayout, or a RAID-5 volume that you have converted to a  
special layered volume (see “Using a DCO and DCO volume with a RAID-5  
volume” on page 277) and then mirrored. The plexes in a volume with a  
stripe-mirrorlayout are mirrored at the subvolume level, and cannot be  
broken off.  
The attributes for a snapshot are specified as a tuple to the vxsnapmake  
command. This command accepts multiple tuples. One tuple is required for each  
snapshot that is being created. Each element of a tuple is separated from the  
next by a slash character (/). Tuples are separated by white space.  
To create and manage a third-mirror break-off snapshot  
1
To create the snapshot, you can either take some of the existing ACTIVE  
plexes in the volume, or you can use the following command to add new  
snapshot mirrors to the volume:  
# vxsnap [-b] [-g diskgroup] addmir volume [nmirror=N] \  
[alloc=storage_attributes]  
By default, the vxsnapaddmircommand adds one snapshot mirror to a  
volume unless you use the nmirrorattribute to specify a different number  
of mirrors. The mirrors remain in the SNAPATTstate until they are fully  
synchronized. The -boption can be used to perform the synchronization in  
the background. Once synchronized, the mirrors are placed in the  
SNAPDONEstate.  
For example, the following command adds 2 mirrors to the volume, vol1,  
on disks mydg10and mydg11:  
# vxsnap -g mydg addmir vol1 nmirror=2 alloc=mydg10,mydg11  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
330 Administering volume snapshots  
Creating instant snapshots  
If you specify the -boption to the vxsnapaddmircommand, you can use the  
vxsnapsnapwaitcommand to wait for synchronization of the snapshot  
plexes to complete, as shown in this example:  
# vxsnap -g mydg snapwait vol1 nmirror=2  
2
To create a third-mirror break-off snapshot, use the following form of the  
vxsnapmakecommand.  
# vxsnap [-g diskgroup] make source=volume[/newvol=snapvol]\  
{/plex=plex1[,plex2,...]|/nmirror=number]}  
Either of the following attributes may be specified to create the new  
snapshot volume, snapvol, by breaking off one or more existing plexes in  
the original volume:  
plex  
Specifies the plexes in the existing volume that are to be broken  
off. This attribute can only be used with plexes that are in the  
ACTIVEstate.  
nmirror Specifies how many plexes are to be broken off. This attribute  
can only be used with plexes that are in the SNAPDONEstate.  
(Such plexes could have been added to the volume by using the  
vxsnapaddmircommand.)  
Snapshots that are created from one or more ACTIVEor SNAPDONEplexes  
in the volume are already synchronized by definition.  
For backup purposes, a snapshot volume with one plex should be sufficient.  
For example, to create the instant snapshot volume, snap2myvol, of the  
volume, myvol, in the disk group, mydg, from a single existing plex in the  
volume, use the following command:  
# vxsnap -g mydg make source=myvol/newvol=snap2myvol\  
/nmirror=1  
The next example shows how to create a mirrored snapshot from two  
existing plexes in the volume:  
# vxsnap -g mydg make source=myvol/newvol=snap2myvol\  
/plex=myvol-03,myvol-04  
3
4
Use fsck(or some utility appropriate for the application running on the  
volume) to clean the temporary volume’s contents. For example, you can use  
this command with a VxFS file system:  
# fsck -F vxfs /dev/vx/rdsk/diskgroup/snapshot  
If you require a backup of the data in the snapshot, use an appropriate  
utility or operating system command to copy the contents of the snapshot to  
tape, or to some other backup medium.  
5
You now have the following choices of what to do with a third-mirror break-  
off snapshot:  
■
Refresh the contents of the snapshot. This creates a new point-in-time  
image of the original volume ready for another backup. If  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Administering volume snapshots 331  
Creating instant snapshots  
synchronization was already in progress on the snapshot, this  
operation may result in large portions of the snapshot having to be  
resynchronized. See “Refreshing an instant snapshot” on page 337 for  
details.  
■
■
Reattach some or all of the plexes of the snapshot volume with the  
original volume. See “Reattaching an instant snapshot” on page 338 for  
details.  
Restore the contents of the original volume from the snapshot volume.  
You can choose whether none, a subset, or all of the plexes of the  
snapshot volume are returned to the original volume as a result of the  
operation. See “Restoring a volume from an instant snapshot” on  
page 340 for details.  
■
■
Dissociate the snapshot volume entirely from the original volume. This  
may be useful if you want to use the copy for other purposes such as  
testing or report generation. If desired, you can delete the dissociated  
volume. See “Dissociating an instant snapshot” on page 340 for details.  
If the snapshot is part of a snapshot hierarchy, you can also choose to  
split this hierarchy from its parent volumes. See “Splitting an instant  
snapshot hierarchy” on page 341 for details.  
Creating and managing linked break-off snapshot volumes  
Note: Break-off snapshots are suitable for write-intensive volumes, such as  
database redo logs.  
For linked break-off snapshots, you must prepare a volume that is to be used as  
the snapshot volume. This must be the same size as the volume for which the  
snapshot is being created, and it must also have the same region size. See  
“Creating a volume for use as a full-sized instant or linked break-off snapshot”  
on page 323 for details.  
The attributes for a snapshot are specified as a tuple to the vxsnapmake  
command. This command accepts multiple tuples. One tuple is required for each  
snapshot that is being created. Each element of a tuple is separated from the  
next by a slash character (/). Tuples are separated by white space.  
To create and manage a linked break-off snapshot  
1
Use the following command to link the prepared snapshot volume, snapvol,  
to the data volume:  
# vxsnap [-g diskgroup] [-b] addmir volume mirvol=snapvol \  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
332 Administering volume snapshots  
Creating instant snapshots  
[mirdg=snapdg]  
The optional mirdgattribute can be used to specify the snapshot volume’s  
current disk group, snapdg. The -boption can be used to perform the  
synchronization in the background. If the -boption is not specified, the  
command does not return until the link becomes ACTIVE.  
For example, the following command links the prepared volume,  
prepsnap, in the disk group, mysnapdg, to the volume, vol1, in the disk  
group, mydg:  
# vxsnap -g mydg -b addmir vol1 mirvol=prepsnap \  
mirdg=mysnapdg  
If the -boption is specified, you can use the vxsnapsnapwaitcommand to  
wait for the synchronization of the linked snapshot volume to complete, as  
shown in this example:  
# vxsnap -g mydg snapwait vol1 mirvol=prepsnap \  
mirdg=mysnapvoldg  
2
To create a linked break-off snapshot, use the following form of the vxsnap  
makecommand.  
# vxsnap [-g diskgroup] make source=volume/snapvol=snapvol\  
[/snapdg=snapdiskgroup]  
The snapdgattribute must be used to specify the snapshot volume’s disk  
group if this is different from that of the data volume.  
For example, to use the prepared volume, prepsnap, as the snapshot for  
the volume, vol1, in the disk group, mydg, use the following command:  
# vxsnap -g mydg make \  
source=vol1/snapvol=prepsnap/snapdg=mysnapdg  
3
4
Use fsck(or some utility appropriate for the application running on the  
volume) to clean the temporary volume’s contents. For example, you can use  
this command with a VxFS file system:  
# fsck -F vxfs /dev/vx/rdsk/diskgroup/snapshot  
If you require a backup of the data in the snapshot, use an appropriate  
utility or operating system command to copy the contents of the snapshot to  
tape, or to some other backup medium.  
5
You now have the following choices of what to do with a linked break-off  
snapshot:  
■
Refresh the contents of the snapshot. This creates a new point-in-time  
image of the original volume ready for another backup. If  
synchronization was already in progress on the snapshot, this  
operation may result in large portions of the snapshot having to be  
resynchronized. See “Refreshing an instant snapshot” on page 337 for  
details.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Administering volume snapshots 333  
Creating instant snapshots  
Note: This operation is not possible if the linked volume and snapshot  
are in different disk groups.  
■
■
Reattach the snapshot volume with the original volume. See  
“Reattaching a linked break-off snapshot volume” on page 339 for  
details.  
Dissociate the snapshot volume entirely from the original volume. This  
may be useful if you want to use the copy for other purposes such as  
testing or report generation. If desired, you can delete the dissociated  
volume. See “Dissociating an instant snapshot” on page 340 for details.  
■
If the snapshot is part of a snapshot hierarchy, you can also choose to  
split this hierarchy from its parent volumes. See “Splitting an instant  
snapshot hierarchy” on page 341 for details.  
Creating multiple instant snapshots  
To make it easier to create snapshots of several volumes at the same time, the  
vxsnapmakecommand accepts multiple tuples that define the source and  
snapshot volumes names as their arguments. For example, to create three  
instant snapshots, each with the same redundancy, from specified storage, the  
following form of the command can be used:  
# vxsnap [-g diskgroup] make source=vol1/snapvol=snapvol1\  
source=vol2/snapvol=snapvol2 source=vol3/snapvol=snapvol3  
The snapshot volumes (snapvol1, snapvol2 and so on) must have been prepared  
in advance as described in “Creating a volume for use as a full-sized instant or  
linked break-off snapshot” on page 323. The specified source volumes (vol1, vol2  
and so on) may be the same volume or they can be different volumes.  
If all the snapshots are to be space-optimized and to share the same cache, the  
following form of the command can be used:  
# vxsnap [-g diskgroup] make \  
source=vol1/newvol=snapvol1/cache=cacheobj \  
source=vol2/newvol=snapvol2/cache=cacheobj \  
source=vol3/newvol=snapvol3/cache=cacheobj \  
[alloc=storage_attributes]  
The vxsnapmakecommand also allows the snapshots to be of different types,  
have different redundancy, and be configured from different storage, as shown  
here:  
# vxsnap [-g diskgroup] make source=vol1/snapvol=snapvol1 \  
source=vol2[/newvol=snapvol2]/cache=cacheobj\  
[/alloc=storage_attributes2][/nmirror=number2]  
source=vol3[/newvol=snapvol3][/alloc=storage_attributes3]\  
/nmirror=number3  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
334 Administering volume snapshots  
Creating instant snapshots  
In this example, snapvol1 is a full-sized snapshot that uses a prepared volume,  
snapvol2 is a space-optimized snapshot that uses a prepared cache, and  
snapvol3 is a break-off full-sized snapshot that is formed from plexes of the  
original volume.  
An example of where you might want to create mixed types of snapshots at the  
same time is when taking snapshots of volumes containing database redo logs  
and database tables:  
# vxsnap -g mydg make \  
source=logv1/newvol=snplogv1/drl=sequential/nmirror=1 \  
source=logv2/newvol=snplogv2/drl=sequential/nmirror=1 \  
source=datav1/newvol=snpdatav1/cache=mydgcobj/drl=on \  
source=datav2/newvol=snpdatav2/cache=mydgcobj/drl=on  
In this example, sequential DRL is enabled for the snapshots of the redo log  
volumes, and normal DRL is applied to the snapshots of the volumes that  
contain the database tables. The two space-optimized snapshots are configured  
to share the same cache object in the disk group. Also note that break-off  
snapshots are used for the redo logs as such volumes are write intensive.  
Creating instant snapshots of volume sets  
Volume set names can be used in place of volume names with the following  
vxsnapoperations on instant snapshots: addmir, dis, make, prepare, reattach,  
refresh, restore, rmmir, split, syncpause, syncresume, syncstart, syncstop,  
syncwait, and unprepare.  
The procedure for creating an instant snapshot of a volume set is the same as  
that for a standalone volume. However, there are certain restrictions if a full-  
sized instant snapshot is to be created from a prepared volume set. A full-sized  
instant snapshot of a volume set must itself be a volume set with the same  
number of volumes, and the same volume sizes and index numbers as the  
parent. For example, if a volume set contains three volumes with sizes 1GB, 2GB  
and 3GB, and indexes 0, 1 and 2 respectively, then the snapshot volume set must  
have three volumes with the same sizes matched to the same set of index  
numbers. The corresponding volumes in the parent and snapshot volume sets  
are also subject to the same restrictions as apply between standalone volumes  
and their snapshots.  
You can use the vxvsetlistcommand to verify that the volume sets have  
identical characteristics as shown in this example:  
# vxvset -g mydg list vset1  
VOLUME  
vol_0  
vol_1  
vol_2  
INDEX  
LENGTH  
204800  
409600  
614400  
KSTATE  
CONTEXT  
0
1
2
ENABLED  
ENABLED  
ENABLED  
-
-
-
# vxvset -g mydg list snapvset1  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Administering volume snapshots 335  
Creating instant snapshots  
VOLUME  
svol_0  
svol_1  
svol_2  
INDEX  
LENGTH  
204800  
409600  
614400  
KSTATE  
CONTEXT  
0
1
2
ENABLED  
ENABLED  
ENABLED  
-
-
-
A full-sized instant snapshot of a volume set can be created using a prepared  
volume set in which each volume is the same size as the corresponding volume  
in the parent volume set. Alternatively, you can use the nmirrorsattribute to  
specify the number of plexes that are to be broken off provided that sufficient  
plexes exist for each volume in the volume set.  
The following example shows how to prepare a source volume set, vset1, and  
an identical volume set, snapvset1, which is then used to create the snapshot:  
# vxsnap -g mydg prepare vset1  
# vxsnap -g mydg prepare snapvset1  
# vxsnap -g mydg make source=vset1/snapvol=snapvset1  
To create a full-sized third-mirror break-off snapshot, you must ensure that  
each volume in the source volume set contains sufficient plexes. The following  
example shows how to achieve this by using the vxsnapcommand to add the  
required number of plexes before breaking off the snapshot:  
# vxsnap -g mydg prepare vset2  
# vxsnap -g mydg addmir vset2 nmirror=1  
# vxsnap -g mydg make source=vset2/newvol=snapvset2/nmirror=1  
See “Adding snapshot mirrors to a volume” on page 336 for more information  
about adding plexes to volumes or to volume sets.  
To create a space-optimized instant snapshot of a volume set, the commands are  
again identical to those for a standalone volume as shown in these examples:  
# vxsnap -g mydg prepare vset3  
# vxsnap -g mydg make source=vset3/newvol=snapvset3/  
cachesize=20m  
# vxsnap -g mydg prepare vset4  
# vxsnap -g mydg make source=vset4/newvol=snapvset4/cache=mycobj  
Here a new cache object is created for the volume set, vset3, and an existing  
cache object, mycobj, is used for vset4.  
See “Creating and administering volume sets” on page 361 for more information  
on creating and administering volume sets.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
336 Administering volume snapshots  
Creating instant snapshots  
Adding snapshot mirrors to a volume  
If you are going to create a full-sized break-off snapshot volume, you can use the  
following command to add new snapshot mirrors to a volume:  
# vxsnap [-b] [-g diskgroup] addmir volume|volume_set \  
[nmirror=N] [alloc=storage_attributes]  
Note: The volume must have been prepared using the vxsnappreparecommand  
as described in “Preparing a volume for DRL and instant snapshots” on  
page 275.  
If a volume set name is specified instead of a volume, the specified number of  
plexes is added to each volume in the volume set.  
By default, the vxsnapaddmircommand adds one snapshot mirror to a volume  
unless you use the nmirrorattribute to specify a different number of mirrors.  
The mirrors remain in the SNAPATTstate until they are fully synchronized. The  
-boption can be used to perform the synchronization in the background. Once  
synchronized, the mirrors are placed in the SNAPDONEstate.  
For example, the following command adds 2 mirrors to the volume, vol1, on  
disks mydg10and mydg11:  
# vxsnap -g mydg addmir vol1 nmirror=2 alloc=mydg10,mydg11  
Note: This command is similar in usage to the vxassistsnapstartcommand,  
and supports the traditional third-mirror break-off snapshot model. As such, it  
does not provide an instant snapshot capability.  
Once you have added one or more snapshot mirrors to a volume, you can use the  
vxsnapmakecommand with either the nmirrorattribute or the plexattribute  
to create the snapshot volumes.  
Removing a snapshot mirror  
To remove a single snapshot mirror from a volume, use this command:  
# vxsnap [-g diskgroup] rmmir volume|volume_set  
For example, the following command removes a snapshot mirror from the  
volume, vol1:  
# vxsnap -g mydg rmmir vol1  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Administering volume snapshots 337  
Creating instant snapshots  
Note: This command is similar in usage to the vxassistsnapabortcommand.  
If a volume set name is specified instead of a volume, a mirror is removed from  
each volume in the volume set.  
Removing a linked break-off snapshot volume  
To remove a linked break-off snapshot volume from a volume, use this  
command:  
# vxsnap [-g diskgroup] rmmir volume|volume_set mirvol=snapvol \  
[mirdg=snapdiskgroup]  
The mirvoland optional mirdgattributes specify the snapshot volume, snapvol,  
and its disk group, snapdiskgroup. For example, the following command  
removes a linked snapshot volume, prepsnap, from the volume, vol1:  
# vxsnap -g mydg rmmir vol1 mirvol=prepsnap mirdg=mysnapdg  
Adding a snapshot to a cascaded snapshot hierarchy  
To create a snapshot and push it onto a snapshot hierarchy between the original  
volume and an existing snapshot volume, specify the name of the existing  
snapshot volume as the value of the infrontofattribute to the vxsnapmake  
command. The following example shows how to place the space-optimized  
snapshot, thurs_bu, of the volume, dbvol, in front of the earlier snapshot,  
wed_bu:  
# vxsnap -g dbdg make source=dbvol/newvol=thurs_bu/\  
infrontof=wed_bu/cache=dbdgcache  
Similarly, the next snapshot that is taken, fri_bu, is placed in front of  
thurs_bu:  
# vxsnap -g dbdg make source=dbvol/newvol=fri_bu/\  
infrontof=thurs_bu/cache=dbdgcache  
For more information on the application of cascaded snapshots, see “Cascaded  
snapshots” on page 312.  
Refreshing an instant snapshot  
Refreshing an instant snapshot replaces it with another point-in-time copy of a  
parent volume. To refresh one or more snapshots and make them immediately  
available for use, use the following command:  
# vxsnap [-g diskgroup] refresh snapvolume|snapvolume_set \  
source=volume|volume_set [[snapvol2 source=vol2]...] \  
[syncing=yes|no]  
If the source volume is not specified, the immediate parent of the snapshot is  
used. For full-sized instant snapshots, resynchronization is started by default.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                   
338 Administering volume snapshots  
Creating instant snapshots  
To disable resynchronization, specify the syncing=noattribute. This attribute is  
not supported for space-optimized snapshots.  
Note: The snapshot being refreshed must not be open to any application. For  
example, any file system configured on the volume must first be unmounted.  
It is possible to refresh a volume from an unrelated volume provided that their  
sizes are compatible.  
You can use the vxsnapsyncwaitcommand to wait for the synchronization of  
the snapshot volume to be completed, as shown here:  
# vxsnap [-g diskgroup] syncwait snapvol  
See “Controlling instant snapshot synchronization” on page 344 for more  
information.  
Reattaching an instant snapshot  
Note: This operation is not supported for space-optimized instant snapshots.  
Using the following command, some or all plexes of an instant snapshot may be  
reattached to the specified original volume, or to a source volume in the  
snapshot hierarchy above the snapshot volume:  
# vxsnap [-g diskgroup] reattach snapvolume|snapvolume_set \  
source=volume|volume_set [nmirror=number]  
By default, all the plexes are reattached, which results in the removal of the  
snapshot. If required, the number of plexes to be reattached may be specified as  
the value assigned to the nmirrorattribute.  
Note: The snapshot being reattached must not be open to any application. For  
example, any file system configured on the snapshot volume must first be  
unmounted.  
It is possible to reattach a volume to an unrelated volume provided that their  
volume sizes and region sizes are compatible.  
For example the following command reattaches one plex from the snapshot  
volume, snapmyvol, to the volume, myvol:  
# vxsnap -g mydg reattach snapmyvol source=myvol nmirror=1  
While the reattached plexes are being resynchronized from the data in the  
parent volume, they remain in the SNAPTMPstate. After resynchronization is  
complete, the plexes are placed in the SNAPDONEstate. You can use the vxsnap  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Administering volume snapshots 339  
Creating instant snapshots  
snapwaitcommand (but not vxsnapsyncwait) to wait for the resynchronization  
of the reattached plexes to complete, as shown here:  
# vxsnap -g mydg snapwait myvol nmirror=1  
Note: If the volume and its snapshot have both been resized (to an identical  
smaller or larger size) before performing the reattachment, a fast  
resynchronization can still be performed. A full resynchronization is not  
required. Version 20 DCO volumes are resized proportionately when the  
associated data volume is resized. For version 0 DCO volumes, the FastResync  
maps stay the same size, but the region size is recalculated, and the locations of  
the dirty bits in the existing maps are adjusted. In both cases, new regions are  
marked as dirty in the maps.  
Reattaching a linked break-off snapshot volume  
Unlike other types of snapshot, the reattachment operation for linked break-off  
snapshot volumes does not return the plexes of the snapshot volume to the  
parent volume. The link relationship is re-established that makes the snapshot  
volume a mirror of the parent volume, and this allows the snapshot data to be  
resynchronized. However, the snapshot volume is only readopted by its parent  
volume if they are both in the same disk group.  
To reattach a linked break-off snapshot volume, use the following form of the  
vxsnapreattachcommand:  
# vxsnap [-g snapdiskgroup] reattach snapvolume|snapvolume_set \  
source=volume|volume_set [sourcedg=diskgroup]  
The sourcedgattribute must be used to specify the data volume’s disk group if  
this is different from the snapshot volume’s disk group, snapdiskgroup.  
Note: The snapshot being reattached must not be open to any application. For  
example, any file system configured on the snapshot volume must first be  
unmounted.  
It is possible to reattach a volume to an unrelated volume provided that their  
sizes and region sizes are compatible.  
For example the following command reattaches the snapshot volume,  
prepsnap, in the disk group, snapdg, to the volume, myvol, in the disk group,  
mydg:  
# vxsnap -g snapdg reattach prepsnap source=myvol sourcedg=mydg  
After resynchronization of the snapshot volume is complete, the link is placed  
in the ACTIVEstate. You can use the vxsnapsnapwaitcommand (but not vxsnap  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
340 Administering volume snapshots  
Creating instant snapshots  
syncwait) to wait for the resynchronization of the reattached volume to  
complete, as shown here:  
# vxsnap -g snapdg snapwait myvol mirvol=prepsnap  
Restoring a volume from an instant snapshot  
It may sometimes be desirable to reinstate the contents of a volume from a  
backup or modified replica in a snapshot volume. The following command may  
be used to restore one or more volumes from the specified snapshots:  
# vxsnap [-g diskgroup] restore volume|volume_set \  
source=snapvolume|snapvolume_set \  
[[volume2|volume_set2 source=snapvolume2|snapvolume_set2]...]\  
[destroy=yes|no] [syncing=yes|no] [nmirror=number]  
For a full-sized instant snapshot, some or all of its plexes may be reattached to  
the parent volume or to a specified source volume in the snapshot hierarchy  
above the snapshot volume. If destroy=yesis specified, all the plexes of the full-  
sized instant snapshot are reattached and the snapshot volume is removed.  
For a space-optimized instant snapshot, the cached data is used to recreate the  
contents of the specified volume. The space-optimized instant snapshot remains  
unchanged by the restoreoperation.  
Note: For this operation to succeed, the volume that is being restored and the  
snapshot volume must not be open to any application. For example, any file  
systems that are configured on either volume must first be unmounted.  
It is not possible to restore a volume from an unrelated volume.  
The destroyand nmirrorattributes are not supported for space-optimized  
instant snapshots.  
The following example demonstrates how to restore the volume, myvol, from  
the space-optimized snapshot, snap3myvol.  
# vxsnap -g mydg restore myvol source=snap3myvol  
Dissociating an instant snapshot  
The following command breaks the association between a full-sized instant  
snapshot volume, snapvol, and its parent volume, so that the snapshot may be  
used as an independent volume:  
# vxsnap [-f] [-g diskgroup] dis snapvolume|snapvolume_set  
This operation fails if the snapshot, snapvol, has a snapshot hierarchy below it  
that contains unsynchronized snapshots. If this happens, the dependent  
snapshots must be fully synchronized from snapvol. When no dependent  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
Administering volume snapshots 341  
Creating instant snapshots  
snapshots remain, snapvol may be dissociated. The snapshot hierarchy is then  
adopted by snapvol’s parent volume.  
Note: To be usable after dissociation, the snapshot volume and any snapshots in  
the hierarchy must have been fully synchronized. See “Controlling instant  
snapshot synchronization” on page 344 for more information. In addition, you  
cannot dissociate a snapshot if synchronization of any of the dependent  
snapshots in the hierarchy is incomplete. If an incomplete snapshot is  
dissociated, it is unusable and should be deleted as described in “Removing an  
instant snapshot” on page 341.  
The following command dissociates the snapshot, snap2myvol, from its parent  
volume:  
# vxsnap -g mydg dis snap2myvol  
Note: When applied to a volume set or to a component volume of a volume set,  
this operation can result in inconsistencies in the snapshot hierarchy in the case  
of a system crash or hardware failure. If the operation is applied to a volume set,  
the -f(force) option must be specified.  
Removing an instant snapshot  
When you have dissociated a full-sized instant snapshot, you can use the vxedit  
command to delete it altogether, as shown in this example:  
# vxedit -g mydg -r rm snap2myvol  
You can also use this command to remove a space-optimized instant snapshot  
from its cache. For details of how to remove a cache, see “Removing a cache” on  
page 347.  
Splitting an instant snapshot hierarchy  
Note: This operation is not supported for space-optimized instant snapshots.  
The following command breaks the association between a snapshot hierarchy  
that has the snapshot volume, snapvol, at its head, and its parent volume, so that  
the snapshot hierarchy may be used independently of the parent volume:  
# vxsnap [-f] [-g diskgroup] split snapvolume|snapvolume_set  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
342 Administering volume snapshots  
Creating instant snapshots  
Note: The topmost snapshot volume in the hierarchy must have been fully  
synchronized for this command to succeed. Snapshots that are lower down in  
the hierarchy need not have been fully resynchronized. See “Controlling instant  
snapshot synchronization” on page 344 for more information.  
The following command splits the snapshot hierarchy under snap2myvolfrom  
its parent volume:  
# vxsnap -g mydg split snap2myvol  
Note: When applied to a volume set or to a component volume of a volume set,  
this operation can result in inconsistencies in the snapshot hierarchy in the case  
of a system crash or hardware failure. If the operation is applied to a volume set,  
the -f(force) option must be specified.  
Displaying instant snapshot information  
The vxsnapprintcommand may be used to display information about the  
snapshots that are associated with a volume.  
# vxsnap [-g diskgroup] print [vol]  
This command shows the percentage progress of the synchronization of a  
snapshot or volume. If no volume is specified, information about the snapshots  
for all the volumes in a disk group is displayed. The following example shows a  
volume, vol1, which has a full-sized snapshot, snapvol1whose contents have  
not been synchronized with vol1:  
# vxsnap -g mydg print  
NAME  
SNAPOBJECT  
TYPE  
PARENT  
SNAPSHOT  
%DIRTY  
%VALID  
vol1  
--  
volume --  
--  
--  
1.30  
1.30  
100  
--  
1.30  
snapvol1_snp1volume --  
snapvol1 vol1_snp1 volume vol1  
snapvol1  
--  
The %DIRTYvalue for snapvol1shows that its contents have changed by 1.30%  
when compared with the contents of vol1. As snapvol1has not been  
synchronized with vol1, the %VALIDvalue is the same as the %DIRTYvalue. If  
the snapshot were partly synchronized, the %VALIDvalue would lie between the  
%DIRTYvalue and 100%. If the snapshot were fully synchronized, the %VALID  
value would be 100%. The snapshot could then be made independent or moved  
into another disk group.  
Additional information about the snapshots of volumes and volume sets can be  
obtained by using the -noption with the vxsnapprintcommand:  
# vxsnap [-g diskgroup] -n [-l] [-v] [-x] print [vol]  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Administering volume snapshots 343  
Creating instant snapshots  
Alternatively, you can use the vxsnaplistcommand, which is an alias for the  
vxsnap-nprintcommand:  
# vxsnap [-g diskgroup] [-l] [-v] [-x] list [vol]  
The following output is an example of using this command on the disk group  
dg1:  
# vxsnap -g dg -vx list  
NAME  
vol  
DG  
OBJTYPESNAPTYPE PARENT PARENTDG SNAPDATE  
CHANGE_DATA SYNCED_DATA  
10G (100%)  
dg1 vol  
-
-
-
-
-
svol1 dg2 vol  
svol2 dg1 vol  
svol3 dg2 vol  
svol21 dg1 vol  
vol-02 dg1 plex  
fullinst vol  
dg1  
dg1  
dg1  
2006/2/1 12:29 20M (0.2%)  
2006/2/1 12:29 120M (1.2%) 10G (100%)  
2006/2/1 12:29 105M (1.1%) 10G (100%)  
60M (0.6%)  
mirbrk  
volbrk  
vol  
vol  
spaceopt svol2 dg1  
snapmir vol  
2006/2/1 12:29 52M (0.5%)  
52M (0.5%)  
56M (0.6%)  
58M (0.6%)  
2G (100%)  
1G (100%)  
1G (100%)  
2G (100%)  
1G (100%)  
1G (100%)  
-
dg1  
dg1  
-
-
-
dg1  
dg1  
dg1  
dg1  
dg1  
-
-
-
-
-
-
-
-
-
-
mvol  
dg2 vol  
mirvol  
-
vol  
-
vset1 dg1 vset  
v1  
v2  
dg1 compvol-  
dg1 compvol-  
-
-
svset1 dg1 vset  
mirbrk  
vset  
v1  
2006/2/1 12:29 1G (50%)  
2006/2/1 12:29 512M (50%)  
2006/2/1 12:29 512M (50%)  
sv1  
sv2  
dg1 compvolmirbrk  
dg1 compvolmirbrk  
v2  
vol-03 dg1 plex  
mvol2 dg2 vol  
detmir  
detvol  
vol  
vol  
-
-
20M (0.2%)  
20M (0.2%)  
-
This shows that the volume volhas three full-sized snapshots, svol1, svol2  
and svol3, which are of types full-sized instant (fullinst), mirror break-off  
(mirbrk) and linked break-off (volbrk). It also has one snapshot plex  
(snapmir), vol-02, and one linked mirror volume (mirvol), mvol. The  
snapshot svol2itself has a space-optimized instant snapshot (spaceopt),  
svol21. There is also a volume set, vset1, with component volumes v1and v2.  
This volume set has a mirror break-off snapshot, svset1, with component  
volumes sv1and sv2. The last two entries show a detached plex, vol-03, and a  
detached mirror volume, mvol2, which have volas their parent volume. These  
snapshot objects may have become detached due to an I/O error, or, in the case  
of the plex, by running the vxplexdetcommand.  
The CHANGE_DATAcolumn shows the approximate difference between the  
current contents of the snapshot and its parent volume. This corresponds to the  
amount of data that would have to be resynchronized to make the contents the  
same again.  
The SYNCED_DATAcolumn shows the approximate progress of synchronization  
since the snapshot was taken.  
The -loption can be used to obtain a longer form of the output listing instead of  
the tabular form.  
The -xoption expands the output to include the component volumes of volume  
sets.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
344 Administering volume snapshots  
Creating instant snapshots  
See the vxsnap(1M) manual page for more information about using the vxsnap  
printand vxsnaplistcommands.  
Controlling instant snapshot synchronization  
Note: Synchronization of the contents of a snapshot with its original volume is  
not possible for space-optimized instant snapshots.  
The commands in this section cannot be used to control the synchronization of  
linked break-off snapshots.  
By default, synchronization is enabled for the vxsnapreattach, refreshand  
restoreoperations on instant snapshots. Otherwise, synchronization is  
disabled unless you specify the syncing=yesattribute to the vxsnapcommand.  
The following table shows the commands that are provided for controlling the  
synchronization manually.  
Command  
Description  
vxsnap [-g diskgroup] syncpause vol|vol_set Pause synchronization of a volume.  
vxsnap [-g diskgroup] syncresume \  
vol|vol_set  
Resume synchronization of a  
volume.  
vxsnap [-b] [-g diskgroup] syncstart \  
vol|vol_set  
Start synchronization of a volume.  
The -boption puts the operation  
in the background.  
vxsnap [-g diskgroup] syncstop vol|vol_set  
vxsnap [-g diskgroup] syncwait vol|vol_set  
Stop synchronization of a volume.  
Exit when synchronization of a  
volume is complete. An error is  
returned if volis invalid (for  
example, it is a space-optimized  
snapshot), or if volis not being  
synchronized.  
Note: You cannot use this  
command to wait for  
synchronization of reattached  
plexes to complete.  
The vxsnapsnapwaitcommand is provided to wait for the link between new  
linked break-off snapshots to become ACTIVE, or for reattached snapshot plexes  
to reach the SNAPDONE state following resynchronization. See “Creating and  
managing linked break-off snapshot volumes” on page 331, “Reattaching an  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                 
Administering volume snapshots 345  
Creating instant snapshots  
instant snapshot” on page 338 and “Reattaching a linked break-off snapshot  
volume” on page 339 for details.  
Improving the performance of snapshot synchronization  
Two optional arguments to the -ooption are provided to help optimize the  
performance of synchronization when using the make, refresh, restoreand  
syncstartoperations:  
iosize=size Specifies the size of each I/O request that is used when  
synchronizing the regions of a volume. Specifying a larger size  
causes synchronization to complete sooner, but with greater  
impact on the performance of other processes that are  
accessing the volume. The default size of 1m (1MB) is  
suggested as the minimum value for high-performance array  
and controller hardware. The specified value is rounded to a  
multiple of the volume’s region size.  
slow=iodelay Specifies the delay in milliseconds between synchronizing  
successive sets of regions as specified by the value of iosize.  
This can be used to change the impact of synchronization on  
system performance. The default value of iodelay is 0  
milliseconds (no delay). Increasing this value slows down  
synchronization, and reduces the competition for I/O  
bandwidth with other processes that may be accessing the  
volume.  
Options may be combined as shown in the following examples:  
# vxsnap -g mydg -o iosize=2m,slow=100 make \  
source=myvol/snapvol=snap2myvol/syncing=on  
# vxsnap -g mydg -o iosize=10m,slow=250 syncstart snap2myvol  
Note: These optional parameters only affect the synchronization of full-sized  
instant snapshots. They are not supported for space-optimized snapshots.  
Listing the snapshots created on a cache  
To list the space-optimized instant snapshots that have been created on a cache  
object, use the following command:  
# vxcache [-g diskgroup] listvol cache_object  
The snapshot names are printed as a space-separated list ordered by timestamp.  
If two or more snapshots have the same timestamp, these snapshots are sorted  
in order of decreasing size.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
346 Administering volume snapshots  
Creating instant snapshots  
Tuning the autogrow attributes of a cache  
The highwatermark, autogrowbyand maxautogrowattributes determine  
how the VxVM cache daemon (vxcached) maintains the cache if the autogrow  
feature has been enabled and vxcachedis running:  
■
When cache usage reaches the high watermark value, highwatermark  
(default value is 90 percent), vxcachedgrows the size of the cache volume by  
the value of autogrowby(default value is 20% of the size of the cache  
volume in blocks). The new required cache size cannot exceed the value of  
maxautogrow(default value is twice the size of the cache volume in blocks).  
■
When cache usage reaches the high watermark value, and the new required  
cache size would exceed the value of maxautogrow, vxcacheddeletes the  
oldest snapshot in the cache. If there are several snapshots with the same  
age, the largest of these is deleted.  
If the autogrowfeature has been disabled:  
■
When cache usage reaches the high watermark value, vxcacheddeletes the  
oldest snapshot in the cache. If there are several snapshots with the same  
age, the largest of these is deleted. If there is only a single snapshot, this  
snapshot is detached and marked as invalid.  
Note: The vxcacheddaemon does not remove snapshots that are currently  
open, and it does not remove the last or only snapshot in the cache.  
If the cache space becomes exhausted, the snapshot is detached and marked as  
invalid. If this happens, the snapshot is unrecoverable and must be removed.  
Enabling the autogrowfeature on the cache helps to avoid this situation  
occurring. However, for very small caches (of the order of a few megabytes), it is  
possible for the cache to become exhausted before the system has time to  
respond and grow the cache. In such cases, either increase the size of the cache  
manually as described in “Growing and shrinking a cache” on page 347, or use  
the vxcachesetcommand to reduce the value of highwatermarkas shown in  
this example:  
# vxcache -g mydg set highwatermark=60 cobjmydg  
You can use the maxautogrowattribute to limit the maximum size to which a  
cache can grow. To estimate this size, consider how much the contents of each  
source volume are likely to change between snapshot refreshes, and allow some  
additional space for contingency.  
If necessary, you can use the vxcachesetcommand to change other autogrow  
attribute values for a cache. See the vxcache(1M) manual page for details.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Administering volume snapshots 347  
Creating instant snapshots  
Caution: Ensure that the cache is sufficiently large, and that the autogrow  
attributes are configured correctly for your needs.  
Growing and shrinking a cache  
You can use the vxcachecommand to increase the size of the cache volume that  
is associated with a cache object:  
# vxcache [-g diskgroup] growcacheto cache_object size  
For example, to increase the size of the cache volume associated with the cache  
object, mycache, to 2GB, you would use the following command:  
# vxcache -g mydg growcacheto mycache 2g  
To grow a cache by a specified amount, use the following form of the command  
shown here:  
# vxcache [-g diskgroup] growcacheby cache_object size  
For example, the following command increases the size of mycacheby 1GB:  
# vxcache -g mydg growcacheby mycache 1g  
You can similarly use the shrinkcachebyand shrinkcachetooperations to  
reduce the size of a cache. See the vxcache(1M) manual page for more  
information.  
Removing a cache  
To remove a cache completely, including the cache object, its cache volume  
and all space-optimized snapshots that use the cache:  
1
Run the following command to find out the names of the top-level snapshot  
volumes that are configured on the cache object:  
# vxprint -g diskgroup -vne \  
"v_plex.pl_subdisk.sd_dm_name ~ /cache_object/"  
where cache_object is the name of the cache object.  
2
Remove all the top-level snapshots and their dependent snapshots (this can  
be done with a single command):  
# vxedit -g diskgroup -r rm snapvol ...  
where snapvol is the name of a top-level snapshot volume.  
3
4
Stop the cache object:  
# vxcache -g diskgroup stop cache_object  
Finally, remove the cache object and its cache volume:  
# vxedit -g diskgroup -r rm cache_object  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                   
348 Administering volume snapshots  
Creating traditional third-mirror break-off snapshots  
Creating traditional third-mirror break-off  
snapshots  
VxVM provides third-mirror break-off snapshot images of volume devices using  
vxassistand other commands.  
Note: To enhance the efficiency and usability of volume snapshots, turn on  
FastResync as described in “Enabling FastResync on a volume” on page 292. If  
Persistent FastResync is required, you must associate a version 0 DCO with the  
volume as described in “Adding a version 0 DCO and DCO volume” on page 356.  
You need a full license to use this feature.  
The procedure described in this section requires a plex that is large enough to  
store the complete contents of the volume. For details of a method that uses  
space-optimized snapshots, see “Creating instant snapshots” on page 319.  
The recommended approach to performing volume backup from the command  
line, or from a script, is to use the vxassistcommand. The vxassist  
snapstart, snapwait, and snapshottasks allow you to back up volumes online  
with minimal disruption to users.  
The vxassistsnapshotprocedure consists of two steps:  
■
■
Run vxassistsnapstartto create a snapshot mirror.  
Run vxassistsnapshotto create a snapshot volume.  
The vxassistsnapstartstep creates a write-only backup plex which gets  
attached to and synchronized with the volume. When synchronized with the  
volume, the backup plex is ready to be used as a snapshotmirror. The end of  
the update procedure is indicated by the new snapshotmirror changing its  
state to SNAPDONE. This change can be tracked by the vxassist snapwait  
task, which waits until at least one of the mirrors changes its state to  
SNAPDONE. If the attach process fails, the snapshotmirror is removed and its  
space is released.  
Note: If the snapstartprocedure is interrupted, the snapshot mirror is  
automatically removed when the volume is started.  
Once the snapshotmirror is synchronized, it continues being updated until it  
is detached. You can then select a convenient time at which to create a  
snapshotvolume as an image of the existing volume. You can also ask users to  
refrain from using the system during the brief time required to perform the  
snapshot(typically less than a minute). The amount of time involved in  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Administering volume snapshots 349  
Creating traditional third-mirror break-off snapshots  
creating the snapshotmirror is long in contrast to the brief amount of time  
that it takes to create the snapshotvolume.  
The online backup procedure is completed by running the vxassistsnapshot  
command on a volume with a SNAPDONE mirror. This task detaches the  
finished snapshot(which becomes a normal mirror), creates a new normal  
volume and attaches the snapshotmirror to the snapshotvolume. The  
snapshotthen becomes a normal, functioning volume and the state of the  
snapshotis set to ACTIVE.  
To back up a volume using the vxassist command  
1
Create a snapshot mirror for a volume using the following command:  
# vxassist [-b] [-g diskgroup] snapstart [nmirror=N] volume  
For example, to create a snapshot mirror of a volume called voldef, use the  
following command:  
# vxassist [-g diskgroup] snapstart voldef  
The vxassistsnapstarttask creates a write-only mirror, which is attached  
to and synchronized from the volume to be backed up.  
Note: By default, VxVM attempts to avoid placing snapshot mirrors on a  
disk that already holds any plexes of a data volume. However, this may be  
impossible if insufficient space is available in the disk group. In this case,  
VxVM uses any available space on other disks in the disk group. If the  
snapshot plexes are placed on disks which are used to hold the plexes of  
other volumes, this may cause problems when you subsequently attempt to  
move a snapshot volume into another disk group as described in “Moving  
DCO volumes between disk groups” on page 200. To override the default  
storage allocation policy, you can use storage attributes to specify explicitly  
which disks to use for the snapshot plexes. See “Creating a volume on  
specific disks” on page 244 for more information.  
If you start vxassistsnapstartin the background using the -boption, you  
can use the vxassistsnapwaitcommand to wait for the creation of the  
mirror to complete as shown here:  
# vxassist [-g diskgroup] snapwait volume  
If vxassistsnapstartis not run in the background, it does not exit until  
the mirror has been synchronized with the volume. The mirror is then  
ready to be used as a plex of a snapshot volume. While attached to the  
original volume, its contents continue to be updated until you take the  
snapshot.  
Use the nmirrorattribute to create as many snapshot mirrors as you need  
for the snapshot volume. For a backup, you should usually only require the  
default of one.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
350 Administering volume snapshots  
Creating traditional third-mirror break-off snapshots  
It is also possible to make a snapshot plex from an existing plex in a volume.  
See “Converting a plex into a snapshot plex” on page 351 for details.  
2
3
Choose a suitable time to create a snapshot. If possible, plan to take the  
snapshot at a time when users are accessing the volume as little as possible.  
Create a snapshot volume using the following command:  
# vxassist [-g diskgroup] snapshot [nmirror=N] volumesnapshot  
If required, use the nmirrorattribute to specify the number of mirrors in  
the snapshot volume.  
For example, to create a snapshot of voldef, use the following command:  
# vxassist [-g diskgroup] snapshot voldef snapvol  
The vxassistsnapshottask detaches the finished snapshot mirror, creates  
a new volume, and attaches the snapshot mirror to it. This step should only  
take a few minutes. The snapshot volume, which reflects the original  
volume at the time of the snapshot, is now available for backing up, while  
the original volume continues to be available for applications and users.  
If required, you can make snapshot volumes for several volumes in a disk  
group at the same time. See “Creating multiple snapshots” on page 352 for  
more information.  
4
5
Use fsck(or some utility appropriate for the application running on the  
volume) to clean the temporary volume’s contents. For example, you can use  
this command with a VxFS file system:  
# fsck -F vxfs /dev/vx/rdsk/diskgroup/snapshot  
If you require a backup of the data in the snapshot, use an appropriate  
utility or operating system command to copy the contents of the snapshot to  
tape, or to some other backup medium.  
When the backup is complete, you have the following choices for what to do with  
the snapshot volume:  
■
Reattach some or all of the plexes of the snapshot volume with the original  
volume as described in “Reattaching a snapshot volume” on page 352. If  
FastResync was enabled on the volume before the snapshot was taken, this  
speeds resynchronization of the snapshot plexes before the backup cycle  
starts again at step 3.  
■
■
Dissociate the snapshot volume entirely from the original volume as  
described in “Dissociating a snapshot volume” on page 354. This may be  
useful if you want to use the copy for other purposes such as testing or  
report generation.  
Remove the snapshot volume to save space with this command:  
# vxedit [-g diskgroup] -rf rm snapshot  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Administering volume snapshots 351  
Creating traditional third-mirror break-off snapshots  
Note: Dissociating or removing the snapshot volume loses the advantage of  
fast resynchronization if FastResync was enabled. If there are no further  
snapshot plexes available, any subsequent snapshots that you take require  
another complete copy of the original volume to be made.  
Converting a plex into a snapshot plex  
Note: The procedure described in this section cannot be used with layered  
volumes or any volume that has an associated version 20 DCO volume.  
It is recommended that the instant snapshot feature is used in preference to the  
procedure described in this section.  
In some circumstances, you may find it more convenient to convert an existing  
plex in a volume into a snapshot plex rather than running vxassistsnapstart.  
For example, you may want to do this if you are short of disk space for creating  
the snapshot plex and the volume that you want to snapshot contains more than  
two plexes.  
The procedure can also be used to speed up the creation of a snapshot volume  
when a mirrored volume is created with more than two plexes and init=active  
is specified.  
Note: It is advisable to retain at least two plexes in a volume to maintain data  
redundancy.  
To convert an existing plex into a snapshot plex for a volume on which  
Persistent FastResync is enabled, use the following command:  
# vxplex [-g diskgroup] -o dcoplex=dcologplex convert \  
state=SNAPDONE plex  
dcologplex is the name of an existing DCO plex that is to be associated with the  
new snapshot plex. You can use the vxprintcommand to find out the name of  
the DCO volume as described in “Adding a version 0 DCO and DCO volume” on  
page 356.  
For example, to make a snapshot plex from the plex trivol-03in the 3-plex  
volume trivol, you would use the following command:  
# vxplex -o dcoplex=trivol_dco-03 convert state=SNAPDONE \  
trivol-03  
Here the DCO plex trivol_dco_03is specified as the DCO plex for the new  
snapshot plex.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
352 Administering volume snapshots  
Creating traditional third-mirror break-off snapshots  
To convert an existing plex into a snapshot plex in the SNAPDONE state for a  
volume on which Non-Persistent FastResync is enabled, use the following  
command:  
# vxplex [-g diskgroup] convert state=SNAPDONE plex  
A converted plex is in the SNAPDONE state, and can be used immediately to  
create a snapshot volume.  
Note: The last complete regular plex in a volume, an incomplete regular plex, or  
a dirty region logging (DRL) log plex cannot be converted into a snapshot plex.  
Creating multiple snapshots  
To make it easier to create snapshots of several volumes at the same time, the  
snapshot option accepts more than one volume name as its argument, for  
example:  
# vxassist [-g diskgroup] snapshot volume1volume2...  
By default, the first snapshot volume is named SNAP-volume, and each  
subsequent snapshot is named SNAPnumber-volume, where number is a unique  
serial number, and volume is the name of the volume for which the snapshot is  
being taken. This default pattern can be overridden by using the option -o  
name=pattern, as described on the vxassist(1M) manual page. For example, the  
pattern SNAP%v-%dreverses the order of the number and volume components in  
the name.  
To snapshot all the volumes in a single disk group, specify the option -o  
allvolsto vxassist:  
# vxassist -g diskgroup -o allvols snapshot  
This operation requires that all snapstartoperations are complete on the  
volumes. It fails if any of the volumes in the disk group do not have a complete  
snapshot plex in the SNAPDONE state.  
Reattaching a snapshot volume  
Note: The information in this section does not apply to RAID-5 volumes unless  
they have been converted to a special layered volume layout by the addition of a  
DCO and DCO volume. See “Adding a version 0 DCO and DCO volume” on  
page 356 for details.  
Snapback merges a snapshot copy of a volume with the original volume. One or  
more snapshot plexes are detached from the snapshot volume and re-attached  
to the original volume. The snapshot volume is removed if all its snapshot  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Administering volume snapshots 353  
Creating traditional third-mirror break-off snapshots  
plexes are snapped back. This task resynchronizes the data in the volume so that  
the plexes are consistent.  
Note: To enhance the efficiency of the snapback operation, enable FastResync  
on the volume before taking the snapshot, as described in “Enabling FastResync  
on a volume” on page 292.  
To merge one snapshot plex with the original volume, use the following  
command:  
# vxassist [-g diskgroup] snapback snapshot  
where snapshot is the snapshot copy of the volume.  
To merge all snapshot plexes in the snapshot volume with the original volume,  
use the following command:  
# vxassist [-g diskgroup] -o allplexes snapback snapshot  
To merge a specified number of plexes from the snapshot volume with the  
original volume, use the following command:  
# vxassist [-g diskgroup] snapback nmirror=number snapshot  
Here the nmirrorattribute specifies the number of mirrors in the snapshot  
volume that are to be re-attached.  
Once the snapshot plexes have been reattached and their data resynchronized,  
they are ready to be used in another snapshotoperation.  
By default, the data in the original volume is used to update the snapshot plexes  
that have been re-attached. To copy the data from the replica volume instead,  
use the following command:  
# vxassist [-g diskgroup] -o resyncfromreplica snapback snapshot  
Caution: Always unmount the snapshot volume (if mounted) before performing  
a snapback. In addition, you must unmount the file system corresponding to the  
primary volume before using the resyncfromreplicaoption.  
Adding plexes to a snapshot volume  
If you want to retain the existing plexes in a snapshot volume after a snapback  
operation, you can create additional snapshot plexes that are to be used for the  
snapback.  
To add plexes to a snapshot volume  
1
Use the following vxprintcommands to discover the names of the snapshot  
volume’s data change object (DCO) and DCO volume:  
# DCONAME=‘vxprint [-g diskgroup] -F%dco_name snapshot‘  
# DCOVOL=‘vxprint [-g diskgroup] -F%log_vol $DCONAME‘  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
354 Administering volume snapshots  
Creating traditional third-mirror break-off snapshots  
2
Use the vxassistmirrorcommand to create mirrors of the existing  
snapshot volume and its DCO volume:  
# vxassist -g diskgroup mirror snapshot  
# vxassist -g diskgroup mirror $DCOVOL  
Note: The new plex in the DCO volume is required for use with the new data  
plex in the snapshot.  
3
4
5
Use the vxprintcommand to find out the name of the additional snapshot  
plex:  
# vxprint -g diskgroup snapshot  
Use the vxprintcommand to find out the record ID of the additional DCO  
plex:  
# vxprint -g diskgroup -F%rid $DCOVOL  
Use the vxeditcommand to set the dco_plex_ridfield of the new data  
plex to the name of the new DCO plex:  
# vxedit -g diskgroup set dco_plex_rid=dco_plex_rid new_plex  
The new data plex is now ready to be used to perform a snapback operation.  
Dissociating a snapshot volume  
The link between a snapshot and its original volume can be permanently broken  
so that the snapshot volume becomes an independent volume. Use the following  
command to dissociate the snapshot volume, snapshot:  
# vxassist snapclear snapshot  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Administering volume snapshots 355  
Creating traditional third-mirror break-off snapshots  
Displaying snapshot information  
The vxassistsnapprintcommand displays the associations between the  
original volumes and their respective replicas (snapshot copies):  
# vxassist snapprint [volume]  
Output from this command is shown in the following examples:  
# vxassist -g mydg snapprint v1  
V NAME  
SS SNAPOBJ  
DP NAME  
USETYPE  
NAME  
VOLUME  
LENGTH  
LENGTH  
LENGTH  
%DIRTY  
%DIRTY  
v v1  
fsgen  
20480  
20480  
20480  
20480  
ss SNAP-v1_snp SNAP-v1  
dp v1-01  
dp v1-02  
4
0
0
v1  
v1  
v SNAP-v1  
ss v1_snp  
fsgen  
v1  
20480  
20480  
0
# vxassist -g mydg snapprint v2  
V NAME  
SS SNAPOBJ  
DP NAME  
USETYPE  
NAME  
VOLUME  
LENGTH  
LENGTH  
LENGTH  
%DIRTY  
%DIRTY  
v v2  
ss --  
dp v2-01  
fsgen  
SNAP-v2  
v2  
20480  
20480  
20480  
0
0
v SNAP-v2  
ss --  
fsgen  
v2  
20480  
20480  
0
In this example, Persistent FastResync is enabled on volume v1, and Non-  
Persistent FastResync on volume v2. Lines beginning with v, dpand ssindicate  
a volume, detached plex and snapshot plex respectively. The %DIRTYfield  
indicates the percentage of a snapshot plex or detached plex that is dirty with  
respect to the original volume. Notice that no snap objects are associated with  
volume v2or with its snapshot volume SNAP-v2. See “How persistent  
FastResync works with snapshots” on page 70 for more information about snap  
objects.  
If a volume is specified, the snapprintcommand displays an error message if no  
FastResync maps are enabled for that volume.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
356 Administering volume snapshots  
Adding a version 0 DCO and DCO volume  
Adding a version 0 DCO and DCO volume  
Note: The procedure described in this section adds a DCO log volume that has a  
version 0 layout as introduced in VxVM 3.2. The version 0 layout supports  
traditional (third-mirror break-off) snapshots, but not full-sized or space-  
optimized instant snapshots. See “Version 0 DCO volume layout” on page 69 and  
“Version 20 DCO volume layout” on page 69 for a description of the differences  
between the old and new DCO volume layouts.  
See “Determining the DCO version number” on page 277 for details of how to  
determine the version number of a volume’s DCO.  
To put Persistent FastResync into effect for a volume, a Data Change Object  
(DCO) and DCO volume must first be associated with that volume. When you  
have added a DCO object and DCO volume to a volume, you can then enable  
Persistent FastResync on the volume as described in “Enabling FastResync on a  
volume” on page 292.  
TM  
Note: You need a Veritas FlashSnap or FastResync license to use the  
FastResync feature. Even if you do not have a license, you can configure a DCO  
object and DCO volume so that snap objects are associated with the original and  
snapshot volumes. For more information about snap objects, see “How  
persistent FastResync works with snapshots” on page 70.  
To add a DCO object and DCO volume to an existing volume (which may already  
have dirty region logging (DRL) enabled), use the following procedure:  
1
Ensure that the disk group containing the existing volume has been  
upgraded to at least version 90. Use the following command to check the  
version of a disk group:  
# vxdg list diskgroup  
To upgrade a disk group to the latest version, use the following command:  
# vxdg upgrade diskgroup  
For more information, see “Upgrading a disk group” on page 208.  
2
Use the following command to turn off Non-Persistent FastResync on the  
original volume if it is currently enabled:  
# vxvol [-g diskgroup] set fastresync=off volume  
If you are uncertain about which volumes have Non-Persistent FastResync  
enabled, use the following command to obtain a listing of such volumes:  
# vxprint [-g diskgroup] -F “%name” \  
-e “v_fastresync=on && !v_hasdcolog”  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Administering volume snapshots 357  
Adding a version 0 DCO and DCO volume  
3
Use the following command to add a DCO and DCO volume to the existing  
volume:  
# vxassist [-g diskgroup] addlog volume logtype=dco \  
[ndcomirror=number] [dcolen=size] [storage_attributes]  
For non-layered volumes, the default number of plexes in the mirrored DCO  
volume is equal to the lesser of the number of plexes in the data volume or  
2. For layered volumes, the default number of DCO plexes is always 2. If  
required, use the ndcomirrorattribute to specify a different number. It is  
recommended that you configure as many DCO plexes as there are existing  
data and snapshot plexes in the volume. For example, specify ndcomirror=3  
when adding a DCO to a 3-way mirrored volume.  
The default size of each plex is 132 blocks. You can use the dcolen  
attribute to specify a different size. If specified, the size of the plex must be  
an integer multiple of 33 blocks from 33 up to a maximum of 2112 blocks.  
You can specify vxassist-style storage attributes to define the disks that can  
and/or cannot be used for the plexes of the DCO volume. See “Specifying storage  
for version 0 DCO plexes” on page 357 for details.  
Specifying storage for version 0 DCO plexes  
Note: The operations in this section relate to version 0 DCO volumes. They are  
not supported for the version 20 DCO volume layout that was introduced in  
VxVM 4.0.  
If the disks that contain volumes and their snapshots are to be moved or split  
into different disk groups, the disks that contain their respective DCO plexes  
must be able to accompany them. By default, VxVM attempts to place the DCO  
plexes on the same disks as the data plexes of the parent volume. However, this  
may be impossible if there is insufficient space available on those disks. In this  
case, VxVM uses any available space on other disks in the disk group. If the DCO  
plexes are placed on disks which are used to hold the plexes of other volumes,  
this may cause problems when you subsequently attempt to move volumes into  
other disk groups.  
You can use storage attributes to specify explicitly which disks to use for the  
DCO plexes. If possible, specify the same disks as those on which the volume is  
configured. For example, to add a DCO object and DCO volume with plexes on  
mydg05and mydg06, and a plex size of 264 blocks to the volume, myvol, in the  
disk group, mydg, use the following command:  
# vxassist -g mydg addlog myvol logtype=dco dcolen=264 \  
mydg05 mydg06  
To view the details of the DCO object and DCO volume that are associated with a  
volume, use the vxprintcommand. The following is partial vxprintoutput for  
Download from Www.Somanuals.com. All Manuals Search And Download.  
               
358 Administering volume snapshots  
Adding a version 0 DCO and DCO volume  
the volume named vol1(the TUTIL0 and PUTIL0 columns are omitted for  
clarity):  
TY NAME  
v vol1  
ASSOC  
fsgen  
vol1  
vol1-01  
vol1  
vol1-02  
vol1  
gen  
KSTATE  
LENGTH PLOFFS  
STATE ...  
ACTIVE  
ACTIVE  
-
ACTIVE  
-
-
ACTIVE  
ACTIVE  
-
ACTIVE  
-
ENABLED 1024  
ENABLED 1024  
ENABLED 1024  
ENABLED 1024  
ENABLED 1024  
-
-
0
-
0
-
-
-
0
-
pl vol1-01  
sd disk01-01  
pl vol1-02  
sd disk02-01  
dc vol1_dco  
v vol1_dcl  
pl vol1_dcl-01 vol1_dcl  
sd disk03-01  
pl vol1_dcl-02 vol1_dcl  
sd disk04-01 vol1_dcl-02 ENABLED 132  
-
-
ENABLED 132  
ENABLED 132  
vol1_dcl-01 ENABLED 132  
ENABLED 132  
0
In this output, the DCO object is shown as vol1_dco, and the DCO volume as  
vol1_dclwith 2 plexes, vol1_dcl-01and vol1_dcl-02.  
If required, you can use the vxassistmovecommand to relocate DCO plexes to  
different disks. For example, the following command moves the plexes of the  
DCO volume, vol1_dcl, for volume vol1from disk03and disk04to disk07  
and disk08:  
# vxassist -g mydg move vol1_dcl !disk03 !disk04 disk07 disk08  
For more information, see “Moving DCO volumes between disk groups” on  
page 200, and the vxassist(1M) manual page.  
Removing a version 0 DCO and DCO volume  
Note: The operations in this section relate to version 0 DCO volumes. They are  
not supported for the version 20 DCO volume layout that was introduced in  
VxVM 4.0.  
To dissociate a DCO object, DCO volume and any snap objects from a volume, use  
the following command:  
# vxassist [-g diskgroup] remove log volume logtype=dco  
This completely removes the DCO object, DCO volume and any snap objects. It  
also has the effect of disabling FastResync for the volume.  
Alternatively, you can use the vxdcocommand to the same effect:  
# vxdco [-g diskgroup] [-o rm] dis dco_obj  
The default name of the DCO object, dco_obj, for a volume is usually formed by  
appending the string _dcoto the name of the parent volume. To find out the  
name of the associated DCO object, use the vxprintcommand on the volume.  
To dissociate, but not remove, the DCO object, DCO volume and any snap objects  
from the volume, myvol, in the disk group, mydg, use the following command:  
# vxdco -g mydg dis myvol_dco  
Download from Www.Somanuals.com. All Manuals Search And Download.  
               
Administering volume snapshots 359  
Adding a version 0 DCO and DCO volume  
This form of the command dissociates the DCO object from the volume but does  
not destroy it or the DCO volume. If the -ormoption is specified, the DCO object,  
DCO volume and its plexes, and any snap objects are also removed.  
Note: Dissociating a DCO and DCO volume disables Persistent FastResync on the  
volume. A full resynchronization of any remaining snapshots is required when  
they are snapped back.  
For more information, see the vxassist(1M) and vxdco(1M) manual pages.  
Reattaching a version 0 DCO and DCO volume  
Note: The operations in this section relate to version 0 DCO volumes. They are  
not supported for version 20 DCO volume layout that was introduced in VxVM  
4.0.  
If the DCO object and DCO volume are not removed by specifying the -orm  
option to vxdco, they can be reattached to the parent volume using the following  
command:  
# vxdco [-g diskgroup] att volume dco_obj  
For example, to reattach the DCO object, myvol_dco, to the volume, myvol, use  
the following command:  
# vxdco -g mydg att myvol myvol_dco  
For more information, see the vxdco(1M) manual page.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
360 Administering volume snapshots  
Adding a version 0 DCO and DCO volume  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter  
10  
Creating and  
administering  
volume sets  
This chapter describes how to use the vxvsetcommand to create and administer  
volume sets in Veritas Volume Manager (VxVM). Volume sets enable the use of  
the Multi-Volume Support feature with Veritas File System (VxFS). It is also  
possible to use the Veritas Enterprise Administrator (VEA) to create and  
administer volumes sets. For more information, see the VEA online help.  
For full details of the usage of the vxvsetcommand, see the vxvset(1M) manual  
page.  
Note: Most VxVM commands require superuser or equivalent privileges.  
Please note the following limitation of volume sets:  
■
■
■
A maximum of 2048 volumes may be configured in a volume set.  
Only Veritas File System is supported on a volume set.  
The first volume (index 0) in a volume set must be larger than the sum of the  
total volume size divided by 4000, the size of the VxFS intent log, and 1MB.  
■
■
Raw I/O from and to a volume set is not supported.  
Raw I/O from and to the component volumes of a volume set is supported  
under certain conditions. See “Raw device node access to component  
volumes” on page 364 for more information.  
■
Resizing a volume setwith an unmounted file system is not supported.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
362 Creating and administering volume sets  
Creating a volume set  
■
Volume sets can be used in place of volumes with the following vxsnap  
operations on instant snapshots: addmir, dis, make, prepare, reattach,  
refresh, restore, rmmir, split, syncpause, syncresume, syncstart,  
syncstop, syncwait, and unprepare. The third-mirror break-off usage  
model for full-sized instant snapshots is supported for volume sets provided  
that sufficient plexes exist for each volume in the volume set. See  
“Administering volume snapshots” on page 303 and “Creating instant  
snapshots of volume sets” on page 334 for more information.  
■
A full-sized snapshot of a volume set must itself be a volume set with the  
same number of volumes and the same volume index numbers as the parent.  
The corresponding volumes in the parent and snapshot volume sets are also  
subject to the same restrictions as apply between standalone volumes and  
their snapshots.  
Creating a volume set  
To create a volume set for use by Veritas File System (VxFS), use the following  
command:  
# vxvset [-g diskgroup] -t vxfs make volset volume  
Here volset is the name of the volume set, and volume is the name of the first  
volume in the volume set. The -toption defines the content handler  
subdirectory for the application that is to be used with the volume. This  
subdirectory contains utilities that an application uses to operate on the volume  
set. As the operation of these utilities is determined by the requirements of the  
application and not by VxVM, it is not discussed further here.  
For example, to create a volume set named myvsetthat contains the volume  
vol1, in the disk group mydg, you would use the following command:  
# vxvset -g mydg -t vxfs make myvset vol1  
Adding a volume to a volume set  
Having created a volume set containing a single volume, you can use the  
following command to add further volumes to the volume set:  
# vxvset [-g diskgroup] [-f] addvol volset volume  
For example, to add the volume vol2, to the volume set myvset, use the  
following command:  
# vxvset -g mydg addvol myvset vol2  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Creating and administering volume sets 363  
Listing details of volume sets  
Caution: The -f(force) option must be specified if the volume being added, or  
any volume in the volume set, is either a snapshot or the parent of a snapshot.  
Using this option can potentially cause inconsistencies in a snapshot hierarchy  
if any of the volumes involved in the operation is already in a snapshot chain.  
Listing details of volume sets  
To list the details of the component volumes of a volume set, use the following  
command:  
# vxvset [-g diskgroup] list [volset]  
If the name of a volume set is not specified, the command lists the details of all  
volume sets in a disk group, as shown in the following example:  
# vxvset -g mydg list  
NAME  
set1  
set2  
GROUP  
mydg  
mydg  
NVOLS  
CONTEXT  
-
-
3
2
To list the details of each volume in a volume set, specify the name of the  
volume set as an argument to the command:  
# vxvset -g mydg list set1  
VOLUME  
vol1  
vol2  
INDEX  
LENGTH  
12582912  
12582912  
12582912  
KSTATE CONTEXT  
ENABLED -  
ENABLED -  
0
1
2
vol3  
ENABLED -  
The context field contains details of any string that the application has set up  
for the volume or volume set to tag its purpose.  
Stopping and starting volume sets  
Under some circumstances, you may need to stop and restart a volume set. For  
example, a volume within the set may have become detached, as shown here:  
# vxvset -g mydg list set1  
VOLUME  
vol1  
vol2  
INDEX  
LENGTH  
12582912  
12582912  
12582912  
KSTATE CONTEXT  
DETACHED -  
ENABLED -  
0
1
2
vol3  
ENABLED -  
To stop and restart one or more volume sets, use the following commands:  
# vxvset [-g diskgroup] stop volset ...  
# vxvset [-g diskgroup] start volset ...  
For the example given previously, the effect of running these commands on the  
component volumes is shown below:  
# vxvset -g mydg stop set1  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
364 Creating and administering volume sets  
Removing a volume from a volume set  
# vxvset -g mydg list set1  
VOLUME  
vol1  
vol2  
INDEX  
LENGTH  
12582912  
12582912  
12582912  
KSTATE CONTEXT  
DISABLED -  
DISABLED -  
0
1
2
vol3  
DISABLED -  
# vxvset -g mydg start set1  
# vxvset -g mydg list set1  
VOLUME  
vol1  
vol2  
INDEX  
LENGTH  
12582912  
12582912  
12582912  
KSTATE CONTEXT  
ENABLED -  
ENABLED -  
0
1
2
vol3  
ENABLED -  
Removing a volume from a volume set  
To remove a component volume from a volume set, use the following command:  
# vxvset [-g diskgroup] [-f] rmvol volset volume  
For example, the following commands remove the volumes, vol1and vol2,  
from the volume set myvset:  
# vxvset -g mydg rmvol myvset vol1  
# vxvset -g mydg rmvol myvset vol2  
Note: When the final volume is removed, this deletes the volume set.  
Caution: The -f(force) option must be specified if the volume being removed, or  
any volume in the volume set, is either a snapshot or the parent of a snapshot.  
Using this option can potentially cause inconsistencies in a snapshot hierarchy  
if any of the volumes involved in the operation is already in a snapshot chain.  
Raw device node access to component volumes  
To guard against accidental file system and data corruption, the device nodes of  
the component volumes are configured by default not to have raw and block  
entries in the /dev/vx/rdsk/diskgroupand /dev/vx/dsk/diskgroup  
directories. As a result, applications are prevented from directly reading from or  
writing to the component volumes of a volume set.  
If some applications, such as the raw volume backup and restore feature of the  
TM  
Veritas NetBackup software, need to read from or write to the component  
volumes by accessing raw device nodes in the /dev/vx/rdsk/diskgroup  
directory, this is supported by specifying additional command-line options to  
the vxvsetcommand. Access to the block device nodes of the component  
volumes of a volume set is unsupported.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Creating and administering volume sets 365  
Raw device node access to component volumes  
Caution: Writing directly to or reading from the raw device node of a component  
volume of a volume set should only be performed if it is known that the volume's  
data will not otherwise change during the period of access.  
All of the raw device nodes for the component volumes of a volume set can be  
created or removed in a single operation. Raw device nodes for any volumes  
added to a volume set are created automatically as required, and inherit the  
access mode of the existing device nodes.  
Access to the raw device nodes for the component volumes can be configured to  
be read-only or read-write. This mode is shared by all the raw device nodes for  
the component volumes of a volume set. The read-only access mode implies that  
any writes to the raw device will fail, however writes using the ioctlinterface  
or by VxFS to update metadata are not prevented. The read-write access mode  
allows direct writes via the raw device. The access mode to the raw device nodes  
of a volume set can be changed as required.  
The presence of raw device nodes and their access mode is persistent across  
system reboots.  
Note the following limitations of this feature:  
■
■
The disk group version must be 120 or greater.  
Access to the raw device nodes of the component volumes of a volume set is  
only supported for private disk groups; it is not supported for shared disk  
groups in a cluster.  
Enabling raw device access when creating a volume set  
To enable raw device access when creating a volume set, use the following form  
of the vxvsetmakecommand:  
# vxvset [-g diskgroup] -o makedev=on \  
[-o compvol_access={read-only|read-write}] \  
[-o index] [-c "ch_addopt"] make vset vol [index]  
The -o makedev=onoption enables the creation of raw device nodes for the  
component volumes at the same time that the volume set is created. The default  
is setting is off.  
If the -o compvol_access=read-writeoption is specified, direct writes are  
allowed to the raw device of each component volume. If the value is set to read-  
only, only reads are allowed from the raw device of each component volume.  
If the -o makedev=onoption is specified, but -o compvol_accessis not  
specified, the default access mode is read-only.  
If the vxvsetaddvolcommand is subsequently used to add a volume to a volume  
set, a new raw device node is created in /dev/vx/rdsk/diskgroup if the  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
366 Creating and administering volume sets  
Raw device node access to component volumes  
value of the makedevattribute is currently set to on. The access mode is  
determined by the current setting of the compvol_accessattribute.  
The following example creates a volume set, myvset1, containing the volume,  
myvol1, in the disk group, mydg, with raw device access enabled in read-write  
mode:  
# vxvset -g mydg -o makedev=on -o compvol_access=read-write \  
make myvset1 myvol1  
Displaying the raw device access settings for a volume set  
You can use the vxprint-m command to display the current settings for a  
volume set. If the makedevattribute is set to on, one of the following is displayed  
in the output:  
vset_devinfo=on:read-only Raw device nodes in read-only mode.  
vset_devinfo=on:read-write Raw device nodes in read-write mode.  
This field is not displayed if makedevis set to off.  
Note: If the output from the vxprint-m command is fed to the vxmakecommand  
to recreate a volume set, the vset_devinfoattribute must set to off. Use the  
vxvset set command to re-enable raw device access with the desired access mode  
as described in “Controlling raw device access for an existing volume set” on  
page 366.  
Controlling raw device access for an existing volume set  
To enable or disable raw device node access for an existing volume set, use the  
following command:  
# vxvset [-g diskgroup] [-f] set makedev={on|off} vset  
The makedevattribute can be specified to the vxvsetsetcommand to create  
(makedev=on) or remove (makedev=off) the raw device nodes for the component  
volumes of a volume set. If any of the component volumes are open, the -f  
(force) option must be specified to set the attribute to off.  
Note: Specifying makedev=offremoves the existing raw device nodes from the  
/dev/vx/rdsk/diskgroup directory.  
If the makedevattribute is set to off, and you use the mknodcommand to create  
the raw device nodes, you cannot read from or write to those nodes unless you  
set the value of makedevto on.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Creating and administering volume sets 367  
Raw device node access to component volumes  
The syntax for setting the compvol_accessattribute on a volume set is:  
# vxvset [-g diskgroup] [-f] set \  
compvol_access={read-only|read-write} vset  
The compvol_accessattribute can be specified to the vxvsetsetcommand to  
change the access mode to the component volumes of a volume set. If any of the  
component volumes are open, the -f(force) option must be specified to set the  
attribute to read-only.  
The following example sets the makedev=onand compvol_access=read-only  
attributes on a volume set, myvset2, in the disk group, mydg:  
# vxvset -g mydg set makedev=on myvset2  
The next example sets the compvol_access=read-writeattribute on the volume  
set, myvset2:  
# vxvset -g mydg set compvol_access=read-write myvset2  
The final example removes raw device node access for the volume set, myvset2:  
# vxvset -g mydg set makedev=off myvset2  
Download from Www.Somanuals.com. All Manuals Search And Download.  
368 Creating and administering volume sets  
Raw device node access to component volumes  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter  
11  
Configuring off-host  
processing  
Off-host processing allows you to implement the following activities:  
Data backup  
As the requirement for 24 x 7 availability becomes essential  
for many businesses, organizations cannot afford the  
downtime involved in backing up critical data offline. By  
taking a snapshot of the data, and backing up from this  
snapshot, business-critical applications can continue to run  
without extended down time or impacted performance.  
Decision support analysis  
and reporting  
Because snapshots hold a point-in-time copy of a  
production database, a replica of the database can be set up  
using the snapshots. Operations such as decision support  
analysis and business reporting do not require access to up-  
to-the-minute information. This means that they can use a  
database copy that is running on a host other than the  
primary. When required, the database copy can quickly be  
synchronized with the data in the primary database.  
Testing and training  
Development or service groups can use snapshots as test  
data for new applications. Snapshot data provides  
developers, system testers and QA groups with a realistic  
basis for testing the robustness, integrity and performance  
of new applications.  
Database error recovery  
Logic errors caused by an administrator or an application  
program can compromise the integrity of a database. By  
restoring the database table files from a snapshot copy, the  
database can be recovered more quickly than by full  
restoration from tape or other backup media.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
370 Configuring off-host processing  
Implementing off-host processing solutions  
Off-host processing is made simpler by using linked break-off snapshots, which  
are described in “Linked break-off snapshot volumes” on page 311.  
Implementing off-host processing solutions  
As shown in Figure 11-1, by accessing snapshot volumes from a lightly loaded  
host (shown here as the off-host processing (OHP) host), CPU- and I/O-intensive  
operations for online backup and decision support do not degrade the  
performance of the primary host that is performing the main production  
activity (such as running a database). Also, if you place the snapshot volumes on  
disks that are attached to different host controllers than the disks in the  
primary volumes, it is possible to avoid contending with the primary host for I/O  
resources.  
Figure 11-1  
Example implementation of off-host processing  
OHP host  
Primary host  
SCSI or Fibre Channel  
connectivity  
Disks containing primary  
volumes used to hold  
production databases or file  
systems  
Disks containing snapshot  
volumes used to implement  
off-host processing solutions  
The following sections describe how you can apply off-host processing to  
implement regular online backup of a volume in a private disk group, and to set  
up a replica of a production database for decision support. Two applications are  
outlined in the following sections:  
■
Implementing off-host online backup  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Configuring off-host processing 371  
Implementing off-host processing solutions  
■
Implementing decision support  
These applications use the Persistent FastResync feature of VxVM in  
conjunction with linked break-off snapshots.  
Note: A volume snapshot represents the data that exists in a volume at a given  
point in time. As such, VxVM does not have any knowledge of data that is cached  
by the overlying file system, or by applications such as databases that have files  
open in the file system. If the fsgenvolume usage type is set on a volume that  
contains a Veritas File System (VxFS), intent logging of the file system metadata  
ensures the internal consistency of the file system that is backed up. For other  
file system types, depending on the intent logging capabilities of the file system,  
there may potentially be inconsistencies between in-memory data and the data  
in the snapshot image.  
For databases, a suitable mechanism must additionally be used to ensure the  
integrity of tablespace data when the volume snapshot is taken. The facility to  
temporarily suspend file system I/O is provided by most modern database  
software. For ordinary files in a file system, which may be open to a wide variety  
of different applications, there may be no way to ensure the complete integrity  
of the file data other than by shutting down the applications and temporarily  
unmounting the file system. In many cases, it may only be important to ensure  
the integrity of file data that is not in active use at the time that you take the  
snapshot.  
Implementing off-host online backup  
This section describes a procedure for implementing off-host online backup for  
a volume in a private disk group. The intention is to present an outline of how to  
set up a regular backup cycle. It is beyond the scope of this guide to describe how  
to configure a database to use this procedure, or how to perform the backup  
itself.  
To back up a volume in a private disk group  
1
Use the following command on the primary host to see if the volume is  
associated with a version 20 data change object (DCO) and DCO volume that  
allow instant snapshots and Persistent FastResync to be used with the  
volume:  
# vxprint -g volumedg -F%instant volume  
This command returns onif the volume can be used for instant snapshot  
operations; otherwise, it returns off.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
372 Configuring off-host processing  
Implementing off-host processing solutions  
Note: If the volume was created under VxVM 4.0 or a later release, and it is  
not associated with a new-style DCO object and DCO volume, follow the  
procedure described in “Preparing a volume for DRL and instant snapshots”  
on page 275.  
If the volume was created before release 4.0 of VxVM, and has any attached  
snapshot plexes, or is associated with any snapshot volumes, follow the  
procedure given in “Upgrading existing volumes to use version 20 DCOs” on  
page 279.  
2
Use the following command on the primary host to check whether  
FastResync is enabled on the volume:  
# vxprint -g volumedg -F%fastresync volume  
This command returns onif FastResync is enabled; otherwise, it returns  
off.  
If FastResync is disabled, enable it using the following command on the  
primary host:  
# vxvol -g volumedg set fastresync=on volume  
3
4
On the primary host, create a new volume in a separate disk group for use as  
the snapshot volume as described in “Creating a volume for use as a full-  
sized instant or linked break-off snapshot” on page 323. It is recommended  
that a snapshot disk group is dedicated to maintaining only those disks that  
are used for off-host processing.  
On the primary host, link the snapshot volume in the snapshot disk group to  
the data volume:  
# vxsnap -g volumedg -b addmir volume mirvol=snapvol \  
mirdg=snapvoldg  
You can use the vxsnapsnapwaitcommand to wait for synchronization of  
the linked snapshot volume to complete:  
# vxsnap -g volumedg snapwait volume mirvol=snapvol \  
mirdg=snapvoldg  
Note: This step sets up the snapshot volumes. When you are ready to create  
a backup, proceed to step 5.  
5
6
On the primary host, suspend updates to the volume that contains the  
database tables. A database may have a hot backup mode that allows you to  
do this by temporarily suspending writes to its tables.  
Create the snapshot volume, snapvol, by running the following command on  
the primary host:  
# vxsnap -g volumedg make \  
source=volume/snapvol=snapvol/snapdg=snapvoldg  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Configuring off-host processing 373  
Implementing off-host processing solutions  
If a database spans more than one volume, you can specify all the volumes  
and their snapshot volumes using one command, as shown in this example:  
# vxsnap -g dbasedg make \  
source=vol1/snapvol=snapvol1/snapdg=sdg \  
source=vol2/snapvol=snapvol2/snapdg=sdg \  
source=vol3/snapvol=snapvol3/snapdg=sdg  
This step sets up the snapshot volumes ready for the backup cycle, and  
starts tracking changes to the original volumes.  
7
8
On the primary host, if you temporarily suspended updates to a volume in  
step 5, release all the database tables from hot backup mode.  
On the primary host, deport the snapshot volume’s disk group using the  
following command:  
# vxdg deport snapvoldg  
9
On the OHP host where the backup is to be performed, use the following  
command to import the snapshot volume’s disk group:  
# vxdg import snapvoldg  
10 The snapshot volume is initially disabled following the join. Use the  
following commands on the OHP host to recover and restart the snapshot  
volume:  
# vxrecover -g snapvoldg -m snapvol  
# vxvol -g snapvoldg start snapvol  
11 On the OHP host, back up the snapshot volume. If you need to remount the  
file system in the volume to back it up, first run fsckon the volume. The  
following are sample commands for checking and mounting a file system:  
# fsck -F vxfs /dev/vx/rdsk/snapvoldg/snapvol  
# mount -F vxfs /dev/vx/dsk/snapvoldg/snapvol mount_point  
Back up the file system at this point, and then use the following command  
to unmount it.  
# umount mount_point  
12 On the OHP host, use the following command to deport the snapshot  
volume’s disk group:  
# vxdg deport snapvoldg  
13 On the primary host, re-import the snapshot volume’s disk group using the  
following command:  
# vxdg import snapvoldg  
14 The snapshot volume is initially disabled following the join. Use the  
following command on the primary host to recover and restart the snapshot  
volume:  
# vxrecover -g snapvoldg -m snapvol  
15 On the primary host, reattach the snapshot volume to its original volume  
using the following command:  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
374 Configuring off-host processing  
Implementing off-host processing solutions  
# vxsnap -g snapvoldg reattach snapvol source=vol \  
sourcedg=volumedg  
For example, to reattach the snapshot volumes svol1, svol2and svol3:  
# vxsnap -g sdg reattach svol1 \  
source=vol1 sourcedg=dbasedg \  
svol2 source=vol2 sourcedg=dbasedg \  
svol3 source=vol3 sourcedg=dbasedg  
You can use the vxsnapsnapwaitcommand to wait for synchronization of  
the linked snapshot volume to complete:  
# vxsnap -g volumedg snapwait volume mirvol=snapvol  
Repeat step 5 through step 15 each time that you need to back up the volume.  
Implementing decision support  
This section describes a procedure for implementing off-host decision support  
for a volume in a private disk group. The intention is to present an outline of  
how to set up a replica database. It is beyond the scope of this guide to describe  
how to configure a database to use this procedure, or how to perform the  
decision support processing itself.  
To set up a replica database using the table files that are configured within a  
volume in a private disk group  
1
Use the following command on the primary host to see if the volume is  
associated with a version 20 data change object (DCO) and DCO volume that  
allow instant snapshots and Persistent FastResync to be used with the  
volume:  
# vxprint -g volumedg -F%instant volume  
This command returns onif the volume can be used for instant snapshot  
operations; otherwise, it returns off.  
Note: If the volume was created under VxVM 4.0 or a later release, and it is  
not associated with a new-style DCO object and DCO volume, follow the  
procedure described in “Preparing a volume for DRL and instant snapshots”  
on page 275.  
If the volume was created before release 4.0 of VxVM, and has any attached  
snapshot plexes, or is associated with any snapshot volumes, follow the  
procedure given in “Upgrading existing volumes to use version 20 DCOs” on  
page 279.  
2
Use the following command on the primary host to check whether  
FastResync is enabled on a volume:  
# vxprint -g volumedg -F%fastresync volume  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Configuring off-host processing 375  
Implementing off-host processing solutions  
This command returns onif FastResync is enabled; otherwise, it returns  
off.  
If FastResync is disabled, enable it using the following command on the  
primary host:  
# vxvol -g volumedg set fastresync=on volume  
3
4
Prepare the OHP host to receive the snapshot volume that contains the copy  
of the database tables. This may involve setting up private volumes to  
contain any redo logs, and configuring any files that are used to initialize  
the database.  
On the primary host, create a new volume in a separate disk group for use as  
the snapshot volume as described in “Creating a volume for use as a full-  
sized instant or linked break-off snapshot” on page 323. It is recommended  
that a snapshot disk group is dedicated to maintaining only those disks that  
are used for off-host processing.  
5
On the primary host, link the snapshot volume in the snapshot disk group to  
the data volume:  
# vxsnap -g volumedg -b addmir volume mirvol=snapvol \  
mirdg=snapvoldg  
You can use the vxsnapsnapwaitcommand to wait for synchronization of  
the linked snapshot volume to complete:  
# vxsnap -g volumedg snapwait volume mirvol=snapvol \  
mirdg=snapvoldg  
Note: This step sets up the snapshot volumes. When you are ready to create  
a replica database, proceed to step 6.  
6
7
On the primary host, suspend updates to the volume that contains the  
database tables. A database may have a hot backup mode that allows you to  
do this by temporarily suspending writes to its tables.  
Create the snapshot volume, snapvol, by running the following command on  
the primary host:  
# vxsnap -g volumedg make \  
source=volume/snapvol=snapvol/snapdg=snapvoldg  
If a database spans more than one volume, you can specify all the volumes  
and their snapshot volumes using one command, as shown in this example:  
# vxsnap -g dbasedg make \  
source=vol1/snapvol=snapvol1/snapdg=sdg \  
source=vol2/snapvol=snapvol2/snapdg=sdg \  
source=vol3/snapvol=snapvol3/snapdg=sdg  
This step makes the snapshot volumes ready for the decision-support  
processing cycle, and starts tracking changes to the original volumes.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
376 Configuring off-host processing  
Implementing off-host processing solutions  
8
9
On the primary host, if you temporarily suspended updates to a volume in  
step 6, release all the database tables from hot backup mode.  
On the primary host, deport the snapshot volume’s disk group using the  
following command:  
# vxdg deport snapvoldg  
10 On the OHP host where the replica database is to be set up, use the following  
command to import the snapshot volume’s disk group:  
# vxdg import snapvoldg  
11 The snapshot volume is initially disabled following the join. Use the  
following command on the OHP host to recover and restart the snapshot  
volume:  
# vxrecover -g snapvoldg -m snapvol  
12 On the OHP host, check and mount the snapshot volume. The following are  
sample commands for checking and mounting a file system:  
# fsck -F vxfs /dev/vx/rdsk/snapvoldg/snapvol  
# mount -F vxfs /dev/vx/dsk/snapvoldg/snapvol \  
mount_point  
13 On the OHP host, use the appropriate database commands to recover and  
start the replica database for its decision support role.  
When you want to resynchronize the snapshot volume’ s data with the primary  
database, you can reattach the snapshot volume to the original volume as  
follows:  
1
2
3
4
On the OHP host, shut down the replica database, and use the following  
command to unmount the snapshot volume:  
# umount mount_point  
On the OHP host, use the following command to deport the snapshot  
volume’s disk group:  
# vxdg deport snapvoldg  
On the primary host, re-import the snapshot volume’s disk group using the  
following command:  
# vxdg import snapvoldg  
The snapshot volume is initially disabled following the join. Use the  
following command on the primary host to recover and restart the snapshot  
volume:  
# vxrecover -g snapvoldg -m snapvol  
5
On the primary host, reattach the snapshot volume to its original volume  
using the following command:  
# vxsnap -g snapvoldg reattach snapvol source=vol \  
sourcedg=volumedg  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Configuring off-host processing 377  
Implementing off-host processing solutions  
For example, to reattach the snapshot volumes svol1, svol2and svol3:  
# vxsnap -g sdg reattach svol1 \  
source=vol1 sourcedg=dbasedg \  
svol2 source=vol2 sourcedg=dbasedg \  
svol3 source=vol3 sourcedg=dbasedg  
You can use the vxsnapsnapwaitcommand to wait for synchronization of  
the linked snapshot volume to complete:  
# vxsnap -g volumedg snapwait volume mirvol=snapvol  
You can then resume the procedure from step 6 on page 375.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
378 Configuring off-host processing  
Implementing off-host processing solutions  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter  
12  
Administering  
hot-relocation  
If a volume has a disk I/O failure (for example, the disk has an uncorrectable  
error), Veritas Volume Manager (VxVM) can detach the plex involved in the  
failure. I/O stops on that plex but continues on the remaining plexes of the  
volume.  
If a disk fails completely, VxVM can detach the disk from its disk group. All  
plexes on the disk are disabled. If there are any unmirrored volumes on a disk  
when it is detached, those volumes are also disabled.  
Note: Apparent disk failure may not be due to a fault in the physical disk media  
or the disk controller, but may instead be caused by a fault in an intermediate or  
ancillary component such as a cable, host bus adapter, or power supply.  
The hot-relocation feature in VxVM automatically detects disk failures, and  
notifies the system administrator and other nominated users of the failures by  
electronic mail. Hot-relocation also attempts to use spare disks and free disk  
space to restore redundancy and to preserve access to mirrored and RAID-5  
volumes. For more information, see the section, “How hot-relocation works” on  
page 380.  
If hot-relocation is disabled or you miss the electronic mail, you can use the  
vxprintcommand or the graphical user interface to examine the status of the  
disks. You may also see driver error messages on the console or in the system  
messages file.  
Failed disks must be removed and replaced manually as described in “Removing  
and replacing disks” on page 112.  
For more information about recovering volumes and their data after hardware  
failure, see the Veritas Volume Manager Troubleshooting Guide.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
380 Administering hot-relocation  
How hot-relocation works  
How hot-relocation works  
Hot-relocation allows a system to react automatically to I/O failures on  
redundant (mirrored or RAID-5) VxVM objects, and to restore redundancy and  
access to those objects. VxVM detects I/O failures on objects and relocates the  
affected subdisks to disks designated as spare disks or to free space within the  
disk group. VxVM then reconstructs the objects that existed before the failure  
and makes them redundant and accessible again.  
When a partial disk failure occurs (that is, a failure affecting only some subdisks  
on a disk), redundant data on the failed portion of the disk is relocated. Existing  
volumes on the unaffected portions of the disk remain accessible.  
Note: Hot-relocation is only performed for redundant (mirrored or RAID-5)  
subdisks on a failed disk. Non-redundant subdisks on a failed disk are not  
relocated, but the system administrator is notified of their failure.  
Hot-relocation is enabled by default and takes effect without the intervention of  
the system administrator when a failure occurs.  
The hot-relocation daemon, vxrelocd, detects and reacts to VxVM events that  
signify the following types of failures:  
Disk failure  
This is normally detected as a result of an I/O failure from a  
VxVM object. VxVM attempts to correct the error. If the  
error cannot be corrected, VxVM tries to access  
configuration information in the private region of the disk.  
If it cannot access the private region, it considers the disk  
failed.  
Plex failure  
This is normally detected as a result of an uncorrectable I/O  
error in the plex (which affects subdisks within the plex).  
For mirrored volumes, the plex is detached.  
RAID-5 subdisk failure  
This is normally detected as a result of an uncorrectable I/O  
error. The subdisk is detached.  
When vxrelocddetects such a failure, it performs the following steps:  
■
vxrelocdinforms the system administrator (and other nominated users) by  
electronic mail of the failure and which VxVM objects are affected.  
See “Partial disk failure mail messages” on page 383.  
See “Complete disk failure mail messages” on page 384.  
See “Modifying the behavior of hot-relocation” on page 395.  
■ vxrelocdnext determines if any subdisks can be relocated. vxrelocd  
looks for suitable space on disks that have been reserved as hot-relocation  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Administering hot-relocation 381  
How hot-relocation works  
spares (marked spare) in the disk group where the failure occurred. It then  
relocates the subdisks to use this space.  
■
■
If no spare disks are available or additional space is needed, vxrelocduses  
free space on disks in the same disk group, except those disks that have been  
excluded for hot-relocation use (marked nohotuse). When vxrelocdhas  
relocated the subdisks, it reattaches each relocated subdisk to its plex.  
Finally, vxrelocdinitiates appropriate recovery procedures. For example,  
recovery includes mirror resynchronization for mirrored volumes or data  
recovery for RAID-5 volumes. It also notifies the system administrator of  
the hot-relocation and recovery actions that have been taken.  
If relocation is not possible, vxrelocdnotifies the system administrator and  
takes no further action.  
Note: Hot-relocation does not guarantee the same layout of data or the same  
performance after relocation. The system administrator can make configuration  
changes after hot-relocation occurs.  
Relocation of failing subdisks is not possible in the following cases:  
■
The failing subdisks are on non-redundant volumes (that is, volumes of  
types other than mirrored or RAID-5).  
■
■
There are insufficient spare disks or free disk space in the disk group.  
The only available space is on a disk that already contains a mirror of the  
failing plex.  
■
The only available space is on a disk that already contains the RAID-5 log  
plex or one of its healthy subdisks. Failing subdisks in the RAID-5 plex  
cannot be relocated.  
■
■
If a mirrored volume has a dirty region logging (DRL) log subdisk as part of  
its data plex, failing subdisks belonging to that plex cannot be relocated.  
If a RAID-5 volume log plex or a mirrored volume DRL log plex fails, a new  
log plex is created elsewhere. There is no need to relocate the failed subdisks  
of the log plex.  
See the vxrelocd(1M) manual page for more information about the  
hot-relocation daemon.  
Figure 12-1 illustrates the hot-relocation process in the case of the failure of a  
single subdisk of a RAID-5 volume.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
382 Administering hot-relocation  
How hot-relocation works  
Figure 12-1  
Example of hot-relocation for a subdisk in a RAID-5 volume  
a) Disk group contains five disks. Two RAID-5 volumes are configured across four of the disks.  
One spare disk is available for hot-relocation.  
mydg01  
mydg02  
mydg03  
mydg04  
mydg05  
mydg01-01  
mydg02-01  
mydg02-02  
mydg03-01  
mydg03-02  
mydg04-01  
Spare Disk  
b) Subdisk mydg02-01 in one RAID-5 volume fails. Hot-relocation replaces it with subdisk mydg05-0  
that it has created on the spare disk, and then initiates recovery of the RAID-5 volume.  
mydg01  
mydg02  
mydg03  
mydg04  
mydg05  
mydg01-01  
mydg02-01  
mydg03-01  
mydg03-02  
mydg04-01  
mydg05-01  
X
mydg02-02  
c) RAID-5 recovery recreates subdisk mydg02-01’s data and parity on subdisk mydg05-01 from  
the data and parity information remaining on subdisks mydg01-01 and mydg03-01.  
mydg01  
mydg02  
mydg03  
mydg04  
mydg05  
mydg01-01  
mydg02-01  
mydg03-01  
mydg03-02  
mydg04-01  
mydg05-01  
X
mydg02-02  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Administering hot-relocation 383  
How hot-relocation works  
Partial disk failure mail messages  
If hot-relocation is enabled when a plex or disk is detached by a failure, mail  
indicating the failed objects is sent to root. If a partial disk failure occurs, the  
mail identifies the failed plexes. For example, if a disk containing mirrored  
volumes fails, you can receive mail information as shown in the following  
example:  
To: root  
Subject: Volume Manager failures on host teal  
Failures have been detected by the Veritas Volume Manager:  
failed plexes:  
home-02  
src-02  
See “Modifying the behavior of hot-relocation” on page 395 for information on  
how to send the mail to users other than root.  
You can determine which disk is causing the failures in the above example  
message by using the following command:  
# vxstat -g mydg -s -ff home-02 src-02  
The -soption asks for information about individual subdisks, and the -ff  
option displays the number of failed read and write operations. The following  
output display is typical:  
FAILED  
TYP NAME  
READS  
WRITES  
sd mydg01-04  
sd mydg01-06  
sd mydg02-03  
sd mydg02-04  
0
0
1
1
0
0
0
0
This example shows failures on reading from subdisks mydg02-03and  
mydg02-04of disk mydg02.  
Hot-relocation automatically relocates the affected subdisks and initiates any  
necessary recovery procedures. However, if relocation is not possible or the  
hot-relocation feature is disabled, you must investigate the problem and  
attempt to recover the plexes. Errors can be caused by cabling failures, so check  
the cables connecting your disks to your system. If there are obvious problems,  
correct them and recover the plexes using the following command:  
# vxrecover -b -g mydg home src  
This starts recovery of the failed plexes in the background (the command  
prompt reappears before the operation completes). If an error message appears  
later, or if the plexes become detached again and there are no obvious cabling  
failures, replace the disk (see “Removing and replacing disks” on page 112).  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
384 Administering hot-relocation  
How hot-relocation works  
Complete disk failure mail messages  
If a disk fails completely and hot-relocation is enabled, the mail message lists  
the disk that failed and all plexes that use the disk. For example, you can receive  
mail as shown in this example display:  
To: root  
Subject: Volume Manager failures on host teal  
Failures have been detected by the Veritas Volume Manager:  
failed disks:  
mydg02  
failed plexes:  
home-02  
src-02  
mkting-01  
failing disks:  
mydg02  
This message shows that mydg02was detached by a failure. When a disk is  
detached, I/O cannot get to that disk. The plexes home-02, src-02, and  
mkting-01were also detached (probably because of the failure of the disk).  
As described in “Partial disk failure mail messages” on page 383, the problem  
can be a cabling error. If the problem is not a cabling error, replace the disk (see  
“Removing and replacing disks” on page 112).  
How space is chosen for relocation  
A spare disk must be initialized and placed in a disk group as a spare before it  
can be used for replacement purposes. If no disks have been designated as  
spares when a failure occurs, VxVM automatically uses any available free space  
in the disk group in which the failure occurs. If there is not enough spare disk  
space, a combination of spare space and free space is used.  
The free space used in hot-relocation must not have been excluded from  
hot-relocation use. Disks can be excluded from hot-relocation use by using  
vxdiskadm, vxeditor the Veritas Enterprise Administrator (VEA).  
You can designate one or more disks as hot-relocation spares within each disk  
group. Disks can be designated as spares by using vxdiskadm, vxedit, or the  
VEA. Disks designated as spares do not participate in the free space model and  
should not have storage space allocated on them.  
When selecting space for relocation, hot-relocation preserves the redundancy  
characteristics of the VxVM object to which the relocated subdisk belongs. For  
example, hot-relocation ensures that subdisks from a failed plex are not  
relocated to a disk containing a mirror of the failed plex. If redundancy cannot  
be preserved using any available spare disks and/or free space, hot-relocation  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Administering hot-relocation 385  
Configuring a system for hot-relocation  
does not take place. If relocation is not possible, the system administrator is  
notified and no further action is taken.  
From the eligible disks, hot-relocation attempts to use the disk that is “closest”  
to the failed disk. The value of “closeness” depends on the controller, target, and  
disk number of the failed disk. A disk on the same controller as the failed disk is  
closer than a disk on a different controller. A disk under the same target as the  
failed disk is closer than one on a different target.  
Hot-relocation tries to move all subdisks from a failing drive to the same  
destination disk, if possible.  
When hot-relocation takes place, the failed subdisk is removed from the  
configuration database, and VxVM ensures that the disk space used by the failed  
subdisk is not recycled as free space.  
Configuring a system for hot-relocation  
By designating spare disks and making free space on disks available for use by  
hot relocation, you can control how disk space is used for relocating subdisks in  
the event of a disk failure. If the combined free space and space on spare disks is  
not sufficient or does not meet the redundancy constraints, the subdisks are not  
relocated.  
■
To find out which disks are spares or are excluded from hot-relocation, see  
“Displaying spare disk information” on page 386.  
You can prepare for hot-relocation by designating one or more disks per disk  
group as hot-relocation spares.  
■
To designate a disk as being a hot-relocation spare for a disk group, see  
“Marking a disk as a hot-relocation spare” on page 387.  
■
To remove a disk from use as a hot-relocation spare, see “Removing a disk  
from use as a hot-relocation spare” on page 388.  
If no spares are available at the time of a failure or if there is not enough space  
on the spares, free space on disks in the same disk group as where the failure  
occurred is automatically used, unless it has been excluded from hot-relocation  
use.  
■
To exclude a disk from hot-relocation use, see “Excluding a disk from  
hot-relocation use” on page 388.  
■
To make a disk available for hot-relocation use, see “Making a disk available  
for hot-relocation use” on page 389.  
Depending on the locations of the relocated subdisks, you can choose to move  
them elsewhere after hot-relocation occurs (see “Configuring hot-relocation to  
use only spare disks” on page 390).  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
386 Administering hot-relocation  
Displaying spare disk information  
After a successful relocation, remove and replace the failed disk as described in  
“Removing and replacing disks” on page 112).  
Displaying spare disk information  
Use the following command to display information about spare disks that are  
available for relocation:  
# vxdg [-g diskgroup] spare  
The following is example output:  
GROUP DISK  
mydg mydg02  
DEVICE  
c0t2d0  
TAG  
c0t2d0  
OFFSET  
0
LENGTH  
658007  
FLAGS  
s
Here mydg02is the only disk designated as a spare in the mydgdisk group. The  
LENGTHfield indicates how much spare space is currently available on mydg02  
for relocation.  
The following commands can also be used to display information about disks  
that are currently designated as spares:  
■ vxdisklistlists disk information and displays spare disks with a spare  
flag.  
■ vxprintlists disk and other information and displays spare disks with a  
SPAREflag.  
■
The listmenu item on the vxdiskadmmain menu lists all disks including  
spare disks.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Administering hot-relocation 387  
Marking a disk as a hot-relocation spare  
Marking a disk as a hot-relocation spare  
Hot-relocation allows the system to react automatically to I/O failure by  
relocating redundant subdisks to other disks. Hot-relocation then restores the  
affected VxVM objects and data. If a disk has already been designated as a spare  
in the disk group, the subdisks from the failed disk are relocated to the spare  
disk. Otherwise, any suitable free space in the disk group is used.  
To designate a disk as a hot-relocation spare, enter the following command:  
# vxedit [-g diskgroup] set spare=on diskname  
where diskname is the disk media name.  
For example, to designate mydg01as a spare in the disk group, mydg, enter the  
following command:  
# vxedit -g mydg set spare=on mydg01  
You can use the vxdisklistcommand to confirm that this disk is now a spare;  
mydg01should be listed with a spareflag.  
Any VM disk in this disk group can now use this disk as a spare in the event of a  
failure. If a disk fails, hot-relocation automatically occurs (if possible). You are  
notified of the failure and relocation through electronic mail. After successful  
relocation, you may want to replace the failed disk.  
To use vxdiskadm to designate a disk as a hot-relocation spare  
1
Select menu item 11 (Mark a disk as a spare for a disk group)  
from the vxdiskadmmain menu.  
2
At the following prompt, enter a disk media name (such as mydg01):  
Menu: VolumeManager/Disk/MarkSpareDisk  
Use this operation to mark a disk as a spare for a disk group.  
This operation takes, as input, a disk name. This is the same  
name that you gave to the disk when you added the disk to the  
disk group.  
Enter disk name [<disk>,list,q,?] mydg01  
The following notice is displayed when the disk has been marked as spare:  
VxVM NOTICE V-5-2-219 Marking of mydg01 in mydg as a spare disk  
is complete.  
3
At the following prompt, indicate whether you want to add more disks as  
spares (y) or return to the vxdiskadmmain menu (n):  
Mark another disk as a spare? [y,n,q,?] (default: n)  
Any VM disk in this disk group can now use this disk as a spare in the event  
of a failure. If a disk fails, hot-relocation should automatically occur (if  
possible). You should be notified of the failure and relocation through  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
388 Administering hot-relocation  
Removing a disk from use as a hot-relocation spare  
electronic mail. After successful relocation, you may want to replace the  
failed disk.  
Removing a disk from use as a hot-relocation spare  
While a disk is designated as a spare, the space on that disk is not used for the  
creation of VxVM objects within its disk group. If necessary, you can free a spare  
disk for general use by removing it from the pool of hot-relocation disks.  
To remove a spare from the hot-relocation pool, use the following command:  
# vxedit [-g diskgroup] set spare=off diskname  
where diskname is the disk media name.  
For example, to make mydg01available for normal use in the disk group, mydg,  
use the following command:  
# vxedit -g mydg set spare=off mydg01  
To use vxdiskadm to remove a disk from the hot-relocation pool  
1
Select menu item 12 (Turn off the spare flag on a disk)from  
the vxdiskadmmain menu.  
2
At the following prompt, enter the disk media name of a spare disk (such as  
mydg01):  
Menu: VolumeManager/Disk/UnmarkSpareDisk  
Use this operation to turn off the spare flag on a disk.  
This operation takes, as input, a disk name. This is the same  
name that you gave to the disk when you added the disk to the  
disk group.  
Enter disk name [<disk>,list,q,?] mydg01  
The following confirmation is displayed:  
VxVM NOTICE V-5-2-143 Disk mydg01 in mydg no longer marked as  
a spare disk.  
3
At the following prompt, indicate whether you want to disable more spare  
disks (y) or return to the vxdiskadmmain menu (n):  
Turn-off spare flag on another disk? [y,n,q,?] (default: n)  
Excluding a disk from hot-relocation use  
To exclude a disk from hot-relocation use, use the following command:  
# vxedit [-g diskgroup] set nohotuse=on diskname  
where diskname is the disk media name.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                 
Administering hot-relocation 389  
Making a disk available for hot-relocation use  
To use vxdiskadm to exclude a disk from hot-relocation use  
1
Select menu item 15 (Exclude a disk from hot-relocation use)  
from the vxdiskadmmain menu.  
2
At the following prompt, enter the disk media name (such as mydg01):  
Exclude a disk from hot-relocation use  
Menu: VolumeManager/Disk/UnmarkSpareDisk  
Use this operation to exclude a disk from hot-relocation use.  
This operation takes, as input, a disk name. This is the same  
name that you gave to the disk when you added the disk to the  
disk group.  
Enter disk name [<disk>,list,q,?] mydg01  
The following confirmation is displayed:  
VxVM INFO V-5-2-925 Excluding mydg01 in mydg from hot-  
relocation use is complete.  
3
At the following prompt, indicate whether you want to add more disks to be  
excluded from hot-relocation (y) or return to the vxdiskadmmain menu (n):  
Exclude another disk from hot-relocation use? [y,n,q,?]  
(default: n)  
Making a disk available for hot-relocation use  
Free space is used automatically by hot-relocation in case spare space is not  
sufficient to relocate failed subdisks. You can limit this free space usage by  
hot-relocation by specifying which free disks should not be touched by  
hot-relocation. If a disk was previously excluded from hot-relocation use, you  
can undo the exclusion and add the disk back to the hot-relocation pool.  
To make a disk available for hot-relocation use, use the following command:  
# vxedit [-g diskgroup] set nohotuse=off diskname  
To use vxdiskadm to make a disk available for hot-relocation use  
1
Select menu item 16 (Make a disk available for  
hot-relocation use) from the vxdiskadmmain menu.  
2
At the following prompt, enter the disk media name (such as mydg01):  
Menu: VolumeManager/Disk/UnmarkSpareDisk  
Use this operation to make a disk available for hot-relocation  
use. This only applies to disks that were previously excluded  
from hot-relocation use. This operation takes, as input, a  
disk name. This is the same name that you gave to the disk when  
you added the disk to the disk group.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
390 Administering hot-relocation  
Configuring hot-relocation to use only spare disks  
Enter disk name [<disk>,list,q,?] mydg01  
The following confirmation is displayed:  
V-5-2-932 Making mydg01 in mydg available for hot-relocation  
use is complete.  
3
At the following prompt, indicate whether you want to add more disks to be  
excluded from hot-relocation (y) or return to the vxdiskadmmain menu (n):  
Make another disk available for hot-relocation use? [y,n,q,?]  
(default: n)  
Configuring hot-relocation to use only spare disks  
If you want VxVM to use only spare disks for hot-relocation, add the following  
line to the file /etc/default/vxassist:  
spare=only  
If not enough storage can be located on disks marked as spare, the relocation  
fails. Any free space on non-spare disks is not used.  
Moving and unrelocating subdisks  
When hot-relocation occurs, subdisks are relocated to spare disks and/or  
available free space within the disk group. The new subdisk locations may not  
provide the same performance or data layout that existed before hot-relocation  
took place. You can move the relocated subdisks (after hot-relocation is  
complete) to improve performance.  
You can also move the relocated subdisks off the spare disks to keep the spare  
disk space free for future hot-relocation needs. Another reason for moving  
subdisks is to recreate the configuration that existed before hot-relocation  
occurred.  
During hot-relocation, one of the electronic mail messages sent to rootis  
shown in the following example:  
To: root  
Subject: Volume Manager failures on host teal  
Attempting to relocate subdisk mydg02-03 from plex home-02.  
Dev_offset 0 length 1164 dm_name mydg02 da_name c0t5d0.  
The available plex home-01 will be used to recover the data.  
This message has information about the subdisk before relocation and can be  
used to decide where to move the subdisk after relocation.  
Here is an example message that shows the new location for the relocated  
subdisk:  
To: root  
Subject: Attempting VxVM relocation on host teal  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
Administering hot-relocation 391  
Moving and unrelocating subdisks  
Volume home Subdisk mydg02-03 relocated to mydg05-01,  
but not yet recovered.  
Before you move any relocated subdisks, fix or replace the disk that failed (as  
described in “Removing and replacing disks” on page 112). Once this is done,  
you can move a relocated subdisk back to the original disk as described in the  
following sections.  
Caution: During subdisk move operations, RAID-5 volumes are not redundant.  
Moving and unrelocating subdisks using vxdiskadm  
To move the hot-relocated subdisks back to the disk where they originally  
resided after the disk has been replaced following a failure  
1
Select menu item 14 (Unrelocate subdisks back to a disk)from  
the vxdiskadmmain menu.  
2
This option prompts for the original disk media name first.  
Enter the disk media name where the hot-relocated subdisks originally  
resided at the following prompt:  
Enter the original disk name [<disk>,list,q,?]  
If there are no hot-relocated subdisks in the system, vxdiskadmdisplays  
Currently there are no hot-relocated disks, and asks you to  
press Return to continue.  
3
4
You are next asked if you want to move the subdisks to a destination disk  
other than the original disk.  
Unrelocate to a new disk [y,n,q,?] (default: n)  
If moving subdisks to their original offsets is not possible, you can choose to  
unrelocate the subdisks forcibly to the specified disk, but not necessarily to  
the same offsets.  
Use -f option to unrelocate the subdisks if moving to the exact  
offset fails? [y,n,q,?] (default: n)  
5
If you entered yat step 4 to unrelocate the subdisks forcibly, enter yor press  
Return at the following prompt to confirm the operation:  
Requested operation is to move all the subdisks which were  
hot-relocated from mydg10 back to mydg10 of disk group mydg.  
Continue with operation? [y,n,q,?] (default: y)  
A status message is displayed at the end of the operation.  
VxVM INFO V-5-2-954 Unrelocate to disk mydg10 is complete.  
As an alternative to this procedure, use either the vxassistcommand or the  
vxunreloccommand directly, as described in “Moving and unrelocating  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
392 Administering hot-relocation  
Moving and unrelocating subdisks  
subdisks using vxassist” on page 392 and “Moving and unrelocating subdisks  
using vxunreloc” on page 392.  
Moving and unrelocating subdisks using vxassist  
You can use the vxassistcommand to move and unrelocate subdisks. For  
example, to move the relocated subdisks on mydg05belonging to the volume  
homeback to mydg02, enter the following command:  
# vxassist -g mydg move home !mydg05 mydg02  
Here, !mydg05specifies the current location of the subdisks, and mydg02  
specifies where the subdisks should be relocated.  
If the volume is enabled, subdisks within detached or disabled plexes, and  
detached log or RAID-5 subdisks, are moved without recovery of data.  
If the volume is not enabled, subdisks within STALE or OFFLINE plexes, and  
stale log or RAID-5 subdisks, are moved without recovery. If there are other  
subdisks within a non-enabled volume that require moving, the relocation fails.  
For enabled subdisks in enabled plexes within an enabled volume, data is moved  
to the new location, without loss of either availability or redundancy of the  
volume.  
Moving and unrelocating subdisks using vxunreloc  
VxVM hot-relocation allows the system to automatically react to I/O failures on  
a redundant VxVM object at the subdisk level and then take necessary action to  
make the object available again. This mechanism detects I/O failures in a  
subdisk, relocates the subdisk, and recovers the plex associated with the  
subdisk. After the disk has been replaced, vxunrelocallows you to restore the  
system back to the configuration that existed before the disk failure.  
vxunrelocallows you to move the hot-relocated subdisks back onto a disk that  
was replaced due to a failure.  
When vxunrelocis invoked, you must specify the disk media name where the  
hot-relocated subdisks originally resided. When vxunrelocmoves the  
subdisks, it moves them to the original offsets. If you try to unrelocate to a disk  
that is smaller than the original disk that failed,vxunrelocdoes nothing except  
return an error.  
vxunreloc provides an option to move the subdisks to a different disk from  
where they were originally relocated. It also provides an option to unrelocate  
subdisks to a different offset as long as the destination disk is large enough to  
accommodate all the subdisks.  
If vxunreloccannot replace the subdisks back to the same original offsets, a  
force option is available that allows you to move the subdisks to a specified disk  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Administering hot-relocation 393  
Moving and unrelocating subdisks  
without using the original offsets. Refer to the vxunreloc(1M) manual page for  
more information.  
The examples in the following sections demonstrate the use of vxunreloc.  
Moving hot-relocated subdisks back to their original disk  
Assume that mydg01failed and all the subdisks were relocated. After mydg01is  
replaced, vxunreloccan be used to move all the hot-relocated subdisks back to  
mydg01.  
# vxunreloc -g mydg mydg01  
Moving hot-relocated subdisks back to a different disk  
The vxunrelocutility provides the -noption to move the subdisks to a  
different disk from where they were originally relocated.  
Assume that mydg01failed, and that all of the subdisks that resided on it were  
hot-relocated to other disks. vxunrelocprovides an option to move the  
subdisks to a different disk from where they were originally relocated. After the  
disk is repaired, it is added back to the disk group using a different name, for  
example, mydg05. If you want to move all the hot-relocated subdisks back to the  
new disk, the following command can be used:  
# vxunreloc -g mydg -n mydg05 mydg01  
The destination disk should have at least as much storage capacity as was in use  
on the original disk. If there is not enough space, the unrelocate operation will  
fail and none of the subdisks will be moved.  
Forcing hot-relocated subdisks to accept different offsets  
By default, vxunrelocattempts to move hot-relocated subdisks to their  
original offsets. However, vxunrelocfails if any subdisks already occupy part or  
all of the area on the destination disk. In such a case, you have two choices:  
■
■
Move the existing subdisks somewhere else, and then re-run vxunreloc.  
Use the -foption provided by vxunrelocto move the subdisks to the  
destination disk, but leave it to vxunrelocto find the space on the disk. As  
long as the destination disk is large enough so that the region of the disk for  
storing subdisks can accommodate all subdisks, all the hot-relocated  
subdisks will be unrelocated without using the original offsets.  
Assume that mydg01failed and the subdisks were relocated and that you want  
to move the hot-relocated subdisks to mydg05where some subdisks already  
reside. You can use the force option to move the hot-relocated subdisks to  
mydg05, but not to the exact offsets:  
# vxunreloc -g mydg -f -n mydg05 mydg01  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
394 Administering hot-relocation  
Moving and unrelocating subdisks  
Examining which subdisks were hot-relocated from a disk  
If a subdisk was hot relocated more than once due to multiple disk failures, it  
can still be unrelocated back to its original location. For instance, if mydg01  
failed and a subdisk named mydg01-01was moved to mydg02, and then  
mydg02experienced disk failure, all of the subdisks residing on it, including the  
one which was hot-relocated to it, will be moved again. When mydg02was  
replaced, a vxunrelocoperation for mydg02will do nothing to the hot-relocated  
subdisk mydg01-01. However, a replacement of mydg01followed by a  
vxunrelocoperation, moves mydg01-01back to mydg01if vxunrelocis run  
immediately after the replacement.  
After the disk that experienced the failure is fixed or replaced, vxunreloccan  
be used to move all the hot-relocated subdisks back to the disk. When a subdisk  
is hot-relocated, its original disk-media name and the offset into the disk are  
saved in the configuration database. When a subdisk is moved back to the  
original disk or to a new disk using vxunreloc, the information is erased. The  
original disk-media name and the original offset are saved in the subdisk  
records. To print all of the subdisks that were hot-relocated from mydg01in the  
mydgdisk group, use the following command:  
# vxprint -g mydg -se 'sd_orig_dmname="mydg01"'  
Restarting vxunreloc after errors  
vxunrelocmoves subdisks in three phases:  
■
vxunreloccreates as many subdisks on the specified destination disk as  
there are subdisks to be unrelocated. The string UNRELOCis placed in the  
commentfield of each subdisk record.  
Creating the subdisk is an all-or-nothing operation. If vxunreloccannot  
create all the subdisks successfully, none are created, and vxunrelocexits.  
■
■
vxunrelocmoves the data from each subdisk to the corresponding newly  
created subdisk on the destination disk.  
When all subdisk data moves have been completed successfully, vxunreloc  
sets the commentfield to the null string for each subdisk on the destination  
disk whose commentfield is currently set to UNRELOC.  
The commentfields of all the subdisks on the destination disk remain marked as  
UNRELOCuntil phase 3 completes. If its execution is interrupted, vxunreloccan  
subsequently re-use subdisks that it created on the destination disk during a  
previous execution, but it does not use any data that was moved to the  
destination disk.  
If a subdisk data move fails, vxunrelocdisplays an error message and exits.  
Determine the problem that caused the move to fail, and fix it before  
re-executing vxunreloc.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Administering hot-relocation 395  
Modifying the behavior of hot-relocation  
If the system goes down after the new subdisks are created on the destination  
disk, but before all the data has been moved, re-execute vxunrelocwhen the  
system has been rebooted.  
Caution: Do not modify the string UNRELOCin the comment field of a subdisk  
record.  
Modifying the behavior of hot-relocation  
Hot-relocation is turned on as long as the vxrelocdprocess is running. You  
should normally leave hot-relocation turned on so that you can take advantage  
of this feature if a failure occurs. However, if you choose to disable  
hot-relocation (perhaps because you do not want the free space on your disks to  
be used for relocation), you can prevent vxrelocdfrom starting at system  
startup time by editing the startup file that invokes vxrelocd:  
/sbin/init.d/vxvm-recover.  
You can alter the behavior of vxrelocdas follows:  
◆
To prevent vxrelocdstarting, comment out the entry that invokes it in the  
startup file:  
# nohup vxrelocd root &  
◆
By default, vxrelocdsends electronic mail to rootwhen failures are  
detected and relocation actions are performed. You can instruct vxrelocd  
to notify additional users by adding the appropriate user names as shown  
here:  
nohup vxrelocd root user1 user2 &  
◆
To reduce the impact of recovery on system performance, you can instruct  
vxrelocdto increase the delay between the recovery of each region of the  
volume, as shown in the following example:  
nohup vxrelocd -o slow[=IOdelay] root &  
where the optional IOdelay value indicates the desired delay in  
milliseconds. The default value for the delay is 250 milliseconds.  
After making changes to the way vxrelocdis invoked in the startup file, reboot  
the system so that the changes go into effect.  
You can also stop hot-relocation at any time by killing the vxrelocdprocess  
(this should not be done while a hot-relocation attempt is in progress).  
When executing vxrelocdmanually, either include /etc/vx/binin your  
PATH or specify vxrelocd’s absolute pathname, for example:  
# PATH=/etc/vx/bin:$PATH  
# export PATH  
# nohup vxrelocd root &  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
396 Administering hot-relocation  
Modifying the behavior of hot-relocation  
Alternatively, you can use the following command:  
# nohup /etc/vx/bin/vxrelocd root user1 user2 &  
See the vxrelocd(1M) manual page for more information.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter  
13  
Administering  
cluster functionality  
A cluster consists of a number of hosts or nodes that share a set of disks. The  
main benefits of cluster configurations are:  
Availability  
If one node fails, the other nodes can still access the shared  
disks. When configured with suitable software, mission-critical  
applications can continue running by transferring their  
execution to a standby node in the cluster. This ability to provide  
continuous uninterrupted service by switching to redundant  
hardware is commonly termed failover.  
Failover is transparent to users and high-level applications for  
database and file-sharing. You must configure cluster  
management software, such as Veritas Cluster Server (VCS), to  
monitor systems and services, and to restart applications on  
another node in the event of either hardware or software failure.  
VCS also allows you to perform general administration tasks  
such as making nodes join or leave a cluster.  
Note that a standby node need not remain idle. It could be used  
to serve other applications in parallel.  
Off-host processing  
Clusters can reduce contention for system resources by  
performing activities such as backup, decision support and  
report generation on the more lightly loaded nodes of the  
cluster. This allows businesses to derive enhanced value from  
their investment in cluster systems.  
The cluster functionality of Veritas Volume Manager (CVM) allows up to 16  
nodes in a cluster to simultaneously access and manage a set of disks under  
VxVM control (VM disks). The same logical view of disk configuration and any  
changes to this is available on all the nodes. When the cluster functionality is  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
398 Administering cluster functionality  
Overview of cluster volume management  
enabled, all the nodes in the cluster can share VxVM objects such as shared disk  
groups. Private disk groups are supported in the same way as in a non-clustered  
environment. This chapter discusses the cluster functionality that is provided  
with VxVM.  
Note: You need an additional license to use this feature.  
This chapter does not discuss Veritas Storage Foundation Cluster File System  
(SFCFS) nor cluster management software such as Veritas Cluster Server (VCS).  
Such products are separately licensed, and are not included with Veritas Volume  
Manager. See the documentation provided with those products for more  
information about them.  
For information about administering a cluster that is under the control of HP  
Serviceguard, refer to the HP Serviceguard Storage Management Suite  
documentation.  
For additional information about using the Dynamic Multipathing (DMP)  
feature of VxVM in a clustered environment, see “DMP in a clustered  
environment” on page 132.  
For information about administering campus cluster configurations (also known  
as stretch cluster or remote mirror configurations), see “Administering sites and  
remote mirrors” on page 431.  
Overview of cluster volume management  
In recent years, tightly-coupled cluster systems have become increasingly  
popular in the realm of enterprise-scale mission-critical data processing. The  
primary advantage of clusters is protection against hardware failure. Should the  
primary node fail or otherwise become unavailable, applications can continue to  
run by transferring their execution to standby nodes in the cluster. This ability  
to provide continuous availability of service by switching to redundant  
hardware is commonly termed failover.  
Another major advantage of clustered systems is their ability to reduce  
contention for system resources caused by activities such as backup, decision  
support and report generation. Businesses can derive enhanced value from their  
investment in cluster systems by performing such operations on lightly loaded  
nodes in the cluster rather than on the heavily loaded nodes that answer  
requests for service. This ability to perform some operations on the lightly  
loaded nodes is commonly termed load balancing.  
The cluster functionality of VxVM works together with the cluster monitor  
daemon that is provided by VCS or by the host operating system. When  
configured correctly, the cluster monitor informs VxVM of changes in cluster  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Administering cluster functionality 399  
Overview of cluster volume management  
membership. Each node starts up independently and has its own cluster monitor  
plus its own copies of the operating system and VxVM with support for cluster  
functionality. When a node joins a cluster, it gains access to shared disk groups  
and volumes. When a node leaves a cluster, it no longer has access to these  
shared objects. A node joins a cluster when you issue the appropriate command  
on that node. In an HP Serviceguard cluster, a node can join the cluster  
automatically at boot time.  
Caution: The cluster functionality of VxVM is supported only when used in  
conjunction with a cluster monitor that has been configured correctly to work  
with VxVM.  
Figure 13-1 illustrates a simple cluster arrangement consisting of four nodes  
with similar or identical hardware characteristics (CPUs, RAM and host  
adapters), and configured with identical software (including the operating  
system). The nodes are fully connected by a private network and they are also  
separately connected to shared external storage (either disk arrays or JBODs:  
just a bunch of disks) via SCSI or via Fibre Channel in a Storage Area Network  
(SAN).  
Note: In this example, each node has two independent paths to the disks, which  
are configured in one or more cluster-shareable disk groups. Multiple paths  
provide resilience against failure of one of the paths, but this is not a  
requirement for cluster configuration. Disks may also be connected by single  
paths.  
The private network allows the nodes to share information about system  
resources and about each other’s state. Using the private network, any node can  
recognize which other nodes are currently active, which are joining or leaving  
the cluster, and which have failed. The private network requires at least two  
communication channels to provide redundancy against one of the channels  
failing. If only one channel were used, its failure would be indistinguishable  
from node failure—a condition known as network partitioning.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
400 Administering cluster functionality  
Overview of cluster volume management  
Figure 13-1  
Example of a 4-node cluster  
Redundant private network  
Node 0  
(master)  
Node 1  
(slave)  
Node 2  
(slave)  
Node 3  
(slave)  
Redundant SCSI or  
Fibre Channel  
connectivity  
Cluster-shareable  
disks  
Cluster-shareable disk  
groups  
To the cluster monitor, all nodes are the same. VxVM objects configured within  
shared disk groups can potentially be accessed by all nodes that join the cluster.  
However, the cluster functionality of VxVM requires that one node act as the  
master node; all other nodes in the cluster are slave nodes. Any node is capable  
of being the master node, and it is responsible for coordinating certain VxVM  
activities.  
Note: You must run commands that configure or reconfigure VxVM objects on  
the master node. Tasks that must be initiated from the master node include  
setting up shared disk groups, creating and reconfiguring volumes, and  
performing snapshot operations.  
VxVM determines that the first node to join a cluster performs the function of  
master node. If the master node leaves a cluster, one of the slave nodes is chosen  
to be the new master. In “Example of a 4-node cluster,” node 0 is the master  
node and nodes 1, 2 and 3 are slave nodes.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Administering cluster functionality 401  
Overview of cluster volume management  
Private and shared disk groups  
Two types of disk groups are defined:  
Private disk group  
Belongs to only one node. A private disk group can only be  
imported by one system at a time. Disks in a private disk group  
may be physically accessible from one or more systems, but  
access is restricted to one system only. The boot disk group  
(usually aliased by the reserved disk group name bootdg) is  
always a private disk group.  
Shared disk group  
Can be shared by all nodes. A shared (or cluster-shareable) disk  
group is imported by all cluster nodes. Disks in a shared disk  
group must be physically accessible from all systems that may  
join the cluster.  
In a cluster, most disk groups are shared. Disks in a shared disk group are  
accessible from all nodes in a cluster, allowing applications on multiple cluster  
nodes to simultaneously access the same disk. A volume in a shared disk group  
can be simultaneously accessed by more than one node in the cluster, subject to  
licensing and disk group activation mode restrictions.  
You can use the vxdgcommand to designate a disk group as cluster-shareable as  
described in “Importing disk groups as shared” on page 423. When a disk group  
is imported as cluster-shareable for one node, each disk header is marked with  
the cluster ID. As each node subsequently joins the cluster, it recognizes the disk  
group as being cluster-shareable and imports it. As system administrator, you  
can also import or deport a shared disk group at any time; the operation takes  
place in a distributed fashion on all nodes.  
Each physical disk is marked with a unique disk ID. When cluster functionality  
for VxVM starts on the master, it imports all shared disk groups (except for any  
that have the noautoimportattribute set). When a slave tries to join a cluster,  
the master sends it a list of the disk IDs that it has imported, and the slave  
checks to see if it can access them all. If the slave cannot access one of the listed  
disks, it abandons its attempt to join the cluster. If it can access all of the listed  
disks, it joins the cluster and imports the same shared disk groups as the master.  
When a node leaves the cluster gracefully, it deports all its imported shared disk  
groups, but they remain imported on the surviving nodes.  
Reconfiguring a shared disk group is performed with the cooperation of all  
nodes. Configuration changes to the disk group happen simultaneously on all  
nodes and the changes are identical. Such changes are atomic in nature, which  
means that they either occur simultaneously on all nodes or not at all.  
Whether all members of the cluster have simultaneous read and write access to  
a cluster-shareable disk group depends on its activation mode setting as  
discussed in “Activation modes of shared disk groups.” The data contained in a  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
402 Administering cluster functionality  
Overview of cluster volume management  
cluster-shareable disk group is available as long as at least one node is active in  
the cluster. The failure of a cluster node does not affect access by the remaining  
active nodes. Regardless of which node accesses a cluster-shareable disk group,  
the configuration of the disk group looks the same.  
Note: Applications running on each node can access the data on the VM disks  
simultaneously. VxVM does not protect against simultaneous writes to shared  
volumes by more than one node. It is assumed that applications control  
consistency (by using a distributed lock manager, for example).  
Activation modes of shared disk groups  
A shared disk group must be activated on a node in order for the volumes in the  
disk group to become accessible for application I/O from that node. The ability  
of applications to read from or to write to volumes is dictated by the activation  
mode of a shared disk group. Valid activation modes for a shared disk group are  
exclusivewrite, readonly, sharedread, sharedwrite, and off(inactive).  
These activation modes are described in detail in the table “Activation modes for  
shared disk groups.”  
Note: The default activation mode for shared disk groups is off(inactive).  
Special uses of clusters, such as high availability (HA) applications and off-host  
backup, can use disk group activation to explicitly control volume access from  
different nodes in the cluster.  
Table 13-1  
Activation modes for shared disk groups  
Activation mode  
Description  
exclusivewrite (ew)  
The node has exclusive write access to the disk group. No  
other node can activate the disk group for write access.  
readonly (ro)  
The node has read access to the disk group and denies write  
access for all other nodes in the cluster. The node has no  
write access to the disk group. Attempts to activate a disk  
group for either of the write modes on other nodes fail.  
sharedread (sr)  
The node has read access to the disk group. The node has no  
write access to the disk group, however other nodes can  
obtain write access.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
Administering cluster functionality 403  
Overview of cluster volume management  
Table 13-1  
Activation modes for shared disk groups  
Activation mode  
Description  
sharedwrite (sw)  
The node has write access to the disk group. Attempts to  
activate the disk group for shared read and shared write  
access succeed. Attempts to activate the disk group for  
exclusive write and read-only access fail.  
off  
The node has neither read nor write access to the disk  
group. Query operations on the disk group are permitted.  
The following table summarizes the allowed and conflicting activation modes  
for shared disk groups:  
Table 13-2  
Allowed and conflicting activation modes  
Disk group  
Attempt to activate disk group on another node as...  
activated in cluster  
as...  
exclusive-  
write  
readonly  
sharedread  
sharedwrite  
Fails  
Fails  
Succeeds  
Succeeds  
Succeeds  
Succeeds  
Fails  
exclusivewrite  
readonly  
Fails  
Succeeds  
Succeeds  
Fails  
Fails  
Succeeds  
Fails  
Succeeds  
Succeeds  
sharedread  
sharedwrite  
Shared disk groups can be automatically activated in any mode during disk  
group creation or during manual or auto-import. To control auto-activation of  
shared disk groups, the defaults file /etc/default/vxdgmust be created.  
The defaults file /etc/default/vxdgmust contain the following lines:  
enable_activation=true  
default_activation_mode=activation-mode  
The activation-mode is one of exclusivewrite, readonly, sharedread,  
sharedwrite, or off.  
When a shared disk group is created or imported, it is activated in the specified  
mode. When a node joins the cluster, all shared disk groups accessible from the  
node are activated in the specified mode.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
404 Administering cluster functionality  
Overview of cluster volume management  
Note: The activation mode of a disk group controls volume I/O from different  
nodes in the cluster. It is not possible to activate a disk group on a given node if  
it is activated in a conflicting mode on another node in the cluster. When  
enabling activation using the defaults file, it is recommended that this file be  
made identical on all nodes in the cluster. Otherwise, the results of activation  
are unpredictable.  
If the defaults file is edited while the vxconfigddaemon is already running,  
run the vxconfigd -kcommand on all nodes to restart the process.  
If the default activation mode is anything other than off, an activation  
following a cluster join, or a disk group creation or import can fail if another  
node in the cluster has activated the disk group in a conflicting mode.  
To display the activation mode for a shared disk group, use the vxdglist  
diskgroup command as described in “Listing shared disk groups” on page 421.  
You can also use the vxdgcommand to change the activation mode on a shared  
disk group as described in “Changing the activation mode on a shared disk  
group” on page 425.  
For a description of how to configure a volume so that it can only be opened by a  
single node in a cluster, see “Creating volumes with exclusive open access by a  
node” on page 426 and “Setting exclusive open access to a volume by a node” on  
page 426.  
Connectivity policy of shared disk groups  
A shared disk group provides concurrent read and write access to the volumes  
that it contains for all nodes in a cluster. A shared disk group can only be created  
on the master node. This has the following advantages and implications:  
■
■
■
All nodes in the cluster see exactly the same configuration.  
Only the master node can change the configuration.  
Any changes on the master node are automatically coordinated and  
propagated to the slave nodes in the cluster.  
■
■
Any failures that require a configuration change must be sent to the master  
node so that they can be resolved correctly.  
As the master node resolves failures, all the slave nodes are correctly  
updated. This ensures that all nodes have the same view of the  
configuration.  
The practical implication of this design is that I/O failure on any node results in  
the configuration of all nodes being changed. This is known as the global detach  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Administering cluster functionality 405  
Overview of cluster volume management  
policy. However, in some cases, it is not desirable to have all nodes react in this  
way to I/O failure. To address this, an alternate way of responding to I/O  
failures, known as the local detach policy, was introduced in release 3.2 of VxVM.  
The local detach policy is intended for use with shared mirrored volumes in a  
cluster. This policy prevents I/O failure on a single slave node from causing a  
plex to be detached. This would require the plex to be resynchronized when it is  
subsequently reattached. The local detach policy is available for disk groups  
that have a version number of 70 or greater.  
Note: For small mirrored volumes, non-mirrored volumes, volumes that use  
hardware mirrors, and volumes in private disk groups, there is no benefit in  
configuring the local detach policy. In most cases, it is recommended that you  
use the default global detach policy.  
The detach policies have no effect if the master node loses access to all copies of  
the configuration database and logs in a disk group. If this happened in releases  
prior to 4.1, the master node always disabled the disk group. Release 4.1  
introduces the disk group failure policy, which allows you to change this  
behavior for critical disk groups. This policy is only available for disk groups  
that have a version number of 120 or greater.  
The following sections describe the detach and failure policies in greater detail.  
Global detach policy  
Caution: The global detach policy must be selected when Dynamic MultiPathing  
(DMP) is used to manage multipathing on Active/Passive arrays, This ensures  
that all nodes correctly coordinate their use of the active path.  
The global detach policy is the traditional and default policy for all nodes on the  
configuration. If there is a read or write I/O failure on a slave node, the master  
node performs the usual I/O recovery operations to repair the failure, and, if  
required, the plex is detached cluster-wide. All nodes remain in the cluster and  
continue to perform I/O, but the redundancy of the mirrors is reduced. When  
the problem that caused the I/O failure has been corrected, the mirrors that  
were detached must be recovered before the redundancy of the data can be  
restored.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
406 Administering cluster functionality  
Overview of cluster volume management  
Local detach policy  
Caution: Do not use the local detach policy if you use the VCS agents that  
monitor the cluster functionality of Veritas Volume Manager, and which are  
TM  
provided with Veritas Storage Foundation for Cluster File System HA and  
Veritas Storage Foundation for databases HA. These agents do not notify VCS  
about local failures.  
The local detach policy is designed to support failover applications in large  
clusters where the redundancy of the volume is more important than the  
number of nodes that can access the volume. If there is a write failure on a slave  
node, the master node performs the usual I/O recovery operations to repair the  
failure, and additionally contacts all the nodes to see if the disk is still acceptable  
to them. If the write failure is not seen by all the nodes, I/O is stopped for the  
node that first saw the failure, and the application using the volume is also  
notified about the failure. The volume is not disabled.  
If required, configure the cluster management software to move the application  
to a different node, and/or remove the node that saw the failure from the  
cluster. The volume continues to return write errors, as long as one mirror of  
the volume has an error. The volume continues to satisfy read requests as long  
as one good plex is available.  
If the reason for the I/O error is corrected and the node is still a member of the  
cluster, it can resume performing I/O from/to the volume without affecting the  
redundancy of the data.  
See “Setting the disk detach policy on a shared disk group” on page 425 for  
information on how to use the vxdgcommand to set the disk detach policy on a  
shared disk group.  
The table, “Cluster behavior under I/O failure to a mirrored volume for different  
disk detach policies,” summarizes the effect on a cluster of I/O failure to the  
disks in a mirrored volume:  
Table 13-3  
Cluster behavior under I/O failure to a mirrored volume for different  
disk detach policies  
Type of I/O failure  
Local  
Global  
(diskdetpolicy=local)  
(diskdetpolicy=global)  
Failure of path to one Reads fail only if no plexes  
disk in a volume for a remain available to the  
The plex is detached, and I/O  
from/to the volume continues.  
An I/O error is generated if no  
plexes remain.  
single node  
affected node. Writes to the  
volume fail.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Administering cluster functionality 407  
Overview of cluster volume management  
Table 13-3  
Cluster behavior under I/O failure to a mirrored volume for different  
disk detach policies  
Type of I/O failure  
Local  
Global  
(diskdetpolicy=local)  
(diskdetpolicy=global)  
Failure of paths to all  
disks in a volume for a  
single node  
I/O fails for the affected node. The plex is detached, and I/O  
from/to the volume continues.  
An I/O error is generated if no  
plexes remain.  
Failure of one or more The plex is detached, and I/O  
The plex is detached, and I/O  
disks in a volume for  
all nodes.  
from/to the volume continues. from/to the volume continues.  
An I/O error is generated if no An I/O error is generated if no  
plexes remain.  
plexes remain.  
Disk group failure policy  
The local detach policy by itself is insufficient to determine the desired behavior  
if the master node loses access to all disks that contain copies of the  
configuration database and logs. In this case, the disk group is disabled. As a  
result, the other nodes in the cluster also lose access to the volume. In release  
4.1, the disk group failure policy is introduced to determine the behavior of the  
master node in such cases. This policy has two possible settings as shown in the  
following table:  
Table 13-4  
Behavior of master node for different failure policies  
Type of I/O failure  
Leave  
Disable  
(dgfailpolicy=leave)  
(dgfailpolicy=dgdisable)  
Master node loses  
access to all copies of  
the logs.  
The master node panics with  
the message “klog update  
failed” for a failed  
The master node disables the  
disk group.  
kernel-initiated transaction, or  
“cvm config update failed” for  
a failed user-initiated  
transaction.  
The behavior of the master node under the disk group failure policy is  
independent of the setting of the disk detach policy. If the disk group failure  
policy is set to leave, all nodes panic in the unlikely case that none of them can  
access the log copies.  
See “Setting the disk group failure policy on a shared disk group” on page 426  
for information on how to use the vxdgcommand to set the failure policy on a  
shared disk group.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
408 Administering cluster functionality  
Overview of cluster volume management  
Guidelines for choosing detach and failure policies  
In most cases it is recommended that you use the global detach policy, and  
particularly if any of the following conditions apply:  
■
If you are using the VCS agents that monitor the cluster functionality of  
Veritas Volume Manager, and which are provided with Veritas Storage  
Foundation for Cluster File System HA and Veritas Storage Foundation for  
TM  
databases HA. These agents do not notify VCS about local failures.  
■
■
When an array is seen by DMP as Active/Passive. The local detach policy  
causes unpredictable behavior for Active/Passive arrays.  
For clusters with four or fewer nodes. With a small number of nodes in a  
cluster, it is preferable to keep all nodes actively using the volumes, and to  
keep the applications running on all the nodes, rather than keep the  
redundancy of the volume at the same level.  
■
If only non-mirrored, small mirrored, or hardware mirrored volumes are  
configured. This avoids the system overhead of the extra messaging that is  
required by the local detach policy.  
The local detach policy may be suitable in the following cases:  
■
When large mirrored volumes are configured. Resynchronizing a reattached  
plex can degrade system performance. The local detach policy can avoid the  
need to detach the plex at all. (Alternatively, the dirty region logging (DRL)  
feature can be used to reduce the amount of resynchronization that is  
required.)  
■
For clusters with more than four nodes. Keeping an application running on a  
particular node is less critical when there are many nodes in a cluster. It may  
be possible to configure the cluster management software to move an  
application to a node that has access to the volumes. In addition, load  
balancing may be able to move applications to a different volume from the  
one that experienced the I/O problem. This preserves data redundancy, and  
other nodes may still be able to perform I/O from/to the volumes on the  
disk.  
If you have a critical disk group that you do not want to become disabled in the  
case that the master node loses access to the copies of the logs, set the disk  
group failure policy to leave. This prevents I/O failure on the master node  
disabling the disk group. However, critical applications running on the master  
node fail if they lose access to the other shared disk groups. In such a case, it  
may be preferable to set the policy to dgdisable, and to allow the disk group to  
be disabled.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Administering cluster functionality 409  
Overview of cluster volume management  
The default settings for the detach and failure policies are globaland  
dgdisablerespectively. You can use the vxdgcommand to change both the  
detach and failure policies on a shared disk group, as shown in this example:  
# vxdg -g diskgroup set diskdetpolicy=local dgfailpolicy=leave  
Effect of disk connectivity on cluster reconfiguration  
The detach policy, previous I/O errors, or access to disks are not considered  
when a new master node is chosen. When the master node leaves a cluster, the  
node that takes over as master of the cluster may already have seen I/O failures  
for one or more disks. Under the local detach policy, if a node was affected  
before reconfiguration, and this node then becomes the master, the failure is  
treated as described in “Connectivity policy of shared disk groups” on page 404.  
Some failure scenarios do not result in a disk group failure policy being invoked,  
but can potentially impact the cluster. For example, if the local disk detach  
policy is in effect, and the new master node has a failed plex, this results in all  
nodes detaching the plex because the new master is unaffected by the policy.  
The detach policy does not change the requirement that a node joining a cluster  
must have access to all the disks in all shared disk groups. Similarly, a node that  
is removed from the cluster because of an I/O failure cannot rejoin the cluster  
until this requirement is met.  
Limitations of shared disk groups  
Note: The boot disk group (usually aliased as bootdg) cannot be made  
cluster-shareable. It must be private.  
Only raw device access may be performed via the cluster functionality of VxVM.  
It does not support shared access to file systems in shared volumes unless the  
appropriate software is installed and configured.  
The cluster functionality of VxVM does not support RAID-5 volumes, or task  
monitoring for cluster-shareable disk groups. These features can, however, be  
used in private disk groups that are attached to specific nodes of a cluster.  
If you have RAID-5 volumes in a private disk group that you wish to make  
shareable, you must first relayout the volumes as a supported volume type such  
as stripe-mirroror mirror-stripe. Online relayout of shared volumes is  
supported provided that it does not involve RAID-5 volumes.  
If a shared disk group contains RAID-5 volumes, deport it and then reimport the  
disk group as private on one of the cluster nodes. Reorganize the volumes into  
layouts that are supported for shared disk groups, and then deport and reimport  
the disk group as shared.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
410 Administering cluster functionality  
Cluster initialization and configuration  
Cluster initialization and configuration  
Before any nodes can join a new cluster for the first time, you must supply  
certain configuration information during cluster monitor setup. This  
information is normally stored in some form of cluster monitor configuration  
database. The precise content and format of this information depends on the  
characteristics of the cluster monitor. The information required by VxVM is as  
follows:  
■
■
■
■
cluster ID  
node IDs  
network addresses of nodes  
port addresses  
When a node joins the cluster, this information is automatically loaded into  
VxVM on that node at node startup time.  
Note: To make effective use of the cluster functionality of VxVM requires that  
you configure a cluster monitor (such as provided by HP Serviceguard, or by  
GAB (Group Membership and Atomic Broadcast) in VCS). If HP Serviceguard is  
used as the cluster monitor, no additional configuration of VxVM is required,  
apart from the cluster configuration requirements of HP Serviceguard.  
The cluster monitor startup procedure effects node initialization, and brings up  
the various cluster components (such as VxVM with cluster support, the cluster  
monitor, and a distributed lock manager) on the node. Once this is complete,  
applications may be started. The cluster monitor startup procedure must be  
invoked on each node to be joined to the cluster.  
For VxVM in a cluster environment, initialization consists of loading the cluster  
configuration information and joining the nodes in the cluster. The first node to  
join becomes the master node, and later nodes (slaves) join to the master. If two  
nodes join simultaneously, VxVM chooses the master. Once the join for a given  
node is complete, that node has access to the shared disk groups and volumes.  
Cluster reconfiguration  
Cluster reconfiguration occurs if a node leaves or joins a cluster. Each node’s  
cluster monitor continuously watches the other cluster nodes. When the  
membership of the cluster changes, the cluster monitor informs VxVM for it to  
take appropriate action.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Administering cluster functionality 411  
Cluster initialization and configuration  
During cluster reconfiguration, VxVM suspends I/O to shared disks. I/O resumes  
when the reconfiguration completes. Applications may appear to freeze for a  
short time during reconfiguration.  
If other operations, such as VxVM operations or recoveries, are in progress,  
cluster reconfiguration can be delayed until those operations have completed.  
Volume reconfigurations (see “Volume reconfiguration” on page 413) do not  
take place at the same time as cluster reconfigurations. Depending on the  
circumstances, an operation may be held up and restarted later. In most cases,  
cluster reconfiguration takes precedence. However, if the volume  
reconfiguration is in the commit stage, it completes first.  
For more information on cluster reconfiguration, see “vxclustadm utility” on  
page 411.  
vxclustadm utility  
The vxclustadmcommand provides an interface to the cluster functionality of  
VxVM when VCS or HP Serviceguard is used as the cluster monitor. It is also  
called during cluster startup and shutdown.  
The startnodekeyword to vxclustadmstarts cluster functionality on a cluster  
node by passing cluster configuration information to the VxVM kernel. In  
response to this command, the kernel and the VxVM configuration daemon,  
vxconfigd, perform initialization.  
The stopnodekeyword stops cluster functionality on a node. It waits for all  
outstanding I/O to complete and for all applications to close shared volumes.  
The abortnodekeyword terminates cluster activity on a node. It does not wait  
for outstanding I/O to complete nor for applications to close shared volumes.  
The reinitkeyword allows nodes to be added to or removed from a cluster  
without stopping the cluster. Before running this command, the cluster  
configuration file must have been updated with information about the  
supported nodes in the cluster.  
The nidmapkeyword prints a table showing the mapping between CVM node IDs  
in VxVM’s cluster-support subsystem and node IDs in the cluster monitor. It  
also prints the state of the node in the cluster.  
The nodestatekeyword reports the state of a cluster node and also the reason  
for the last abort of the node as shown in this example:  
# /etc/vx/bin/vxclustadm nodestate  
state: out of cluster  
reason: user initiated stop  
The various reasons that may be given are shown in Table 13-5.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
412 Administering cluster functionality  
Cluster initialization and configuration  
Table 13-5  
Reason  
Node abort messages  
Description  
cannot find disk on slave  
node  
Missing disk or bad disk on the slave node.  
cannot obtain configuration  
data  
The node cannot read the configuration data  
due to an error such as disk failure.  
cluster device open failed  
Open of a cluster device failed.  
clustering license mismatch  
with master node  
Clustering license does not match that on the  
master node.  
clustering license not  
available  
Clustering license cannot be found.  
connection refused by master  
Join of a node refused by the master node.  
disk in use by another  
cluster  
A disk belongs to a cluster other than the one  
that a node is joining.  
join timed out during  
reconfiguration  
Join of a node has timed out due to  
reconfiguration taking place in the cluster.  
klog update failed  
Cannot update kernel log copies during the join  
of a node.  
master aborted during join  
minor number conflict  
Master node aborted while another node was  
joining the cluster.  
Minor number conflicts exist between private  
disk groups and shared disk groups that are  
being imported.  
protocol version out of range Cluster protocol version mismatch or  
unsupported version.  
recovery in progress  
transition to role failed  
user initiated abort  
user initiated stop  
Volumes that were opened by the node are still  
recovering.  
Changing the role of a node to be the master  
failed.  
Node is out of cluster due to an abort initiated  
by the user or by the cluster monitor.  
Node is out of cluster due to a stop initiated by  
the user or by the cluster monitor.  
vxconfigd is not enabled  
The VxVM configuration daemon is not  
enabled.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Administering cluster functionality 413  
Cluster initialization and configuration  
See the vxclustadm(1M) manual page for more information about vxclustadm  
and for examples of its usage.  
Volume reconfiguration  
Volume reconfiguration is the process of creating, changing, and removing  
VxVM objects such as disk groups, volumes and plexes. In a cluster, all nodes  
co-operate to perform such operations. The vxconfigddaemons (see  
“vxconfigd daemon” on page 414) play an active role in volume reconfiguration.  
For reconfiguration to succeed, a vxconfigddaemon must be running on each  
of the nodes.  
A volume reconfiguration transaction is initiated by running a VxVM utility on  
the master node. The utility contacts the local vxconfigddaemon on the  
master node, which validates the requested change. For example, vxconfigd  
rejects an attempt to create a new disk group with the same name as an existing  
disk group. The vxconfigddaemon on the master node then sends details of  
the changes to the vxconfigddaemons on the slave nodes. The vxconfigd  
daemons on the slave nodes then perform their own checking. For example,  
each slave node checks that it does not have a private disk group with the same  
name as the one being created; if the operation involves a new disk, each node  
checks that it can access that disk. When the vxconfigddaemons on all the  
nodes agree that the proposed change is reasonable, each notifies its kernel. The  
kernels then co-operate to either commit or to abandon the transaction. Before  
the transaction can be committed, all of the kernels ensure that no I/O is  
underway, and block any I/O issued by applications until the reconfiguration is  
complete. The master node is responsible both for initiating the reconfiguration,  
and for coordinating the commitment of the transaction. The resulting  
configuration changes appear to occur simultaneously on all nodes.  
If a vxconfigddaemon on any node goes away during reconfiguration, all  
nodes are notified and the operation fails. If any node leaves the cluster, the  
operation fails unless the master has already committed it. If the master node  
leaves the cluster, the new master node, which was previously a slave node,  
completes or fails the operation depending on whether or not it received  
notification of successful completion from the previous master node. This  
notification is performed in such a way that if the new master does not receive  
it, neither does any other slave.  
If a node attempts to join a cluster while a volume reconfiguration is being  
performed, the result of the reconfiguration depends on how far it has  
progressed. If the kernel has not yet been invoked, the volume reconfiguration is  
suspended until the node has joined the cluster. If the kernel has been invoked,  
the node waits until the reconfiguration is complete before joining the cluster.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
414 Administering cluster functionality  
Cluster initialization and configuration  
When an error occurs, such as when a check on a slave fails or a node leaves the  
cluster, the error is returned to the utility and a message is sent to the console  
on the master node to identify on which node the error occurred.  
vxconfigd daemon  
The VxVM configuration daemon, vxconfigd, maintains the configuration of  
VxVM objects. It receives cluster-related instructions from the kernel. A  
separate copy of vxconfigdruns on each node, and these copies communicate  
with each other over a network. When invoked, a VxVM utility communicates  
with the vxconfigddaemon running on the same node; it does not attempt to  
connect with vxconfigddaemons on other nodes. During cluster startup, the  
kernel prompts vxconfigdto begin cluster operation and indicates whether it  
is a master node or a slave node.  
When a node is initialized for cluster operation, the vxconfigddaemon is  
notified that the node is about to join the cluster and is provided with the  
following information from the cluster monitor configuration database:  
■
■
■
■
■
cluster ID  
node IDs  
master node ID  
role of the node  
network address of the vxconfigddaemon on each node (if applicable)  
On the master node, the vxconfigddaemon sets up the shared configuration  
by importing shared disk groups, and informs the kernel when it is ready for the  
slave nodes to join the cluster.  
On slave nodes, the vxconfigddaemon is notified when the slave node can join  
the cluster. When the slave node joins the cluster, the vxconfigddaemon and  
the VxVM kernel communicate with their counterparts on the master node to  
set up the shared configuration.  
When a node leaves the cluster, the kernel notifies the vxconfigddaemon on all  
the other nodes. The master node then performs any necessary cleanup. If the  
master node leaves the cluster, the kernels select a new master node and the  
vxconfigddaemons on all nodes are notified of the choice.  
The vxconfigddaemon also participates in volume reconfiguration as  
described in “Volume reconfiguration” on page 413.  
vxconfigd daemon recovery  
In a cluster, the vxconfigddaemons on the slave nodes are always connected  
to the vxconfigddaemon on the master node. If the vxconfigddaemon is  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Administering cluster functionality 415  
Cluster initialization and configuration  
stopped, volume reconfiguration cannot take place. Other nodes can join the  
cluster if the vxconfigddaemon is not running on the slave nodes.  
If the vxconfigddaemon stops, different actions are taken depending on which  
node this occurred:  
■
If the vxconfigddaemon is stopped on the master node, the vxconfigd  
daemons on the slave nodes periodically attempt to rejoin to the master  
node. Such attempts do not succeed until the vxconfigddaemon is  
restarted on the master. In this case, the vxconfigddaemons on the slave  
nodes have not lost information about the shared configuration, so that any  
displayed configuration information is correct.  
■
If the vxconfigddaemon is stopped on a slave node, the master node takes  
no action. When the vxconfigddaemon is restarted on the slave, the slave  
vxconfigddaemon attempts to reconnect to the master daemon and to  
re-acquire the information about the shared configuration. (Neither the  
kernel view of the shared configuration nor access to shared disks is  
affected.) Until the vxconfigddaemon on the slave node has successfully  
reconnected to the vxconfigddaemon on the master node, it has very little  
information about the shared configuration and any attempts to display or  
modify the shared configuration can fail. For example, shared disk groups  
listed using the vxdglistcommand are marked as disabled; when the  
rejoin completes successfully, they are marked as enabled.  
■
If the vxconfigddaemon is stopped on both the master and slave nodes,  
the slave nodes do not display accurate configuration information until  
vxconfigdis restarted on the master and slave nodes, and the daemons  
have reconnected.  
If the CVM agent for VCS determines that the vxconfigddaemon is not  
running on a node during a cluster reconfiguration, vxconfigdis restarted  
automatically.  
If it is necessary to restart vxconfigdmanually in a VCS controlled cluster to  
resolve a VxVM issue, use this procedure:  
1
2
3
Use the following command to disable failover on any service groups that  
contain VxVM objects:  
# hagrp -freeze group  
Enter the following command to stop and restart the VxVM configuration  
daemon on the affected node:  
# vxconfigd -k  
Use the following command to re-enable failover for the service groups that  
you froze in step 1:  
# hagrp -unfreeze group  
Download from Www.Somanuals.com. All Manuals Search And Download.  
416 Administering cluster functionality  
Cluster initialization and configuration  
Note: The -rreset option to vxconfigdrestarts the vxconfigddaemon and  
recreates all states from scratch. This option cannot be used to restart  
vxconfigdwhile a node is joined to a cluster because it causes cluster  
information to be discarded.  
In an HP Serviceguard cluster, use the equivalent Serviceguard functionality to  
stop and restart the appropriate package.  
Node shutdown  
Although it is possible to shut down the cluster on a node by invoking the  
shutdown procedure of the node’s cluster monitor, this procedure is intended  
for terminating cluster components after stopping any applications on the node  
that have access to shared storage. VxVM supports clean node shutdown, which  
allows a node to leave the cluster gracefully when all access to shared volumes  
has ceased. The host is still operational, but cluster applications cannot be run  
on it.  
The cluster functionality of VxVM maintains global state information for each  
volume. This enables VxVM to determine which volumes need to be recovered  
when a node crashes. When a node leaves the cluster due to a crash or by some  
other means that is not clean, VxVM determines which volumes may have  
writes that have not completed and the master node resynchronizes these  
volumes. It can use dirty region logging (DRL) or FastResync if these are active  
for any of the volumes.  
Clean node shutdown must be used after, or in conjunction with, a procedure to  
halt all cluster applications. Depending on the characteristics of the clustered  
application and its shutdown procedure, a successful shutdown can require a lot  
of time (minutes to hours). For instance, many applications have the concept of  
draining, where they accept no new work, but complete any work in progress  
before exiting. This process can take a long time if, for example, a long-running  
transaction is active.  
When the VxVM shutdown procedure is invoked, it checks all volumes in all  
shared disk groups on the node that is being shut down. The procedure then  
either continues with the shutdown, or fails for one of the following reasons:  
■
If all volumes in shared disk groups are closed, VxVM makes them  
unavailable to applications. Because all nodes are informed that these  
volumes are closed on the leaving node, no resynchronization is performed.  
■
If any volume in a shared disk group is open, the shutdown operation in the  
kernel waits until the volume is closed. There is no timeout checking in this  
operation.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Administering cluster functionality 417  
Multiple host failover configurations  
Note: Once shutdown succeeds, the node has left the cluster. It is not possible to  
access the shared volumes until the node joins the cluster again.  
Since shutdown can be a lengthy process, other reconfiguration can take place  
while shutdown is in progress. Normally, the shutdown attempt is suspended  
until the other reconfiguration completes. However, if it is already too far  
advanced, the shutdown may complete first.  
Node abort  
If a node does not leave a cluster cleanly, this is because it crashed or because  
some cluster component made the node leave on an emergency basis. The  
ensuing cluster reconfiguration calls the VxVM abort function. This procedure  
immediately attempts to halt all access to shared volumes, although it does wait  
until pending I/O from or to the disk completes.  
I/O operations that have not yet been started are failed, and the shared volumes  
are removed. Applications that were accessing the shared volumes therefore fail  
with errors.  
After a node abort or crash, shared volumes must be recovered, either by a  
surviving node or by a subsequent cluster restart, because it is very likely that  
there are unsynchronized mirrors.  
Cluster shutdown  
If all nodes leave a cluster, shared volumes must be recovered when the cluster  
is next started if the last node did not leave cleanly, or if resynchronization from  
previous nodes leaving uncleanly is incomplete.  
Multiple host failover configurations  
Outside the context of clustering functionality, VxVM disk groups can be  
“imported” (made available) from only one host at any given time. When a host  
imports a disk group as private, the volumes and configuration of that disk  
group become accessible to the host. If the administrator or system software  
wants to privately use the same disk group from another host, the host that  
already has the disk group imported (importing host) must “deport” (give up  
access to) the disk group. Once deported, the disk group can be imported by  
another host.  
If two hosts are allowed to access a disk group concurrently without proper  
synchronization, such as that provided by the Oracle Parallel Server, the  
configuration of the disk group, and possibly the contents of volumes, can be  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
418 Administering cluster functionality  
Multiple host failover configurations  
corrupted. Similar corruption can also occur if a file system or database on a raw  
disk partition is accessed concurrently by two hosts, so this problem in not  
limited to Veritas Volume Manager.  
Import lock  
When a host in a non-clustered environment imports a disk group, an import  
lock is written on all disks in that disk group. The import lock is cleared when  
the host deports the disk group. The presence of the import lock prevents other  
hosts from importing the disk group until the importing host has deported the  
disk group.  
Specifically, when a host imports a disk group, the import normally fails if any  
disks within the disk group appear to be locked by another host. This allows  
automatic re-importing of disk groups after a reboot (autoimporting) and  
prevents imports by another host, even while the first host is shut down. If the  
importing host is shut down without deporting the disk group, the disk group  
can only be imported by another host by clearing the host ID lock first  
(discussed later).  
The import lock contains a host ID (in Veritas Volume Manager, this is the host  
name) reference to identify the importing host and enforce the lock. Problems  
can therefore arise if two hosts have the same host ID.  
Note: Since Veritas Volume Manager uses the host name as the host ID (by  
default), it is advisable to change the host name of one machine if another  
machine shares its host name. To change the host name, use the vxdctlhostid  
new_hostname command.  
Failover  
The import locking scheme works well in an environment where disk groups are  
not normally shifted from one system to another. However, consider a setup  
where two hosts, Node A and Node B, can access the drives of a disk group. The  
disk group is first imported by Node A, but the administrator wants to access the  
disk group from Node B if Node A crashes. This kind of scenario (failover) can be  
used to provide manual high availability to data, where the failure of one node  
does not prevent access to data. Failover can be combined with a “high  
availability” monitor to provide automatic high availability to data: when Node  
B detects that Node A has crashed or shut down, Node B imports (fails over) the  
disk group to provide access to the volumes.  
Veritas Volume Manager can support failover, but it relies on the administrator  
or on an external high-availability monitor to ensure that the first system is  
shut down or unavailable before the disk group is imported to another system.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Administering cluster functionality 419  
Multiple host failover configurations  
For details on how to clear locks and force an import, see “Moving disk groups  
between systems” on page 185 and the vxdg(1M) manual page.  
Corruption of disk group configuration  
If vxdgimportis used with -C(clears locks) and/or -f(forces import) to  
import a disk group that is still in use from another host, disk group  
configuration corruption is likely to occur. Volume content corruption is also  
likely if a file system or database is started on the imported volumes before the  
other host crashes or shuts down.  
If this kind of corruption occurs, you must probably rebuild your configuration  
from scratch and reload all volumes in the disk group from a backup. To backup  
and rebuild the configuration, if nothing has changed, use vxprint -mspvdand  
store the output which can be fed to vxmaketo restore the layouts. There are  
typically numerous configuration copies for each disk group, but corruption  
nearly always affects all configuration copies, so redundancy does not help in  
this case.  
Disk group configuration corruption usually shows up as missing or duplicate  
records in the configuration databases. This can result in a variety of  
vxconfigderror messages  
VxVM vxconfigd ERROR V-5-1-569 Disk group group,Disk disk:Cannot  
auto-import group: reason  
where the reason can describe errors such as:  
Association not resolved  
Association count is incorrect  
Duplicate record in configuration  
Configuration records are inconsistent  
These errors are typically reported in association with specific disk group  
configuration copies, but usually apply to all copies. The following is usually  
displayed along with the error:  
Disk group has no valid configuration copies  
See the Veritas Volume Manager Troubleshooting Guide for more information on  
Veritas Volume Manager error messages.  
If you use the Veritas Cluster Server product, all disk group failover issues can  
be managed correctly. VCS includes a high availability monitor and includes  
failover scripts for VxVM, VxFS, and for several popular databases.  
The -toption to vxdgprevents automatic re-imports on reboot and is  
necessary when used with a host monitor (such as VCS) that controls imports  
itself, rather than relying on automatic imports by Veritas Volume Manager.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
               
420 Administering cluster functionality  
Administering VxVM in cluster environments  
Administering VxVM in cluster environments  
The following sections describe the administration of VxVM’s cluster  
functionality.  
Note: Most VxVM commands require superuser or equivalent privileges.  
Requesting node status and discovering the master node  
The vxdctlutility controls the operation of the vxconfigdvolume  
configuration daemon. The -coption can be used to request cluster information  
and to find out which node is the master. To determine whether the vxconfigd  
daemon is enabled and/or running, use the following command:  
# vxdctl -c mode  
This produces different output messages depending on the current status of the  
cluster node:  
Table 13-6  
Cluster status messages  
Status message  
Description  
mode: enabled: cluster active - MASTER  
master: mozart  
The node is the master.  
mode: enabled: cluster active - SLAVE  
master: mozart  
The node is a slave.  
mode: enabled: cluster active - role not set The node has not yet been  
master: mozart  
assigned a role, and is in the  
process of joining the  
cluster.  
state: joining  
reconfig: master update  
mode: enabled: cluster active - SLAVE  
master: mozart  
The node is configured as a  
slave, and is in the process  
of joining the cluster.  
state: joining  
mode: enabled: cluster inactive  
The cluster is not active on  
this node.  
Note: If the vxconfigddaemon is disabled, no cluster information is displayed.  
See the vxdctl(1M) manual page for more information.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Administering cluster functionality 421  
Administering VxVM in cluster environments  
Determining if a disk is shareable  
The vxdiskutility manages VxVM disks. To use the vxdiskutility to  
determine whether a disk is part of a cluster-shareable disk group, use the  
following command:  
# vxdisk list accessname  
where accessname is the disk access name (or device name). A portion of the  
output from this command (for the device c4t1d0) is shown here:  
Device:  
devicetag: c4t1d0  
type: auto  
clusterid: cvm2  
c4t1d0  
disk:  
name=shdg01 id=963616090.1034.cvm2  
timeout:  
group:  
flags:  
...  
30  
name=shdg id=963616065.1032.cvm2  
online ready autoconfig shared imported  
Note that the clusteridfield is set to cvm2(the name of the cluster), and the  
flagsfield includes an entry for shared. When a node is not joined to the  
cluster, the flagsfield contains the autoimportflag instead of imported.  
Listing shared disk groups  
vxdgcan be used to list information about shared disk groups. To display  
information for all disk groups, use the following command:  
# vxdg list  
Example output from this command is displayed here:  
NAME  
STATE  
ID  
rootdg  
group2  
group1  
enabled  
enabled,shared  
enabled,shared  
774215886.1025.teal  
774575420.1170.teal  
774222028.1090.teal  
Shared disk groups are designated with the flag shared.  
To display information for shared disk groups only, use the following command:  
# vxdg -s list  
Example output from this command is as follows:  
NAME  
STATE  
ID  
group2  
group1  
enabled,shared  
enabled,shared  
774575420.1170.teal  
774222028.1090.teal  
To display information about one specific disk group, use the following  
command:  
# vxdg list diskgroup  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
422 Administering cluster functionality  
Administering VxVM in cluster environments  
The following is example output for the command vxdglistgroup1on the  
master:  
Group:  
dgid:  
group1  
774222028.1090.teal  
import-id: 32768.1749  
flags:  
version:  
shared  
140  
alignment: 8192 (bytes)  
ssb:  
on  
local-activation: exclusive-write  
cluster-actv-modes: node0=ew node1=off  
detach-policy: local  
private_region_failure: leave  
copies:  
nconfig=2 nlog=2  
config:  
seqno=0.1976 permlen=1456 free=1448 templen=6  
loglen=220  
config disk c1t0d0 copy 1 len=1456 state=clean online  
config disk c1t0d0 copy 1 len=1456 state=clean online  
log disk c1t0d0 copy 1 len=220  
log disk c1t0d0 copy 1 len=220  
Note that the flagsfield is set to shared. The output for the same command  
when run on a slave is slightly different. The local-activationand  
cluster-actv-modesfields display the activation mode for this node and for  
each node in the cluster respectively. The detach-policyand  
private_region_failurefields indicate how the cluster behaves in the  
event of loss of connectivity to the disks, and to the configuration and log copies  
on the disks.  
Creating a shared disk group  
Note: Shared disk groups can only be created on the master node.  
If the cluster software has been run to set up the cluster, a shared disk group can  
be created using the following command:  
# vxdg -s init diskgroup [diskname=]devicename  
where diskgroup is the disk group name, diskname is the administrative name  
chosen for a VM disk, and devicename is the device name (or disk access name).  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Administering cluster functionality 423  
Administering VxVM in cluster environments  
Caution: The operating system cannot tell if a disk is shared. To protect data  
integrity when dealing with disks that can be accessed by multiple systems, use  
the correct designation when adding a disk to a disk group. VxVM allows you to  
add a disk that is not physically shared to a shared disk group if the node where  
the disk is accessible is the only node in the cluster. However, this means that  
other nodes cannot join the cluster. Furthermore, if you attempt to add the same  
disk to different disk groups (private or shared) on two nodes at the same time,  
the results are undefined. Perform all configuration on one node only, and  
preferably on the master node.  
Importing disk groups as shared  
Note: Shared disk groups can only be imported on the master node.  
Disk groups can be imported as shared using the vxdg-simportcommand. If  
the disk groups are set up before the cluster software is run, the disk groups can  
be imported into the cluster arrangement using the following command:  
# vxdg -s import diskgroup  
where diskgroup is the disk group name or ID. On subsequent cluster restarts,  
the disk group is automatically imported as shared. Note that it can be necessary  
to deport the disk group (using the vxdgdeportdiskgroup command) before  
invoking the vxdgutility.  
Forcibly importing a disk group  
You can use the -foption to the vxdgcommand to import a disk group forcibly.  
Caution: The force option(-f) must be used with caution and only if you are fully  
aware of the consequences such as possible data corruption.  
When a cluster is restarted, VxVM can refuse to auto-import a disk group for  
one of the following reasons:  
■
A disk in the disk group is no longer accessible because of hardware errors  
on the disk. In this case, use the following command to forcibly reimport the  
disk group:  
# vxdg -s -f import diskgroup  
Note: After a forced import, the data on the volumes may not be available  
and some of the volumes may be in the disabled state.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
424 Administering cluster functionality  
Administering VxVM in cluster environments  
■
Some of the nodes to which disks in the disk group are attached are not  
currently in the cluster, so the disk group cannot access all of its disks. In  
this case, a forced import is unsafe and must not be attempted because it can  
result in inconsistent mirrors.  
Converting a disk group from shared to private  
Note: Shared disk groups can only be deported on the master node.  
To convert a shared disk group to a private disk group, first deport it on the  
master node using this command:  
# vxdg deport diskgroup  
Then reimport the disk group on any cluster node using this command:  
# vxdg import diskgroup  
Moving objects between disk groups  
As described in “Moving objects between disk groups” on page 203, you can use  
the vxdgmovecommand to move a self-contained set of VxVM objects such as  
disks and top-level volumes between disk groups. In a cluster, you can move  
such objects between private disk groups on any cluster node where those disk  
groups are imported.  
Note: You can only move objects between shared disk groups on the master  
node. You cannot move objects between private and shared disk groups.  
Splitting disk groups  
As described in “Splitting disk groups” on page 205, you can use the vxdgsplit  
command to remove a self-contained set of VxVM objects from an imported disk  
group, and move them to a newly created disk group.  
Splitting a private disk group creates a private disk group, and splitting a shared  
disk group creates a shared disk group. You can split a private disk group on any  
cluster node where that disk group is imported. You can only split a shared disk  
group or create a shared target disk group on the master node.  
See “Moving objects between disk groups” on page 203.  
Joining disk groups  
As described in “Joining disk groups” on page 206, you can use the vxdgjoin  
command to merge the contents of two imported disk groups. In a cluster, you  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                       
Administering cluster functionality 425  
Administering VxVM in cluster environments  
can join two private disk groups on any cluster node where those disk groups are  
imported.  
If the source disk group and the target disk group are both shared, you must  
perform the join on the master node.  
Note: You cannot join a private disk group and a shared disk group.  
Changing the activation mode on a shared disk group  
Note: The activation mode for access by a cluster node to a shared disk group is  
set on that node.  
The activation mode of a shared disk group can be changed using the following  
command:  
# vxdg -g diskgroup set activation=mode  
The activation mode is one of exclusivewriteor ew, readonlyor ro,  
sharedreador sr, sharedwriteor sw, or off.  
If you use this command to change the activation mode of a shared disk group,  
you must first change the activation mode to offbefore setting it to any other  
value, as shown here:  
# vxdg -g myshdg set activation=off  
# vxdg -g myshdg set activation=readonly  
See “Activation modes of shared disk groups” on page 402.  
Setting the disk detach policy on a shared disk group  
Note: The disk detach policy for a shared disk group can only be set on the  
master node.  
The vxdgcommand may be used to set either the globalor localdisk detach  
policy for a shared disk group:  
# vxdg -g diskgroup set diskdetpolicy=global|local  
The default disk detach policy is global.  
See “Connectivity policy of shared disk groups” on page 404.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
426 Administering cluster functionality  
Administering VxVM in cluster environments  
Setting the disk group failure policy on a shared disk group  
Note: The disk group failure policy for a shared disk group can only be set on the  
master node.  
The vxdgcommand may be used to set either the dgdisableor leavefailure  
policy for a shared disk group:  
# vxdg -g diskgroup set dgfailpolicy=dgdisable|leave  
The default failure policy is dgdisable.  
See “Disk group failure policy” on page 407.  
Creating volumes with exclusive open access by a node  
Note: All shared volumes, including those with exclusive open access, can only  
be created on the master node.  
When using the vxassistcommand to create a volume, you can use the  
exclusive=onattribute to specify that the volume may only be opened by one  
node in the cluster at a time. For example, to create the mirrored volume  
volmirin the disk group dskgrp, and configure it for exclusive open, use the  
following command:  
# vxassist -g dskgrp make volmir 5g layout=mirror exclusive=on  
Multiple opens by the same node are also supported. Any attempts by other  
nodes to open the volume fail until the final close of the volume by the node that  
opened it.  
Specifying exclusive=offinstead means that more than one node in a cluster  
can open a volume simultaneously. This is the default behavior.  
Setting exclusive open access to a volume by a node  
Note: Exclusive open access on a volume can only be set on the master node.  
Ensure that none of the nodes in the cluster have the volume open when setting  
this attribute.  
You can set the exclusive=onattribute with the vxvol command to specify that  
an existing volume may only be opened by one node in the cluster at a time.  
For example, to set exclusive open on the volume volmirin the disk group  
dskgrp, use the following command:  
# vxvol -g dskgrp set exclusive=on volmir  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                 
Administering cluster functionality 427  
Administering VxVM in cluster environments  
Multiple opens by the same node are also supported. Any attempts by other  
nodes to open the volume fail until the final close of the volume by the node that  
opened it.  
Specifying exclusive=offinstead means that more than one node in a cluster  
can open a volume simultaneously. This is the default behavior.  
Displaying the cluster protocol version  
The following command displays the cluster protocol version running on a node:  
# vxdctl list  
This command produces output similar to the following:  
Volboot file  
version: 3/1  
seqno: 0.19  
cluster protocol version: 70  
hostid: giga  
entries:  
You can also check the existing cluster protocol version using the following  
command:  
# vxdctl protocolversion  
This produces output similar to the following:  
Cluster running at protocol 70  
Displaying the supported cluster protocol version range  
The following command displays the maximum and minimum protocol version  
supported by the node and the current protocol version:  
# vxdctl support  
This command produces out put similar to the following:  
Support information:  
vxconfigd_vrsn:  
dg_minimum:  
dg_maximum:  
21  
20  
140  
15  
40  
70  
70  
kernel:  
protocol_minimum:  
protocol_maximum:  
protocol_current:  
You can also use the following command to display the maximum and minimum  
cluster protocol version supported by the current Veritas Volume Manager  
release:  
# vxdctl protocolrange  
This produces output similar to the following:  
minprotoversion: 40, maxprotoversion: 70  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
428 Administering cluster functionality  
Administering VxVM in cluster environments  
Upgrading the cluster protocol version  
Note: The cluster protocol version can only be updated on the master node.  
After all the nodes in the cluster have been updated with a new cluster protocol,  
you can upgrade the entire cluster using the following command on the master  
node:  
# vxdctl upgrade  
Recovering volumes in shared disk groups  
Note: Volumes can only be recovered on the master node.  
The vxrecoverutility is used to recover plexes and volumes after disk  
replacement. When a node leaves a cluster, it can leave some mirrors in an  
inconsistent state. The vxrecoverutility can be used to recover such volumes.  
The -coption to vxrecovercauses it to recover all volumes in shared disk  
groups. The vxconfigddaemon automatically calls the vxrecoverutility with  
the -coption when necessary.  
Note: While the vxrecoverutility is active, there can be some degradation in  
system performance.  
Obtaining cluster performance statistics  
The vxstatutility returns statistics for specified objects. In a cluster  
environment, vxstatgathers statistics from all of the nodes in the cluster. The  
statistics give the total usage, by all nodes, for the requested objects. If a local  
object is specified, its local usage is returned.  
You can optionally specify a subset of nodes using the following form of the  
command:  
# vxstat -g diskgroup -n node[,node...]  
where node is the CVM node ID number. You can find out the CVM node ID by  
using the following command:  
# vxclustadm nidmap  
If a comma-separated list of nodes is supplied, the vxstatutility displays the  
sum of the statistics for the nodes in the list.  
For example, to obtain statistics for node 2, volume vol1,use the following  
command:  
# vxstat -g group1 -n 2 vol1  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Administering cluster functionality 429  
Administering VxVM in cluster environments  
This command produces output similar to the following:  
OPERATIONS  
TYP NAME READ  
vol vol1 2421  
BLOCKS  
WRITE READ  
AVG TIME(ms)  
WRITE READ  
0 99.0  
WRITE  
0.0  
0
600000  
To obtain and display statistics for the entire cluster, use the following  
command:  
# vxstat -b  
The statistics for all nodes are summed. For example, if node 1 performed 100  
I/O operations and node 2 performed 200 I/O operations, vxstat-bdisplays a  
total of 300 I/O operations.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
430 Administering cluster functionality  
Administering VxVM in cluster environments  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter  
14  
Administering  
sites and remote mirrors  
In a Remote Mirror configuration (also known as a campus cluster or stretch  
cluster) the hosts and storage of a cluster that would usually be located in one  
place, are instead divided between two or more sites. These sites are typically  
connected via a redundant high-capacity network that provides access to  
storage and private link communication between the cluster nodes. A typical  
two-site remote mirror configuration is illustrated in Figure 14-1.  
Figure 14-1  
Example of a two-site remote mirror configuration  
Private Network  
Site A  
Site B  
Metropolitan or  
Wide Area  
Network Link  
(Fibre Channel or  
DWDM)  
Cluster Nodes  
Cluster Nodes  
Fibre Channel  
Hubs/Switches  
Disk  
Enclosures  
Disk  
Enclosures  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
432 Administering sites and remote mirrors  
If a disk group is configured across the storage at the sites, and inter-site  
communication is disrupted, there is a possibility of a serial split brain condition  
arising if each site continues to update the local disk group configuration copies  
(see “Handling conflicting configuration copies” on page 190). VxVM provides  
mechanisms for dealing with the serial split brain condition, monitoring the  
health of a remote mirror, and testing the robustness of the cluster against  
various types of failure (also known as fire drill).  
For applications and services to function correctly at a site when other sites  
have become inaccessible, at least one complete plex of each volume must be  
configured at each site (site-based allocation), and the consistency of the data in  
the plexes at each site must be ensured (site consistency).  
By tagging disks with site names, storage can be allocated from the correct  
location when creating, resizing or relocating a volume, and when changing a  
volume’s layout.  
Figure 14-2 shows an example of a site-consistent volume with two plexes  
configured at each of two sites. The storage for plexes P1 and P2 is allocated  
storage that is tagged as belonging to site A, and the storage for plexes P3 and P4  
is allocated storage that is tagged as belonging to site B.  
Figure 14-2  
Site-consistent volume with two plexes at each of two sites  
Site A  
Site B  
Disk Group  
Volume V  
Plex P1  
Plex P2  
Plex P3  
Plex P4  
Although not shown in this figure, DCO log volumes are also mirrored across the  
sites, and disk group configuration copies are distributed across the sites.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Administering sites and remote mirrors 433  
To enhance read performance, VxVM will service reads from the plexes at the  
local site where an application is running if the sitereadread policy is set on a  
volume. Writes are written to plexes at all sites.  
The site consistency of a volume is ensured by detaching a site when its last  
complete plex fails at that site. If a site fails, all its plexes are detached and the  
site is said to be detached.  
Configurations with remote storage only are also supported as illustrated in  
Figure 14-3.  
Figure 14-3  
Example of a two-site configuration with remote storage only  
Site A  
Site B  
Cluster or  
Standalone  
System  
Metropolitan  
Area Network Link  
(such as DWDM)  
Fibre Channel  
Hubs/Switches  
Disk  
Disk  
Enclosures  
Enclosures  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
434 Administering sites and remote mirrors  
Configuring sites for hosts and disks  
Configuring sites for hosts and disks  
Note: The Remote Mirror feature requires that the Site Awareness license has  
been installed on all hosts at all sites that are participating in the configuration.  
Use the following command to set the site name for each host:  
# vxdctl set site=sitename  
The name that has been assigned to a site is stored in the /etc/vx/volboot  
file, and can be displayed by using the vxdctllistcommand:  
# vxdctl list | grep siteid  
siteid: building1  
To remove the site name from a host, use this command:  
# vxdctl [-F] unset site  
The -Foption is required if any imported disk groups are registered to the site.  
To tag disks with a site name, use the vxdisksettagcommand as shown here:  
# vxdisk [-g diskgroup] settag disk site=sitename  
where the disk can be specified either by the disk access name or the disk media  
name. You must repeat this command for each of the disks that are to be  
registered to a site.  
To check which disks are registered to a site, use the vxdisklisttagcommand:  
# vxdisk [-g diskgroup] listtag  
To remove the site tag from a disk, use the vxdiskrmtagcommand:  
# vxdisk rmtag disk site=sitename  
Configuring site-based allocation on a disk group  
To turn on the site-based allocation requirement for a site that is registered to a  
disk group, use the vxdgaddsitecommand for each site at which site-based  
allocation is required:  
# vxdg -g diskgroup [-f] addsite sitename  
Each volume on which the allsitesattribute is set to onis checked to ensure  
that it has at least one plex at each site, and the command fails if this condition  
is not met. If the -foption is specified, the command does not fail, but instead it  
sets the allsitesattribute for the volume to off.  
Provided that the Site Awareness license is installed on all the hosts in the  
Remote Mirror configuration, a volume is automatically mirrored across sites  
and its read policy is set to sitereadwhen it is created. If site-based allocation is  
not required, or is not possible (as is the case for RAID-5 volumes), specify the  
allsites=offattribute to the vxassistcommand.  
To remove the site-based allocation requirement from a site, use this command:  
# vxdg -g diskgroup [-f] rmsite sitename  
Download from Www.Somanuals.com. All Manuals Search And Download.  
               
Administering sites and remote mirrors 435  
Configuring site consistency on a disk group  
The -foption allows the requirement to be removed if the site is detached or  
offline.  
The site name is not removed from the disks. If required, use the vxdiskrmtag  
command to remove the site tag as described in “Configuring sites for hosts and  
disks” on page 434.  
Configuring site consistency on a disk group  
Having configured site-based allocation on a disk group as described in  
“Configuring site-based allocation on a disk group” on page 434, turn on the site  
consistency requirement for a disk group by using the vxdgcommand:  
# vxdg -g diskgroup set siteconsistent=on  
Note: This command does not set the siteconsistentattribute on existing  
volumes in the disk group. Although newly created volumes inherit the setting  
from the disk group, it does not change when the disk group setting is changed.  
See “Configuring site consistency on a volume” on page 435.  
All the disks in a disk group must be registered to one of the sites before you can  
set the siteconsistentattribute on the disk group.  
To verify whether site consistency has been enabled for a disk group, use the  
following command:  
# vxdg list diskgroup | grep siteconsistent  
flags: siteconsistent  
To turn off the site consistency requirement for a disk group, use the following  
command:  
# vxdg -g diskgroup set siteconsistent=off  
Configuring site consistency on a volume  
To set the site consistency requirement when creating a volume, specify the  
siteconsistentattribute to the vxassistmakecommand, for example:  
# vxassist [-g diskgroup] make volume size \  
nmirror=4 siteconsistent={on|off}  
By default, a volume inherits the value that is set on its disk group.  
Note: By default, creating a site-consistent volume also creates an associated  
version 20 DCO volume, and enables Persistent FastResync on the volume. This  
allows faster recovery of the volume during the reattachment of a site.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
436 Administering sites and remote mirrors  
Setting the siteread policy on a volume  
To turn on the site consistency requirement for an existing volume, use the  
following form of the vxvolcommand:  
# vxvol [-g diskgroup] set siteconsistent=on volume  
To turn off the site consistency requirement for a volume, use the following  
command:  
# vxvol [-g diskgroup] set siteconsistent=off volume  
Note: The siteconsistentand allsitesattributes must be set to offfor  
RAID-5 volumes in a site-consistent disk group.  
Setting the siteread policy on a volume  
If the Site Awareness license is installed on all the hosts in the Remote Mirror  
configuration, the disk group is configured for site consistency with several  
sites enabled, and the allsites=onattribute is specified for a volume, the  
default read policy is siteread.  
If required, you can use the following command to set the sitereadpolicy on a  
volume:  
# vxvol [-g diskgroup] rdpol siteread volume  
This command has no effect if a site name has not been set for the host.  
See “Changing the read policy for mirrored volumes” on page 289.  
Site-based allocation of storage to volumes  
The vxassistcommand can be used to create volumes only from storage that  
exists at a specified site, as shown in this example:  
# vxassist -g diskgroup make volume size site:site1 \  
[allsites={on|off}] [siteconsistent={on|off}]  
The storage class siteis used in similar way to other storage classes with the  
vxassistcommand, such as enclr, ctlrand disk.  
See “Mirroring across targets, controllers or enclosures” on page 255.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
Administering sites and remote mirrors 437  
Site-based allocation of storage to volumes  
Note: If the Site Awareness license is installed on all the hosts in the Remote  
Mirror configuration, and site consistency is enabled on a volume, the vxassist  
command attempts to allocate storage across the sites that are registered to a  
disk group. If not enough storage is available at all sites, the command fails  
unless you also specify the allsites=offattribute.  
By default, the allsitesattribute is set to onfor volume in a site-consistent disk  
group. The allsitesand siteconsistentattributes must be set to offfor  
RAID-5 volumes in a site-consistent disk group.  
In a similar way to mirroring across controllers, you can also ensure that plexes  
are created at all sites that are registered for a disk group:  
# vxassist -g diskgroup make volume size mirror=site  
The allsitesand siteconsistentattributes can be combined to create a  
non-site-consistent mirrored volume with plexes only at some of the sites:  
# vxassist -g diskgroup make volume size mirror=site \  
site:site1 site:site2 ... allsites=off siteconsistent=off  
a non-site-consistent mirrored volume with plexes at all of the sites:  
# vxassist -g diskgroup make volume size mirror=site \  
allsites=on siteconsistent=off  
a site-consistent mirrored volume with plexes only at some of the sites:  
# vxassist -g diskgroup make volume size mirror=site \  
site:site1 site:site2 ... allsites=off siteconsistent=on  
or a site-consistent mirrored volume with plexes at all of the sites:  
# vxassist -g diskgroup make volume size mirror=site \  
allsites=on siteconsistent=on  
By default, the number of mirrors created is equal to the number of sites that are  
configured in the disk group. There should be a minimum of 2 plexes at each  
site. To override the default behavior, use the nmirrorattribute to specify the  
total number of mirrors:  
# vxassist -g diskgroup make volume size mirror=site \  
nmirror=4 site:site1 site:site2 ... [allsites={on|off}] \  
[siteconsistent={on|off}]  
If a volume is intended to be site consistent, the number of plexes that are  
specified must be a multiple of the number of sites.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
438 Administering sites and remote mirrors  
Site-based allocation of storage to volumes  
Examples of storage allocation using sites  
The examples in the following table demonstrate how to use site names with the  
vxassistcommand to allocate storage. The disk group, ccdg, has been enabled  
for site consistency with disks configured at two sites, site1and site2.  
Command  
Description  
# vxassist -g ccdg make vol 2g \ Create a volume with one mirror at each  
nmirror=2  
site. The value of nmirrormust be greater  
than or equal to the number of sites unless  
allsitesis set to off.  
# vxassist -g ccdg -o ordered \  
make vol 2g \  
Create a mirrored-stripe volume specifying  
allocation order to validate redundancy  
across the sites. The named disks must be  
tagged with the appropriate site name, and  
there must be sufficient disks at each site to  
create the volume.  
layout=mirror-stripe ncol=3 \  
ccdg01 ccdg02 ccdg03 ccdg09 \  
ccdg10 ccdg11  
# vxassist -g ccdg make vol 2g \ Create a volume with one mirror at each of  
nmirror=2 mirror=site \  
site:site1 site:site2  
the named sites. All sites must be named  
and the value of nmirrormust be greater  
than or equal to the number of sites unless  
allsitesis set to off.  
# vxassist -g ccdg make vol 2g \ Create a volume with one mirror on each of  
nmirror=2 ccdg01 ccdg09  
the named disks. The named disks must be  
tagged with the appropriate site name, and  
there must be sufficient disks at each site to  
create the volume.  
# vxassist -g ccdg make vol 2g \ Create a mirrored volume that is not site  
nmirror=2 siteconsistent=off \ consistent. Both mirrors can be allocated  
allsites=off  
from any available storage in the disk  
group.  
# vxassist -g ccdg make vol 2g \ Create a mirrored volume that is not site  
nmirror=2 site:site2 \  
siteconsistent=off \  
allsites=off  
consistent. Both mirrors are allocated from  
any available storage in the disk group that  
is tagged as belonging to site2.  
# vxassist -g ccdg make vol 2g \ Create a mirrored volume that is not site  
nmirror=2 !site:site1 \  
siteconsistent=off \  
allsites=off  
consistent. Both mirrors are allocated from  
any available storage in the disk group that  
is tagged as not belonging to site1.  
# vxassist -g ccdg mirror vol \  
Add a mirror at a specified site. The  
command fails if there is insufficient  
storage available at the site.  
site:site1  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Administering sites and remote mirrors 439  
Making an existing disk group site consistent  
Command  
Description  
# vxassist -g ccdg remove \  
Remove a mirror from a volume at a  
specified site. If the volume is site  
mirror vol site:site1  
consistent, the command fails if this would  
remove the last remaining plex at a site.  
# vxassist -g ccdg growto vol 4g Grow a volume. If the volume is site  
consistent, the command fails if there is  
insufficient storage available at any site.  
Making an existing disk group site consistent  
To make an existing disk group site consistent  
1
Ensure that the disk group is updated to at least version 140, by running the  
vxdgupgradecommand on it:  
# vxdg upgrade diskgroup  
2
3
4
On each host that can access the disk group, define the site name:  
# vxdctl set site=sitename  
Tag all the disks in the disk group with the appropriate site name:  
# vxdisk [-g diskgroup] settag disk site=sitename  
Use the vxdgmovecommand to move any unsupported RAID-5 volumes to  
another disk group. Alternatively, use the vxassistconvertcommand to  
convert the volumes to a supported layout such as mirroror  
mirror-stripe. You can use the siteand mirror=sitestorage allocation  
attribute to ensure that the plexes are created on the correct storage.  
5
Use the vxevaccommand to ensue that the volumes have equal number of  
plexes at each site. You can use the siteand mirror=sitestorage allocation  
attribute to ensure that the plexes are created on the correct storage.  
6
7
8
Register a site record for each site with the disk group:  
# vxdg -g diskgroup addsite sitename  
Turn on site consistency for the disk group:  
# vxdg -g diskgroup set siteconsistent=on  
Turn on site consistency for each volume in the disk group:  
# vxvol [-g diskgroup] set siteconsistent=on volume ...  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
440 Administering sites and remote mirrors  
Fire drill — testing the configuration  
Fire drill — testing the configuration  
Caution: To avoid potential loss of service or data, it is recommended that you do  
not use these procedures on a live system.  
After validating that the consistency of the volumes and disk groups at your  
sites, you should validate the procedures that you will use in the event of the  
various possible types of failure. A fire drill allows you to test that a site can be  
brought up cleanly during recovery from a disaster scenario such as site failure.  
Simulating site failure  
To simulate the failure of a site, use the following command to detach all the  
devices at a specified site:  
# vxdg -g diskgroup [-f] detachsite sitename  
The -foption must be specified if any plexes configured on storage at the site  
are currently online.  
Recovery from simulated site failure  
Use the following commands to reattach a site and recover the disk group:  
# vxdg -g diskgroup [-o overridessb] reattachsite sitename  
# vxrecover -g diskgroup  
It may be necessary to specify the -ooverridessboption if a serial split-brain  
condition is indicated.  
Automatic site reattachment  
The site reattachment daemon, vxsited, provides automatic reattachment of  
sites. vxsiteduses the vxnotifymechanism to monitor storage coming back  
online on a site after a previous failure, and to restore redundancy of mirrors  
across sites.  
If the hot-relocation daemon, vxrelocd, is running, vxsitedattempts to  
reattach the site, and allows vxrelocdto try to use the available disks in the  
disk group to relocate the failed subdisks. If vxrelocdsucceeds in relocating  
the failed subdisks, it starts the recovery of the plexes at the site. When all the  
plexes have been recovered, the plexes are put into the ACTIVE state, and the  
state of the site is set to ACTIVE.  
If vxrelocdis not running, vxsitedreattaches a site only when all the disks  
at that site become accessible. After reattachment succeeds, vxsitedsets the  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                   
Administering sites and remote mirrors 441  
Failure scenarios and recovery procedures  
site state to ACTIVE, and initiates recovery of the plexes. When all the plexes  
have been recovered, the plexes are put into the ACTIVE state.  
Note: vxsiteddoes not try to reattach a site that you have explicitly detached  
by using the vxdgdetachsitecommand.  
The automatic site reattachment feature is enabled by default. The vxsited  
daemon uses email to notify rootof any attempts to reattach sites and to  
initiate recovery of plexes at those sites. To send mail to other users, add the  
user name to the line that starts vxsitedin the /etc/init.d/  
vxvm-recoverstartup script, and reboot the system.  
If you do not want a site to be recovered automatically, kill the vxsited  
daemon, and prevent it from restarting. To kill the daemon, run the following  
command from the command line:  
# ps -afe  
Locate the process table entry for vxsited, and kill it by specifying its process  
ID:  
# kill -9 PID  
If there is no entry in the process table for vxsited, the automatic site  
reattachment feature is disabled.  
To prevent the automatic site reattachment feature from being restarted,  
comment out the line that starts vxsitedin the /etc/init.d/  
vxvm-recoverstartup script.  
Failure scenarios and recovery procedures  
The site reattachment daemon will provide automatic reattachment of detached  
sites. See “Automatic site reattachment” on page 440. This section describes the  
procedures that you can follow if automatic reattachment does not take place.  
For example, this can happen if you have disable automatic reattachment, or if  
you manually detach a site.  
Possible failure scenarios and recovery techniques are listed in the following  
table:  
Failure scenario  
Recovery technique  
Disruption of network link between sites.  
See “Recovery from a loss of site  
connectivity” on page 442.  
Failure of hosts at a site.  
See “Recovery from host failure” on  
page 442.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
442 Administering sites and remote mirrors  
Failure scenarios and recovery procedures  
Failure scenario  
Recovery technique  
Failure of storage at a site.  
See “Recovery from storage failure” on  
page 442.  
Failure of both hosts and storage at a site.  
See “Recovery from site failure” on  
page 443.  
Recovery from a loss of site connectivity  
If the network links between the sites are disrupted, the application  
environments may continue to run in parallel, and this may lead to  
inconsistencies between the disk group configuration copies at the sites. When  
connectivity between the sites is restored, a serial split-brain condition may  
then exist between the sites. One site must be chosen as having the preferred  
version of the disk group configuration copies. The configuration copies at the  
other sites can then be updated from these copies.  
You can use the following commands to reattach a site and recover the disk  
group:  
# vxdg -g diskgroup -o overridessb reattachsite sitename  
# vxrecover -g diskgroup  
In the case that the host systems are configured at a single site with only storage  
at the remote sites, the usual resynchronization mechanism of VxVM is used to  
recover the remote plexes when the storage comes back on line.  
Recovery from host failure  
If one or more cluster nodes fail at a site, but the storage remains online, this is  
handled either by VCS failover in the case of the Storage Foundation HA  
product, or by node takeover in the case that the node was the master for a  
shared disk group as supported by the Storage Foundation Cluster File System  
software.  
Recovery from storage failure  
If storage fails at a site, the plexes that are configured on that storage are  
detached locally if a site-consistent volume still has other mirrors available at  
the site. The hot-relocation feature of VxVM will attempt to recreate the failed  
plexes on other available storage in the disk group. If no plexes of a  
site-consistent volume remain in operation at a site, and hot-relocation cannot  
recreate the plexes at that site, the site is detached. Because site connectivity  
has not been lost, applications running on hosts at the site can still access data  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Administering sites and remote mirrors 443  
Failure scenarios and recovery procedures  
at the other sites. When the storage comes back online, you can use the  
following commands to reattach a site and recover the disk group:  
# vxdg -g diskgroup reattachsite sitename  
# vxrecover -g diskgroup  
Recovery from site failure  
If all the hosts and storage fail at a site, you can use the following commands to  
reattach the site after it comes back online, and to recover the disk group:  
# vxdg -g diskgroup [-o overridessb] reattachsite sitename  
# vxrecover -g diskgroup  
The -ooverridessboption is only required if a serial split-brain condition is  
indicated, which may happen if the site was brought back up while the private  
network link was inoperative. This option updates the configuration database  
on the reattached site with the consistent copies at the other sites.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
444 Administering sites and remote mirrors  
Failure scenarios and recovery procedures  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter  
15  
Using Storage Expert  
About Storage Expert  
System administrators often find that gathering and interpreting data about  
large and complex configurations can be a difficult task. Veritas Storage Expert  
is designed to help in diagnosing configuration problems with VxVM.  
Storage Expert consists of a set of simple commands that collect VxVM  
configuration data and compare it with “best practice.” Storage Expert then  
produces a summary report that shows which objects do not meet these criteria  
and makes recommendations for VxVM configuration improvements.  
These user-configurable tools help you as an administrator to verify and  
validate systems and non-optimal configurations in both small and large VxVM  
installations.  
See “How Storage Expert works” on page 446.  
See “Before using Storage Expert” on page 446.  
See “Running Storage Expert” on page 446.  
See “Identifying configuration problems using Storage Expert” on page 449.  
See “Rule definitions and attributes” on page 456.  
See the vxse(1M) manual page.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
446 Using Storage Expert  
How Storage Expert works  
How Storage Expert works  
Storage Expert components include a set of rule scripts and a rules engine. The  
rules engine runs the scripts and produces ASCII output, which is organized and  
archived by Storage Expert’s report generator. This output contains information  
about areas of VxVM configuration that do not meet the set criteria. By default,  
output is sent to the screen, but you can send it to a file using standard output  
redirection.  
Before using Storage Expert  
Storage Expert is included in the VRTSvxvmpackage. Even if you do not plan to  
use the VEA graphical user interface, you must also have installed the following  
packages to run vxse:  
■ VRTSob  
■ VRTSvmpro  
The VEA service must also be started on the system by running the command  
/opt/VRTS/bin/vxsvc.  
Running Storage Expert  
The executable rule files are located in the directory, /opt/VRTS/vxse/vxvm.  
It is recommended that this directory be added to the PATHvariable.  
The rules are invoked using the following command-line syntax:  
# rulename [options] keyword [attribute=value ...]  
Each of the rules performs a different function.  
See “Rule definitions and attributes” on page 456.  
The following options may be specified:  
-ddefaults_file Specifies an alternate defaults file.  
-gdiskgroup  
Specifies verbose output format.  
-v  
Specifies the disk group to be examined.  
One of the following keywords must be specified:  
check  
Lists the default values used by the rule’s attributes.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
Using Storage Expert 447  
Running Storage Expert  
info  
list  
run  
Describes what the rule does.  
Lists the attributes of the rule that you can set.  
Runs the rule.  
See “Rule definitions and attributes” on page 456.  
Discovering what a rule does  
To obtain details about what a rule does, use the infokeyword, as in the  
following example:  
# vxse_stripes2 info  
VxVM vxse:vxse_stripes2 INFO V-5-1-6153 This rule checks for  
stripe volumes which have too many or too few columns  
Displaying rule attributes and their default values  
To see the attributes that are available for a given rule, use the listkeyword. In  
the following example, the single attribute, mirror_threshold, is shown for  
the rule vxse_drl1:  
# vxse_drl1 list  
VxVM vxse:vxse_drl1 INFO V-5-1-6004  
vxse_drl1 - TUNEABLES default values  
----------------------------------------------------------  
mirror_threshold - large mirror threshold size  
Warn if a mirror is of greater  
than this size and does not have  
an attached DRL log.  
To see the default values of a specified rule’s attributes, use the checkkeyword  
as shown here:  
# vxse_stripes2 check  
vxse_stripes2 - TUNEABLES  
----------------------------------------------------------  
VxVM vxse:vxse_stripes2 INFO V-5-1-5546  
too_wide_stripe - (16) columns in a striped volume  
too_narrow_stripe - (3) columns in a striped volume  
Storage Expert lists the default value of each of the rule’s attributes.  
See “Rule definitions and attributes” on page 456.  
To alter the behavior of rules, you can change the value of their attributes.  
See “Setting rule attributes” on page 448.  
Running a rule  
The runkeyword invokes a default or reconfigured rule on a disk group or file  
name, for example:  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
448 Using Storage Expert  
Running Storage Expert  
# vxse_dg1 -g mydg run  
VxVM vxse:vxse_dg1 INFO V-5-1-5511 vxse_vxdg1 - RESULTS  
----------------------------------------------------------  
vxse_dg1 PASS:  
Disk group (mydg) okay amount of disks in this disk group (4)  
This indicates that the specified disk group (mydg) met the conditions specified  
in the rule.  
See “Rule result types” on page 448.  
You can set Storage Expert to run as a cronjob to notify administrators, and to  
archive reports automatically.  
Rule result types  
Running a rule generates output that shows the status of the objects that have  
been examined against the rule:  
INFO  
Information about the specified object; for example “RAID-5  
does not have a log.”  
PASS  
The object met the conditions of the rule.  
VIOLATION  
The object did not meet the conditions of the rule.  
Setting rule attributes  
You can set attributes in the following ways:  
■
Enter an attribute on the command line, for example:  
# vxse_drl2 -g mydg run large_mirror_size=30m  
■
Create your own defaults file, and specify that file on the command line:  
# vxse_drl2 -d mydefaultsfile run  
Lines in this file contain attribute values definitions for a rule in this  
format:  
rule_name,attribute=value  
For example, the following entry defines a value of 20 gigabytes for the  
attribute large_mirror_sizeof the rule vxse_drl2:  
vxse_drl2,large_mirror_size=20g  
You can specify values that are to be ignored by inserting a # character at  
the start of the line, for example:  
#vxse_drl2,large_mirror_size=20g  
■
Edit the attribute values that are defined in the /etc/default/vxsefile.  
If you do this, make a backup copy of the file in case you need to regress your  
changes.  
Attributes are applied using the following order of precedence from highest to  
lowest:  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Using Storage Expert 449  
Identifying configuration problems using Storage Expert  
■
■
■
■
A value specified on the command line.  
A value specified in a user-defined defaults file.  
A value in the /etc/default/vxsefile that has not been commented out.  
A built-in value defined at compile time.  
Identifying configuration problems using Storage  
Expert  
Storage Expert provides a large number of rules that help you to diagnose  
configuration issues that might cause problems for your storage environment.  
Each rule describes the issues involved, and suggests remedial actions.  
The rules help you to diagnose problems in the following categories:  
■
■
■
■
■
■
■
Recovery time  
Disk groups  
Disk striping  
Disk sparing and relocation management  
Hardware failures  
Rootability  
System name  
See “Rule definitions and attributes” on page 456.  
Recovery time  
Several “best practice” rules enable you to check that your storage configuration  
has the resilience to withstand a disk failure or a system failure.  
Checking for multiple RAID-5 logs on a physical disk  
(vxse_disklog)  
To check whether more than one RAID-5 log exists on the same physical disk,  
run rule vxse_disklog.  
RAID-5 log mirrors for the same physical volume should be located on separate  
physical disks to ensure redundancy. More than one RAID-5 log on a disk also  
makes the recovery process longer and more complicated.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
450 Using Storage Expert  
Identifying configuration problems using Storage Expert  
Checking for large mirror volumes without a dirty region log  
(vxse_drl1)  
To check whether large mirror volumes (larger than 1GB) have an associated  
dirty region log (DRL), run rule vxse_drl1.  
Creating a DRL speeds recovery of mirrored volumes after a system crash. A  
DRL tracks those regions that have changed and uses the tracking information  
to recover only those portions of the volume that need to be recovered. Without  
a DRL, recovery is accomplished by copying the full contents of the volume  
between its mirrors. This process is lengthy and I/O intensive.  
See “Preparing a volume for DRL and instant snapshots” on page 275.  
Checking for large mirrored volumes without a mirrored dirty  
region log (vxse_drl2)  
To check whether a large mirrored volume has a mirrored DRL log, run rule  
vxse_drl2.  
Mirroring the DRL log provides added protection in the event of a disk failure.  
See “Preparing a volume for DRL and instant snapshots” on page 275.  
Checking for RAID-5 volumes without a RAID-5 log  
(vxse_raid5log1)  
To check whether a RAID-5 volume has an associated RAID-5 log, run rule  
vxse_raid5log1.  
In the event of both a system failure and a failure of a disk in a RAID-5 volume,  
data that is not involved in an active write could be lost or corrupted if there is  
no RAID-5 log.  
See “Adding a RAID-5 log” on page 283.  
Checking minimum and maximum RAID-5 log sizes  
(vxse_raid5log2)  
To check that the size of RAID-5 logs falls within the minimum and maximum  
recommended sizes, run rule vxse_raid5log2.  
The recommended minimum and maximum sizes are 64MB and 1GB  
respectively. If vxse_raid5log2reports that the size of the log is outside these  
boundaries, adjust the size by replacing the log.  
Checking for non-mirrored RAID-5 logs (vxse_raid5log3)  
To check that the RAID-5 log of a large volume is mirrored, run the  
vxse_raid5log3rule.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Using Storage Expert 451  
Identifying configuration problems using Storage Expert  
A mirror of the RAID-5 log protects against loss of data due to the failure of a  
single disk. You are strongly advised to mirror the log if vxse_raid5log3reports  
that the log of a large RAID-5 volume does not have a mirror.  
See “Adding a RAID-5 log” on page 283.  
Disk groups  
Disks groups are the basis of VxVM storage configuration so it is critical that the  
integrity and resilience of your disk groups are maintained. Storage Expert  
provides a number of rules that enable you to check the status of disk groups  
and associated objects.  
Checking whether a configuration database is too full  
(vxse_dg1)  
To check whether the disk group configuration database has become too full,  
run rule vxse_dg1.  
By default, this rule suggests a limit of 250 for the number of disks in a disk  
group. If one of your disk groups exceeds this figure, you should consider  
creating a new disk group. The number of objects that can be configured in a  
disk group is limited by the size of the private region which stores configuration  
information about every object in the disk group. Each disk in the disk group  
that has a private region stores a separate copy of this configuration database.  
See “Creating a disk group” on page 170.  
Checking disk group configuration copies and logs (vxse_dg2)  
To check whether a disk group has too many or too few disk group configuration  
copies, and whether a disk group has too many or too few copies of the disk  
group log, run rule vxse_dg2.  
Checking “on disk config” size (vxse_dg3)  
To check whether a disk group has the correct “on disk config” size, run rule  
vxse_dg3.  
Checking the version number of disk groups (vxse_dg4)  
To check the version number of a disk group, run rule vxse_dg4.  
For optimum results, your disk groups should have the latest version number  
that is supported by the installed version of VxVM.  
See “Upgrading a disk group” on page 208.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
452 Using Storage Expert  
Identifying configuration problems using Storage Expert  
Checking the number of configuration copies in a disk group  
(vxse_dg5)  
To find out whether a disk group has only a single VxVM configured disk, run  
rule vxse_dg5.  
See “Creating and administering disk groups” on page 165.  
Checking for non-imported disk groups (vxse_dg6)  
To check for disk groups that are visible to VxVM but not imported, run rule  
vxse_dg6.  
See “Importing a disk group” on page 174.  
Checking for initialized VM disks that are not in a disk group  
(vxse_disk)  
To find out whether there are any initialized disks that are not a part of any disk  
group, run rule vxse_disk. This prints out a list of disks, indicating whether  
they are part of a disk group or unassociated.  
See “Adding a disk to a disk group” on page 171.  
Checking volume redundancy (vxse_redundancy)  
To check whether a volume is redundant, run rule vxse_redundancy.  
This rule displays a list of volumes together with the number of mirrors that are  
associated with each volume. If vxse_redundancyshows that a volume does not  
have an associated mirror, your data is at risk in the event of a disk failure, and  
you should rectify the situation by creating a mirror for the volume.  
See “Adding a mirror to a volume” on page 271.  
Checking states of plexes and volumes (vxse_volplex)  
To check whether your disk groups contain unused objects (such as plexes and  
volumes), run rule vxse_volplex. In particular, this rule notifies you if any of  
the following conditions exist:  
■
■
■
■
■
■
disabled plexes  
detached plexes  
stopped volumes  
disabled volumes  
disabled logs  
failed plexes  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Using Storage Expert 453  
Identifying configuration problems using Storage Expert  
■
volumes needing recovery  
See “Reattaching plexes” on page 231.  
See “Starting a volume” on page 271.  
See the Veritas Volume Manager Troubleshooting Guide.  
Disk striping  
Striping enables you to enhance your system’s performance. Several rules  
enable you to monitor important parameters such as the number of columns in a  
stripe plex or RAID-5 plex, and the stripe unit size of the columns.  
Checking the configuration of large mirrored-stripe volumes  
(vxse_mirstripe)  
To check whether large mirror-striped volumes should be reconfigured as  
striped-mirror volumes, run rule vxse_mirstripe.  
A large mirrored-striped volume should be reconfigured, using relayout, as a  
striped-mirror volume to improve redundancy and enhance recovery time after  
failure.  
See “Converting between layered and non-layered volumes” on page 300.  
Checking the number of columns in RAID-5 volumes  
(vxse_raid5)  
To check whether RAID-5 volumes have too few or too many columns, run rule  
vxse_raid5.  
By default, this rule assumes that a RAID-5 plex should have more than 4  
columns and fewer than 8 columns.  
See “Performing online relayout” on page 294.  
Checking the stripe unit size of striped volumes  
(vxse_stripes1)  
By default, rule vxse_stripes1reports a violation if a volume’s stripe unit size  
is not set to an integer multiple of 8KB.  
See “Performing online relayout” on page 294.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
454 Using Storage Expert  
Identifying configuration problems using Storage Expert  
Checking the number of columns in striped volumes  
(vxse_stripes2)  
The default values for the number of columns in a striped plex are 16 and 3. By  
default, rule vxse_stripes2reports a violation if a striped plex in your volume  
has fewer than 3 columns or more than 16 columns.  
See “Performing online relayout” on page 294.  
Disk sparing and relocation management  
The hot-relocation feature of VxVM uses spare disks in a disk group to recreate  
volume redundancy after disk failure.  
Checking the number of spare disks in a disk group  
(vxse_spares)  
This “best practice” rule assumes that between 10% and 20% of disks in a disk  
group should be allocated as spare disks. By default, vxse_spareschecks that a  
disk group falls within these limits.  
See “Administering hot-relocation” on page 379.  
Hardware failures  
VxVM maintains information about failed disks and disabled controllers.  
Checking for disk failures and disabled controllers  
(vxse_dc_failures)  
Rule vxse_dc_failurescan be used to discover if the system has any failed  
disks or disabled controllers.  
Rootability  
The root disk can be put under VxVM control and mirrored.  
Checking the validity of root mirrors (vxse_rootmir)  
Rule vxse_rootmircan be used to confirm that the root mirrors are set up  
correctly.  
See “Rootability” on page 102.  
System name  
The system name that is known to VxVM can be checked for consistency.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
                 
Using Storage Expert 455  
Identifying configuration problems using Storage Expert  
Checking the system name (vxse_host)  
Rule vxse_hostcan be used to confirm that the system name in the file  
/etc/vx/volbootis the same as the name that was assigned to the system  
when it was booted.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
456 Using Storage Expert  
Rule definitions and attributes  
Rule definitions and attributes  
You can use the infokeyword to show a description of a rule.  
See “Discovering what a rule does” on page 447.  
Table 15-1 lists the available rule definitions, and rule attributes and their  
default values.  
Table 15-1  
Rule definitions in Storage Expert  
Rule  
Description  
vxse_dc_failures  
vxse_dg1  
Checks and points out failed disks and disabled controllers.  
Checks for disk group configurations in which the disk group has  
become too large.  
vxse_dg2  
Checks for disk group configurations in which the disk group has too  
many or too few disk group configuration copies, and if the disk  
group has too many or too few disk group log copies.  
vxse_dg3  
vxse_dg4  
vxse_dg5  
Checks disk group configurations to verify that the disk group has  
the correct “on disk config” size.  
Checks for disk groups that do not have a current version number,  
and which may need to be upgraded.  
Checks for disk groups in which there is only one VxVM configured  
disk.  
vxse_dg6  
vxse_disk  
Checks for disk groups that are seen, but which are not imported.  
Checks for disks that are initialized, but are not part of any disk  
group.  
vxse_disklog  
vxse_drl1  
Checks for physical disks that have more than one RAID-5 log.  
Checks for large mirror volumes that do not have an associated DRL  
log.  
vxse_drl2  
vxse_host  
Checks for large mirror volumes that do not have DRL log that is  
mirrored.  
Checks that the system “hostname” in the /etc/vx/volbootfile  
matches the hostname that was assigned to the system when it was  
booted.  
vxse_mirstripe  
vxse_raid5  
Checks for large mirror-striped volumes that should be  
striped-mirrors.  
Checks for RAID-5 volumes that are too narrow or too wide.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Using Storage Expert 457  
Rule definitions and attributes  
Table 15-1  
Rule definitions in Storage Expert  
Rule  
Description  
vxse_raid5log1  
vxse_raid5log2  
vxse_raid5log3  
Checks for RAID-5 volumes that do not have an associated log.  
Checks for recommended minimum and maximum RAID-5 log sizes.  
Checks for large RAID-5 volumes that do not have a mirrored  
RAID-5 log.  
vxse_redundancy Checks the redundancy of volumes.  
vxse_rootmir  
vxse_spares  
Checks that all root mirrors are set up correctly.  
Checks that the number of spare disks in a disk group is within the  
VxVM “Best Practices” thresholds.  
vxse_stripes1  
Checks for stripe volumes whose stripe unit is not a multiple of the  
default stripe unit size.  
vxse_stripes2  
vxse_volplex  
Checks for stripe volumes that have too many or too few columns.  
Checks for volumes and plexes that are in various states, such as:  
■
■
■
■
■
■
■
disabled plexes  
detached plexes  
stopped volumes  
disabled volumes  
disabled logs  
failed plexes  
volumes needing recovery  
You can use the listand checkkeywords to show what attributes are available  
for a rule and to display the default values of these attributes.  
See “Running a rule” on page 447.  
Table 15-2 lists the available rule attributes and their default values.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
458 Using Storage Expert  
Rule definitions and attributes  
Table 15-2  
Rule  
Rule attributes and default attribute values  
Attribute Default Description  
value  
vxse_dc_failures -  
-
No user-configurable  
variables.  
vxse_dg1  
max_disks_per_dg  
250  
Maximum number of  
disks in a disk group.  
Warn if a disk group  
has more disks than  
this.  
vxse_dg2  
vxse_dg3  
vxse_dg4  
vxse_dg5  
vxse_dg6  
vxse_disk  
vxse_disklog  
vxse_drl1  
-
-
-
-
-
-
-
-
No user-configurable  
variables.  
-
No user-configurable  
variables.  
-
No user-configurable  
variables.  
-
No user-configurable  
variables.  
-
No user-configurable  
variables.  
-
No user-configurable  
variables.  
-
No user-configurable  
variables.  
mirror_threshold  
1g  
Large mirror threshold  
size. Warn if a mirror is  
larger than this and  
does not have an  
(1GB)  
attached DRL log.  
vxse_drl2  
vxse_host  
large_mirror_size  
20g  
Large mirror-stripe  
(20GB) threshold size. Warn if  
a mirror-stripe volume  
is larger than this.  
-
-
No user-configurable  
variables.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Using Storage Expert 459  
Rule definitions and attributes  
Table 15-2  
Rule  
Rule attributes and default attribute values  
Attribute  
Default Description  
value  
vxse_mirstripe  
large_mirror_size  
1g  
Large mirror-stripe  
(1GB)  
threshold size.  
Warn if a mirror-stripe  
volume is larger than  
this.  
nsd_threshold  
8
Large mirror-stripe  
number of subdisks  
threshold. Warn if a  
mirror-stripe volume  
has more subdisks  
than this.  
vxse_raid5  
too_narrow_raid5  
too_wide_raid5  
4
8
Minimum number of  
RAID-5 columns. Warn  
if actual number of  
RAID-5 columns is less  
than this.  
Maximum number of  
RAID-5 columns. Warn  
if the actual number of  
RAID-5 columns is  
greater than this.  
vxse_raid5log1  
vxse_raid5log2  
-
-
No user-configurable  
variables.  
r5_max_size  
1g  
Maximum RAID-5 log  
check size. Warn if a  
RAID-5 log is larger  
than this.  
(1GB)  
r5_min_size  
64m  
Minimum RAID-5 log  
(64MB) check size. Warn if a  
RAID-5 log is smaller  
than this.  
vxse_raid5log3  
large_vol_size  
20g  
Large RAID-5 volume  
(20GB) threshold size. Warn if  
a RAID-5 volume with  
a non-mirrored RAID-5  
log is larger than this.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
460 Using Storage Expert  
Rule definitions and attributes  
Table 15-2  
Rule attributes and default attribute values  
Rule  
Attribute  
Default Description  
value  
vxse_redundancy  
volume_redundancy  
0
Volume redundancy  
check. The value of 2  
performs a mirror  
redundancy check. A  
value of 1 performs a  
RAID-5 redundancy  
check. The default  
value of 0 performs no  
redundancy check.  
vxse_rootmir  
vxse_spares  
-
-
No user-configurable  
variables.  
max_disk_spare_ratio  
20  
Maximum percentage  
of spare disks in a disk  
group. Warn if the  
percentage of spare  
disks is greater than  
this.  
min_disk_spare_ratio  
10  
Minimum percentage  
of spare disks in a disk  
group. Warn if the  
percentage of spare  
disks is less than this.  
vxse_stripes1  
vxse_stripes2  
default _stripeunit  
8k  
Stripe unit size for  
stripe volumes. Warn if  
a stripe does not have a  
stripe unit which is an  
integer multiple of this  
value.  
(8KB)  
too_narrow_stripe  
too_wide_stripe  
3
Minimum number of  
columns in a striped  
plex. Warn if a striped  
volume has fewer  
columns than this.  
16  
Maximum number of  
columns in a striped  
plex. Warn if a striped  
volume has more  
columns than this.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Using Storage Expert 461  
Rule definitions and attributes  
Table 15-2  
Rule  
Rule attributes and default attribute values  
Attribute  
Default Description  
value  
vxse_volplex  
-
-
No user-configurable  
variables.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
462 Using Storage Expert  
Rule definitions and attributes  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter  
16  
Performance monitoring  
and tuning  
Veritas Volume Manager (VxVM) can improve overall system performance by  
optimizing the layout of data storage on the available hardware. This chapter  
contains guidelines establishing performance priorities, for monitoring  
performance, and for configuring your system appropriately.  
Performance guidelines  
VxVM allows you to optimize data storage performance using the following two  
strategies:  
■
■
Balance the I/O load among the available disk drives.  
Use striping and mirroring to increase I/O bandwidth to the most frequently  
accessed data.  
VxVM also provides data redundancy (through mirroring and RAID-5) that  
allows continuous access to data in the event of disk failure.  
Data assignment  
When deciding where to locate file systems, you, as a system administrator,  
typically attempt to balance I/O load among the available disk drives. The  
effectiveness of this approach is limited by the difficulty of anticipating future  
usage patterns, as well as the inability to split file systems across the drives. For  
example, if a single file system receives the most disk accesses, moving the file  
system to another drive also moves the bottleneck to that drive.  
VxVM can split volumes across multiple drives. This permits you a finer level of  
granularity when locating data. After measuring actual access patterns, you can  
adjust your previous decisions on the placement of file systems. You can  
reconfigure volumes online without adversely impacting their availability.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
464 Performance monitoring and tuning  
Performance guidelines  
Striping  
Striping improves access performance by cutting data into slices and storing it  
on multiple devices that can be accessed in parallel. Striped plexes improve  
access performance for both read and write operations.  
Having identified the most heavily accessed volumes (containing file systems or  
databases), you can increase access bandwidth to this data by striping it across  
portions of multiple disks.  
Figure 16-1 shows an example of a single volume (HotVol) that has been  
identified as a data-access bottleneck. This volume is striped across four disks,  
leaving the remaining space on these disks free for use by less-heavily used  
volumes.  
Figure 16-1  
Use of striping for optimal data access  
HotVol  
HotVol  
HotVol  
HotVol  
PL1 SD1  
PL1 SD2  
PL1 SD3  
PL1 SD4  
Cool volume  
Lightly  
used  
Home  
Less  
directory  
volume  
important  
volume  
Another  
volume  
volume  
Disk 1  
Disk 2  
Disk 3  
Disk 4  
Mirroring  
Note: You need a full license to use this feature.  
Mirroring stores multiple copies of data on a system. When properly applied,  
mirroring provides continuous availability of data and protection against data  
loss due to physical media failure. Mirroring improves the chance of data  
recovery in the event of a system crash or the failure of a disk or other  
hardware.  
In some cases, you can also use mirroring to improve I/O performance. Unlike  
striping, the performance gain depends on the ratio of reads to writes in the disk  
accesses. If the system workload is primarily write-intensive (for example,  
greater than 30 percent writes), mirroring can result in reduced performance.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Performance monitoring and tuning 465  
Performance guidelines  
Combining mirroring and striping  
Note: You need a full license to use this feature.  
Mirroring and striping can be used together to achieve a significant  
improvement in performance when there are multiple I/O streams.  
Striping provides better throughput because parallel I/O streams can operate  
concurrently on separate devices. Serial access is optimized when I/O exactly  
fits across all stripe units in one stripe.  
Because mirroring is generally used to protect against loss of data due to disk  
failures, it is often applied to write-intensive workloads which degrades  
throughput. In such cases, combining mirroring with striping delivers both high  
availability and increased throughput.  
A mirrored-stripe volume may be created by striping half of the available disks  
to form one striped data plex, and striping the remaining disks to form the other  
striped data plex in the mirror. This is often the best way to configure a set of  
disks for optimal performance with reasonable reliability. However, the failure  
of a single disk in one of the plexes makes the entire plex unavailable.  
Alternatively, you can arrange equal numbers of disks into separate mirror  
volumes, and then create a striped plex across these mirror volumes to form a  
striped-mirror volume (see “Mirroring plus striping (striped-mirror, RAID-1+0  
or RAID-10)” on page 43). The failure of a single disk in a mirror does not take  
the disks in the other mirrors out of use. A striped-mirror layout is preferred  
over a mirrored-stripe layout for large volumes or large numbers of disks.  
RAID-5  
Note: You need a full license to use this feature.  
RAID-5 offers many of the advantages of combined mirroring and striping, but  
requires less disk space. RAID-5 read performance is similar to that of striping  
and RAID-5 parity offers redundancy similar to mirroring. Disadvantages of  
RAID-5 include relatively slow write performance.  
RAID-5 is not usually seen as a way of improving throughput performance  
except in cases where the access patterns of applications show a high ratio of  
reads to writes.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
466 Performance monitoring and tuning  
Performance guidelines  
Volume read policies  
To help optimize performance for different types of volumes, VxVM supports  
the following read policies on data plexes:  
■ round—a round-robin read policy, where all plexes in the volume take turns  
satisfying read requests to the volume.  
■ prefer—a preferred-plex read policy, where the plex with the highest  
performance usually satisfies read requests. If that plex fails, another plex is  
accessed.  
■ select—default read policy, where the appropriate read policy for the  
configuration is selected automatically. For example, preferis selected  
when there is only one striped plex associated with the volume, and round  
is selected in most other cases.  
Note: You cannot set the read policy on a RAID-5 data plex. RAID-5 plexes have  
their own read policy (RAID).  
For instructions on how to configure the read policy for a volume’s data plexes,  
see “Changing the read policy for mirrored volumes” on page 289.  
In the configuration example shown in Figure 16-2, the read policy of the  
mirrored-stripe volume labeled HotVolis set to preferfor the striped plex  
PL1. This policy distributes the load when reading across the otherwise lightly-  
used disks in PL1, as opposed to the single disk in plex PL2. (HotVolis an  
example of a mirrored-stripe volume in which one data plex is striped and the  
other data plex is concatenated.)  
Figure 16-2  
Use of mirroring and striping for improved performance  
HotVol  
HotVol  
HotVol  
HotVol  
PL1 SD2  
PL1 SD3  
PL1 SD1  
PL2 SD1  
Lightly  
used  
area  
Lightly  
used  
area  
Lightly  
used  
area  
Lightly  
used  
area  
Disk 1  
Disk 2  
Disk 3  
Disk 4  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Performance monitoring and tuning 467  
Performance monitoring  
Note: To improve performance for read-intensive workloads, you can attach up  
to 32 data plexes to the same volume. However, this would usually be an  
ineffective use of disk space for the gain in read performance.  
Performance monitoring  
As a system administrator, you have two sets of priorities for setting priorities  
for performance. One set is physical, concerned with hardware such as disks and  
controllers. The other set is logical, concerned with managing software and its  
operation.  
Setting performance priorities  
The important physical performance characteristics of disk hardware are the  
relative amounts of I/O on each drive, and the concentration of the I/O within a  
drive to minimize seek time. Based on monitored results, you can then move the  
location of subdisks to balance I/O activity across the disks.  
The logical priorities involve software operations and how they are managed.  
Based on monitoring, you may choose to change the layout of certain volumes to  
improve their performance. You might even choose to reduce overall  
throughput to improve the performance of certain critical volumes. Only you  
can decide what is important on your system and what trade-offs you need to  
make.  
Best performance is usually achieved by striping and mirroring all volumes  
across a reasonable number of disks and mirroring between controllers, when  
possible. This procedure tends to even out the load between all disks, but it can  
make VxVM more difficult to administer. For large numbers of disks (hundreds  
or thousands), set up disk groups containing 10 disks, where each group is used  
to create a striped-mirror volume. This technique provides good performance  
while easing the task of administration.  
Obtaining performance data  
VxVM provides two types of performance information: I/O statistics and I/O  
traces. Each of these can help in performance monitoring. You can obtain I/O  
statistics using the vxstatcommand, and I/O traces using the vxtrace  
command. A brief discussion of each of these utilities may be found in the  
following sections.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
468 Performance monitoring and tuning  
Performance monitoring  
Tracing volume operations  
Use the vxtracecommand to trace operations on specified volumes, kernel I/O  
object types or devices. The vxtracecommand either prints kernel I/O errors or  
I/O trace records to the standard output or writes the records to a file in binary  
format. Binary trace records written to a file can also be read back and formatted  
by vxtrace.  
If you do not specify any operands, vxtracereports either all error trace data or  
all I/O trace data on all virtual disk devices. With error trace data, you can select  
all accumulated error trace data, wait for new error trace data, or both of these  
(this is the default action). Selection can be limited to a specific disk group, to  
specific VxVM kernel I/O object types, or to particular named objects or devices.  
For detailed information about how to use vxtrace, refer to the vxtrace(1M)  
manual page.  
Printing volume statistics  
Use the vxstatcommand to access information about activity on volumes,  
plexes, subdisks, and disks under VxVM control, and to print summary statistics  
to the standard output. These statistics represent VxVM activity from the time  
the system initially booted or from the last time the counters were reset to zero.  
If no VxVM object name is specified, statistics from all volumes in the  
configuration database are reported.  
VxVM records the following I/O statistics:  
■
■
count of operations  
number of blocks transferred (one operation can involve more than one  
block)  
■
average operation time (which reflects the total time through the VxVM  
interface and is not suitable for comparison against other statistics  
programs)  
These statistics are recorded for logical I/O including reads, writes, atomic  
copies, verified reads, verified writes, plex reads, and plex writes for each  
volume. As a result, one write to a two-plex volume results in at least five  
operations: one for each plex, one for each subdisk, and one for the volume. Also,  
one read that spans two subdisks shows at least four reads—one read for each  
subdisk, one for the plex, and one for the volume.  
VxVM also maintains other statistical data. For each plex, it records read and  
write failures. For volumes, it records corrected read and write failures in  
addition to read and write failures.  
To reset the statistics information to zero, use the -roption. This can be done  
for all objects or for only those objects that are specified. Resetting just prior to  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Performance monitoring and tuning 469  
Performance monitoring  
an operation makes it possible to measure the impact of that particular  
operation.  
The following is an example of output produced using the vxstatcommand:  
OPERATIONS  
BLOCKS  
AVG TIME(ms)  
TYP NAME  
vol blop  
vol foobarvol  
READ  
0
WRITE  
READ  
WRITE  
READ  
0.0  
0.0  
WRITE  
0.0  
0.0  
0
0
0
0
0
0
0
vol rootvol 73017 181735 718528 1114227  
26.8  
25.8  
0.0  
27.9  
397.0  
0.0  
vol swapvol 13197 20252 105569  
vol testvol  
162009  
0
0
0
0
Additional volume statistics are available for RAID-5 configurations.  
For detailed information about how to use vxstat, refer to the vxstat(1M)  
manual page.  
Using performance data  
When you have gathered performance data, you can use it to determine how to  
configure your system to use resources most effectively. The following sections  
provide an overview of how you can use this data.  
Using I/O statistics  
Examination of the I/O statistics can suggest how to reconfigure your system.  
You should examine two primary statistics: volume I/O activity and disk I/O  
activity.  
Before obtaining statistics, reset the counters for all existing statistics using the  
vxstat-rcommand. This eliminates any differences between volumes or disks  
due to volumes being created, and also removes statistics from boot time (which  
are not usually of interest).  
After resetting the counters, allow the system to run during typical system  
activity. Run the application or workload of interest on the system to measure  
its effect. When monitoring a system that is used for multiple purposes, try not  
to exercise any one application more than usual. When monitoring a time-  
sharing system with many users, let statistics accumulate for several hours  
during the normal working day.  
To display volume statistics, enter the vxstatcommand with no arguments. The  
following is a typical display of volume statistics:  
OPERATIONS  
BLOCKS  
AVG TIME(ms)  
TYP NAME  
vol archive  
vol home  
vol local  
vol rootvol 102906 342664  
READ  
865  
2980  
WRITE  
807  
5287  
49230  
READ  
5722  
6504  
WRITE  
3809  
10550  
204975  
READ  
32.5  
37.7  
28.5  
28.1  
22.4  
25.3  
WRITE  
24.0  
221.1  
33.5  
25.6  
30.9  
49477  
507892  
1085520 1962946  
425472  
182001  
vol src  
vol swapvol 22751  
79174  
23603  
32364  
139302  
258905  
323.2  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
470 Performance monitoring and tuning  
Performance monitoring  
Such output helps to identify volumes with an unusually large number of  
operations or excessive read or write times.  
To display disk statistics, use the vxstat-dcommand. The following is a typical  
display of disk statistics:  
OPERATIONS  
BLOCKS  
AVG TIME(ms)  
TYP NAME  
READ  
40473 174045  
32668  
55249  
11909  
WRITE  
READ  
WRITE  
READ  
29.5  
35.2  
35.3  
25.0  
WRITE  
35.4  
102.9  
61.2  
30.7  
dm  
dm  
dm  
dm  
mydg01  
455898  
470337  
780779  
114508  
951379  
351351  
731979  
128605  
mydg02  
mydg03  
mydg04  
16873  
60043  
13745  
If you need to move the volume named archiveonto another disk, use the  
following command to identify on which disks it lies:  
# vxprint -g mydg -tvh archive  
The following is an extract from typical output:  
V NAME  
PL NAME  
SD NAME  
RVG/VSET/CO KSTATE  
STATE  
STATE  
DISKOFFS LENGTH  
LENGTH  
LENGTH  
READPOL  
LAYOUT  
[COL/]OFF DEVICE  
REFPLEX  
NCOL/WDTH MODE  
UTYPE  
VOLUME  
PLEX  
KSTATE  
DISK  
MODE  
v archive  
pl archive-01 archive  
sd mydg03-03 archive-01  
-
ENABLED ACTIVE  
ENABLED ACTIVE  
20480  
20480  
40960  
SELECT  
CONCAT  
0
-
-
fsgen  
RW  
ENA  
mydg03  
0
c1t2d0  
Note: Your system may use device names that differ from these examples. For  
more information on device names, see “Administering disks” on page 77.  
The subdisks line (beginning sd) indicates that the volume archiveis on disk  
mydg03. To move the volume off mydg03, use the following command:  
# vxassist -g mydg move archive !mydg03 dest_disk  
Here dest_disk is the destination disk to which you want to move the volume. It  
is not necessary to specify a destination disk. If you do not specify a destination  
disk, the volume is moved to an available disk with enough space to contain the  
volume.  
For example, to move a volume from disk mydg03to disk mydg04, in the disk  
group, mydg, use the following command:  
# vxassist -g mydg move archive !mydg03 mydg04  
This command indicates that the volume is to be reorganized so that no part of it  
remains on mydg03.  
Note: The Veritas Enterprise Administrator (VEA) has a graphical user interface  
(GUI), which provides an easier way to move pieces of volumes between disks.  
You may find that approach preferable to using the command line.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Performance monitoring and tuning 471  
Performance monitoring  
If two volumes (other than the root volume) on the same disk are busy, move  
them so that each is on a different disk.  
If one volume is particularly busy (especially if it has unusually large average  
read or write times), stripe the volume (or split the volume into multiple pieces,  
with each piece on a different disk). If done online, converting a volume to use  
striping requires sufficient free space to store an extra copy of the volume. If  
sufficient free space is not available, a backup copy can be made instead. To  
convert a volume, create a striped plex as a mirror of the volume and then  
remove the old plex. For example, the following commands stripe the volume  
archiveacross disks mydg02, mydg03, and mydg04in the disk group, mydg,  
and then remove the original plex archive-01:  
# vxassist -g mydg mirror archive layout=stripe mydg02 mydg03 \  
mydg04  
# vxplex -g mydg -o rm dis archive-01  
After reorganizing any particularly busy volumes, check the disk statistics. If  
some volumes have been reorganized, clear statistics first and then accumulate  
statistics for a reasonable period of time.  
If some disks appear to be excessively busy (or have particularly long read or  
write times), you may want to reconfigure some volumes. If there are two  
relatively busy volumes on a disk, move them closer together to reduce seek  
times on the disk. If there are too many relatively busy volumes on one disk,  
move them to a disk that is less busy.  
Use I/O tracing (or subdisk statistics) to determine whether volumes have  
excessive activity in particular regions of the volume. If the active regions can  
be identified, split the subdisks in the volume and move those regions to a less  
busy disk.  
Caution: Striping a volume, or splitting a volume across multiple disks, increases  
the chance that a disk failure results in failure of that volume. For example, if  
five volumes are striped across the same five disks, then failure of any one of the  
five disks requires that all five volumes be restored from a backup. If each  
volume were on a separate disk, only one volume would need to be restored. Use  
mirroring or RAID-5 to reduce the chance that a single disk failure results in  
failure of a large number of volumes.  
Note that file systems and databases typically shift their use of allocated space  
over time, so this position-specific information on a volume is often not useful.  
Databases are reasonable candidates for moving to non-busy disks if the space  
used by a particularly busy index or table can be identified.  
Examining the ratio of reads to writes helps to identify volumes that can be  
mirrored to improve their performance. If the read-to-write ratio is high,  
mirroring can increase performance as well as reliability. The ratio of reads to  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
472 Performance monitoring and tuning  
Tuning VxVM  
writes where mirroring can improve performance depends greatly on the disks,  
the disk controller, whether multiple controllers can be used, and the speed of  
the system bus. If a particularly busy volume has a high ratio of reads to writes,  
it is likely that mirroring can significantly improve performance of that volume.  
Using I/O tracing  
I/O statistics provide the data for basic performance analysis; I/O traces serve  
for more detailed analysis. With an I/O trace, focus is narrowed to obtain an  
event trace for a specific workload. This helps to explicitly identify the location  
and size of a hot spot, as well as which application is causing it.  
Using data from I/O traces, real work loads on disks can be simulated and the  
results traced. By using these statistics, you can anticipate system limitations  
and plan for additional resources.  
For information on using the vxdmpadmcommand to gather I/O statistics for a  
DMP node, path, or enclosure, see “Gathering and displaying I/O statistics” on  
page 144. You can also use the vxdmpadmcommand to change the I/O load-  
balancing policy for an enclosure as described in “Specifying the I/O policy” on  
page 147.  
Tuning VxVM  
This section describes how to adjust the tunable parameters that control the  
system resources used by VxVM. Depending on the system resources that are  
available, adjustments may be required to the values of some tunable  
parameters to optimize performance.  
General tuning guidelines  
VxVM is optimally tuned for most configurations ranging from small systems to  
larger servers. In cases where tuning can be used to increase performance on  
larger systems at the expense of a valuable resource (such as memory), VxVM is  
generally tuned to run on the smallest supported configuration. Any tuning  
changes must be performed with care, as they may adversely affect overall  
system performance or may even leave VxVM unusable.  
Various mechanisms exist for tuning VxVM. Many parameters can be tuned  
using the System Management Homepage (SMH) or the kctuneutility. Other  
values can only be tuned using the command line interface to VxVM.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Performance monitoring and tuning 473  
Tuning VxVM  
Tuning guidelines for large systems  
On smaller systems (with fewer than a hundred disk drives), tuning is  
unnecessary and VxVM is capable of adopting reasonable defaults for all  
configuration parameters. On larger systems, configurations can require  
additional control over the tuning of these parameters, both for capacity and  
performance reasons.  
Generally, only a few significant decisions must be made when setting up VxVM  
on a large system. One is to decide on the size of the disk groups and the number  
of configuration copies to maintain for each disk group. Another is to choose the  
size of the private region for all the disks in a disk group.  
Larger disk groups have the advantage of providing a larger free-space pool for  
the vxassist(1M) command to select from, and also allow for the creation of  
larger volumes. Smaller disk groups do not require as large a configuration  
database and so can exist with smaller private regions. Very large disk groups  
can eventually exhaust the private region size in the disk group with the result  
that no more configuration objects can be added to that disk group. At that  
point, the configuration either has to be split into multiple disk groups, or the  
private regions have to be enlarged. This involves re-initializing each disk in the  
disk group (and can involve reconfiguring everything and restoring from  
backup).  
A general recommendation for users of disk array subsystems is to create a  
single disk group for each array so the disk group can be physically moved as a  
unit between systems.  
Number of configuration copies for a disk group  
Selection of the number of configuration copies for a disk group is based on a  
trade-off between redundancy and performance. As a general rule, reducing the  
number configuration copies in a disk group speeds up initial access of the disk  
group, initial startup of the vxconfigddaemon, and transactions performed  
within the disk group. However, reducing the number of configuration copies  
also increases the risk of complete loss of the configuration database, which  
results in the loss of all objects in the database and of all data in the disk group.  
The default policy for configuration copies in the disk group is to allocate a  
configuration copy for each controller identified in the disk group, or for each  
target that contains multiple addressable disks. This provides a sufficient  
degree of redundancy, but can lead to a large number of configuration copies  
under some circumstances. If this is the case, we recommended that you limit  
the number of configuration copies to a maximum of 4. Distribute the copies  
across separate controllers or targets to enhance the effectiveness of this  
redundancy.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
474 Performance monitoring and tuning  
Tuning VxVM  
To set the number of configuration copies for a new disk group, use the nconfig  
operand with the vxdginitcommand (see the vxdg(1M) manual page for  
details).  
You can also change the number of copies for an existing group by using the  
vxeditsetcommand (see the vxedit(1M) manual page). For example, to  
configure five configuration copies for the disk group, bigdg, use the following  
command:  
# vxedit set nconfig=5 bigdg  
Changing the values of tunables  
Tunables are modified by using the System Management Homepage (SMH) or  
the kctuneutility. Changed tunables take effect only after relinking the kernel  
and booting the system from the new kernel.  
The values of system tunables can be examined or changed in SMH by selecting  
Home > System Configuration > Kernel Configuration > Tunables.  
Most DMP tunables may be set online (without requiring a reboot) by using the  
vxdmpadmcommand as shown here:  
# vxdmpadm settune dmp_tunable=value  
The values of these tunables can be displayed by using this command:  
# vxdmpadm gettune [dmp_tunable]  
The vxdmpadmcommand also allows you to configure how DMP responds to I/O  
errors at the level of the paths to individual arrays. See “Administering DMP  
using vxdmpadm” on page 139 for details.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Performance monitoring and tuning 475  
Tuning VxVM  
Tunable parameters  
The following sections describe specific tunable parameters.  
dmp_cache_open  
If set to on, the first open of a device that is performed by an array support  
library (ASL) is cached. This enhances the performance of device discovery by  
minimizing the overhead caused by subsequent opens by ASLs. If set to off,  
caching is not performed. The default value is off.  
The value of this tunable is changed by using the vxdmpadm settunecommand.  
dmp_daemon_count  
The number of kernel threads that are available for servicing path error  
handling, path restoration and other DMP administrative tasks.  
The default number of threads is 10.  
The value of this tunable is changed by using the vxdmpadm settunecommand.  
dmp_delayq_interval  
How long DMP should wait before retrying I/O after an array fails over to a  
standby path. Some disk arrays are not be capable of accepting I/O requests  
immediately after failover.  
The default value is 15 seconds.  
The value of this tunable is changed by using the vxdmpadm settunecommand.  
“Configuring DMP path restoration policies” on  
page 160dmp_failed_io_threshold  
The time limit that DMP waits for a failed I/O request to return before the device  
is marked as INSANE, I/O is avoided on the path, and any remaining failed I/O  
requests are returned to the application layer without performing any error  
analysis.  
The default value is 57600 seconds (16 hours).  
The value of this tunable is changed by using the vxdmpadmsettunecommand.  
dmp_fast_recovery  
Whether DMP should attempt to obtain SCSI error information directly from the  
HBA interface. Setting the value to on can potentially provide faster error  
recovery, provided that the HBA interface supports the error enquiry feature.  
If set to off, the HBA interface is not used. This is the default setting.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
476 Performance monitoring and tuning  
Tuning VxVM  
The value of this tunable is changed by using the vxdmpadm settunecommand.  
dmp_health_time  
DMP detects intermittently failing paths, and prevents I/O requests from being  
sent on them. The value of dmp_health_timerepresents the time in seconds  
for which a path must stay healthy. If a path’s state changes back from enabled  
to disabled within this time period, DMP marks the path as intermittently  
failing, and does not re-enable the path for I/O until dmp_path_ageseconds  
elapse. The default value of dmp_health_timeis 60 seconds. A value of 0  
prevents DMP from detecting intermittently failing paths.  
The value of this tunable is changed by using the vxdmpadm settunecommand.  
dmp_log_level  
The level of detail that is displayed for DMP console messages. The following  
level values are defined:  
1
2
3
4
Display all DMP log messages that existed in releases before 5.0. This is  
the default setting.  
Display level 1 messages plus messages that relate to I/O throttling,  
suspected paths, repeated path failures and DMP node migration.  
Display level 1 and 2 messages plus messages that relate to I/O errors,  
I/O error analysis and path media errors.  
Display level 1, 2 and 3 messages plus messages that relate to setting or  
changing attributes on a path.  
The value of this tunable is changed by using the vxdmpadm settunecommand.  
dmp_path_age  
The time for which an intermittently failing path needs to be monitored as  
healthy before DMP once again attempts to schedule I/O requests on it. The  
default value is 300 seconds. A value of 0 prevents DMP from detecting  
intermittently failing paths.  
The value of this tunable is changed by using the vxdmpadm settunecommand.  
dmp_pathswitch_blks_shift  
The default number of contiguous I/O blocks (expressed as the integer exponent  
of a power of 2; for example 10 represents 1024 blocks) that are sent along a  
DMP path to an Active/Active array before switching to the next available path.  
The default value of this parameter is set to 10 so that 1024 blocks (1MB) of  
contiguous I/O are sent over a DMP path before switching. For intelligent disk  
arrays with internal data caches, better throughput may be obtained by  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Performance monitoring and tuning 477  
Tuning VxVM  
increasing the value of this tunable. For example, for the HDS 9960 A/A array,  
the optimal value is between 14 and 16 for an I/O activity pattern that consists  
mostly of sequential reads or writes.  
Note: This parameter only affects the behavior of the balancedI/O policy. A  
value of 0 disables multipathing for the policy unless the vxdmpadmcommand is  
used to specify a different partition size for an array as described in “Specifying  
the I/O policy” on page 147.  
The value of this tunable is changed by using the vxdmpadm settunecommand.  
dmp_probe_idle_lun  
If DMP statistics gathering is enabled, set to on(default) to have the DMP path  
restoration thread probe idle LUNs, or to offto turn off this feature. (Idle LUNs  
are VM disks on which no I/O requests are scheduled.) The value of this tunable  
is only interpreted when DMP statistics gathering is enabled. Turning off  
statistics gathering also disables idle LUN probing.  
The value of this tunable is changed by using the vxdmpadm settunecommand.  
dmp_queue_depth  
The maximum number of queued I/O requests on a path during I/O throttling.  
The default value is 40.  
The value of this tunable is changed by using the vxdmpadm settunecommand.  
See “Configuring the I/O throttling mechanism” on page 157.  
dmp_restore_cycles  
If the DMP restore policy is CHECK_PERIODIC, the number of cycles after which  
the CHECK_ALL policy is called.  
The value of this tunable is only changeable by using the vxdmpadm start  
restorecommand.  
dmp_restore_interval  
The time in seconds between two invocations of the DMP path restoration  
thread.  
The value of this tunable is only changeable by using the vxdmpadm start  
restorecommand.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
478 Performance monitoring and tuning  
Tuning VxVM  
dmp_restore_policy  
The DMP restore policy, which can be set to 0 (CHECK_ALL), 1  
(CHECK_DISABLED), 2 (CHECK_PERIODIC), or 3 (CHECK_ALTERNATE).  
The value of this tunable is only changeable by using the vxdmpadm start  
restorecommand.  
dmp_retry_count  
If an inquiry succeeds on a path, but there is an I/O error, the number of retries  
to attempt on the path. The default number of retries is 30.  
The value of this tunable is changed by an entry in the /kernel/drv/  
vxdmp.conffile, or by using the vxdmpadmsettunecommand.  
See “Configuring the response to I/O failures” on page 156.  
dmp_retry_timeout  
The maximum time period for which DMP retries the SCSI-3 Persistent Reserve  
operation with A/P arrays.  
The default value is 0 seconds, which disables this feature.  
This parameter has no direct effect on I/O processing by DMP.  
Disabling a switch port can trigger a fabric reconfiguration, which can take time  
to stabilize. During this period, attempting to register PGR keys through the  
secondary path to an array may fail with an error condition, such as unit  
attention or device reset, or the return of vendor-specific sense data.  
The retry period allows a fabric reconfiguration (usually a transient condition)  
to not be seen as an error by DMP.  
Do not set the value of the retry period too high. This can delay the failover  
process, and result in I/O sluggishness or suppression of I/O activity during the  
retry period.  
The value of this tunable is changed by using the vxdmpadm settunecommand.  
dmp_scsi_timeout  
Determines the timeout value to be set for any SCSI command that is sent via  
DMP. If the HBA does not receive a response for a SCSI command that it has sent  
to the device within the timeout period, the SCSI command is returned with a  
failure error code.  
The default value is 60 seconds.  
The value of this tunable is changed by using the vxdmpadm settunecommand.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Performance monitoring and tuning 479  
Tuning VxVM  
dmp_stat_interval  
The time interval between gathering DMP statistics. The default and minimum  
value is 1 second.  
The value of this tunable is changed by using the vxdmpadm settunecommand.  
vol_checkpt_default  
The interval at which utilities performing recoveries or resynchronization  
operations load the current offset into the kernel as a checkpoint. A system  
failure during such operations does not require a full recovery, but can continue  
from the last reached checkpoint.  
The default value of the checkpoint is 10240 sectors (10MB).  
Increasing this size reduces the overhead of checkpointing on recovery  
operations at the expense of additional recovery following a system failure  
during a recovery.  
vol_default_iodelay  
The count in clock ticks for which utilities pause if they have been directed to  
reduce the frequency of issuing I/O requests, but have not been given a specific  
delay time. This tunable is used by utilities performing operations such as  
resynchronizing mirrors or rebuilding RAID-5 columns.  
The default for this tunable is 50 ticks.  
Increasing this value results in slower recovery operations and consequently  
lower system impact while recoveries are being performed.  
vol_fmr_logsz  
The maximum size in kilobytes of the bitmap that Non-Persistent FastResync  
uses to track changed blocks in a volume. The number of blocks in a volume that  
are mapped to each bit in the bitmap depends on the size of the volume, and this  
value changes if the size of the volume is changed. For example, if the volume  
size is 1 gigabyte and the system block size is 1024 bytes, a vol_fmr_logsz  
value of 4 yields a map contains 32,768 bits, each bit representing one region of  
32 blocks.  
The larger the bitmap size, the fewer the number of blocks that are mapped to  
each bit. This can reduce the amount of reading and writing required on  
resynchronization, at the expense of requiring more non-pageable kernel  
memory for the bitmap. Additionally, on clustered systems, a larger bitmap size  
increases the latency in I/O performance, and it also increases the load on the  
private network between the cluster members. This is because every other  
member of the cluster must be informed each time a bit in the map is marked.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
480 Performance monitoring and tuning  
Tuning VxVM  
Since the region size must be the same on all nodes in a cluster for a shared  
volume, the value of the vol_fmr_logsztunable on the master node overrides  
the tunable values on the slave nodes, if these values are different. Because the  
value of a shared volume can change, the value of vol_fmr_logszis retained  
for the life of the volume.  
In configurations which have thousands of mirrors with attached snapshot  
plexes, the total memory overhead can represent a significantly higher  
overhead in memory consumption than is usual for VxVM.  
The default value of this tunable is 4KB. The maximum and minimum permitted  
values are 1KB and 8KB.  
Note: The value of this tunable does not have any effect on Persistent  
FastResync.  
vol_max_vol  
The maximum number of volumes that can be created on the system. This value  
can be set to between 1 and the maximum number of minor numbers  
representable in the system.  
The default value for this tunable is 16777215.  
vol_maxio  
The maximum size of logical I/O operations that can be performed without  
breaking up the request. I/O requests to VxVM that are larger than this value are  
broken up and performed synchronously. Physical I/O requests are broken up  
based on the capabilities of the disk device and are unaffected by changes to this  
maximum logical request limit.  
The default value for this tunable is 256 sectors (256KB).  
Note: The value of voliomem_maxpool_szmust be at least 10 times greater  
than the value of vol_maxio.  
If DRL sequential logging is configured, the value of voldrl_min_regionsz  
must be set to at least half the value of vol_maxio.  
vol_maxioctl  
The maximum size of data that can be passed into VxVM via an ioctlcall.  
Increasing this limit allows larger operations to be performed. Decreasing the  
limit is not generally recommended, because some utilities depend upon  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Performance monitoring and tuning 481  
Tuning VxVM  
performing operations of a certain size and can fail unexpectedly if they issue  
oversized ioctlrequests.  
The default value for this tunable is 32768 bytes (32KB).  
vol_maxparallelio  
The number of I/O operations that the vxconfigd(1M) daemon is permitted to  
request from the kernel in a single VOL_VOLDIO_READper  
VOL_VOLDIO_WRITEioctlcall.  
The default value for this tunable is 256. It is not desirable to change this value.  
vol_maxspecialio  
The maximum size of an I/O request that can be issued by an ioctlcall.  
Although the ioctlrequest itself can be small, it can request a large I/O  
request be performed. This tunable limits the size of these I/O requests. If  
necessary, a request that exceeds this value can be failed, or the request can be  
broken up and performed synchronously.  
The default value for this tunable is 256 sectors (256KB).  
Raising this limit can cause difficulties if the size of an I/O request causes the  
process to take more memory or kernel virtual mapping space than exists and  
thus deadlock. The maximum limit for vol_maxspecialiois 20% of the  
smaller of physical memory or kernel virtual memory. It is inadvisable to go  
over this limit, because deadlock is likely to occur.  
If stripes are larger than vol_maxspecialio, full stripe I/O requests are  
broken up, which prevents full-stripe read/writes. This throttles the volume I/O  
throughput for sequential I/O or larger I/O requests.  
This tunable limits the size of an I/O request at a higher level in VxVM than the  
level of an individual disk. For example, for an 8 by 64KB stripe, a value of 256KB  
only allows I/O requests that use half the disks in the stripe; thus, it cuts  
potential throughput in half. If you have more columns or you have used a  
larger interleave factor, then your relative performance is worse.  
This tunable must be set, as a minimum, to the size of your largest stripe (RAID-  
0 or RAID-5).  
vol_subdisk_num  
The maximum number of subdisks that can be attached to a single plex. There is  
no theoretical limit to this number, but it has been limited to a default value of  
4096. This default can be changed, if required.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
482 Performance monitoring and tuning  
Tuning VxVM  
volcvm_smartsync  
If set to 0, volcvm_smartsyncdisables SmartSync on shared disk groups. If  
set to 1, this parameter enables the use of SmartSync with shared disk groups.  
See“SmartSync recovery accelerator” on page 62 for more information.  
voldrl_max_drtregs  
The maximum number of dirty regions that can exist on the system for non-  
sequential DRL on volumes. A larger value may result in improved system  
performance at the expense of recovery time. This tunable can be used to  
regulate the worse-case recovery time for the system following a failure.  
The default value for this tunable is 2048.  
voldrl_max_seq_dirty  
The maximum number of dirty regions allowed for sequential DRL. This is  
useful for volumes that are usually written to sequentially, such as database  
logs. Limiting the number of dirty regions allows for faster recovery if a crash  
occurs.  
The default value for this tunable is 3.  
voldrl_min_regionsz  
The minimum number of sectors for a dirty region logging (DRL) volume region.  
With DRL, VxVM logically divides a volume into a set of consecutive regions.  
Larger region sizes tend to cause the cache hit-ratio for regions to improve. This  
improves the write performance, but it also prolongs the recovery time.  
The VxVM kernel currently sets the default value for this tunable to 512 sectors.  
Note: If DRL sequential logging is configured, the value of  
voldrl_min_regionszmust be set to at least half the value of vol_maxio.  
voliomem_chunk_size  
The granularity of memory chunks used by VxVM when allocating or releasing  
system memory. A larger granularity reduces CPU overhead due to memory  
allocation by allowing VxVM to retain hold of a larger amount of memory.  
The default size for this tunable is 64KB.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Performance monitoring and tuning 483  
Tuning VxVM  
voliomem_maxpool_sz  
The maximum memory requested from the system by VxVM for internal  
purposes. This tunable has a direct impact on the performance of VxVM as it  
prevents one I/O operation from using all the memory in the system.  
VxVM allocates two pools that can grow up to voliomem_maxpool_sz, one for  
RAID-5 and one for mirrored volumes.  
A write request to a RAID-5 volume that is greater than  
voliomem_maxpool_sz/10is broken up and performed in chunks of size  
voliomem_maxpool_sz/10.  
A write request to a mirrored volume that is greater than  
voliomem_maxpool_sz/2is broken up and performed in chunks of size  
voliomem_maxpool_sz/2.  
The default value for this tunable is 4M.  
Note: The value of voliomem_maxpool_szmust be at least 10 times greater  
than the value of vol_maxio.  
voliot_errbuf_dflt  
The default size of the buffer maintained for error tracing events. This buffer is  
allocated at driver load time and is not adjustable for size while VxVM is  
running.  
The default size for this buffer is 16384 bytes (16KB).  
Increasing this buffer can provide storage for more error events at the expense  
of system memory. Decreasing the size of the buffer can result in an error not  
being detected via the tracing device. Applications that depend on error tracing  
to perform some responsive action are dependent on this buffer.  
voliot_iobuf_default  
The default size for the creation of a tracing buffer in the absence of any other  
specification of desired kernel buffer size as part of the trace ioctl.  
The default size of this tunable is 8192 bytes (8KB).  
If trace data is often being lost due to this buffer size being too small, then this  
value can be tuned to a more generous amount.  
voliot_iobuf_limit  
The upper limit to the size of memory that can be used for storing tracing  
buffers in the kernel. Tracing buffers are used by the VxVM kernel to store the  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
484 Performance monitoring and tuning  
Tuning VxVM  
tracing event records. As trace buffers are requested to be stored in the kernel,  
the memory for them is drawn from this pool.  
Increasing this size can allow additional tracing to be performed at the expense  
of system memory usage. Setting this value to a size greater than can readily be  
accommodated on the system is inadvisable.  
The default value for this tunable is 131072 bytes (128KB).  
voliot_iobuf_max  
The maximum buffer size that can be used for a single trace buffer. Requests of  
a buffer larger than this size are silently truncated to this size. A request for a  
maximal buffer size from the tracing interface results (subject to limits of usage)  
in a buffer of this size.  
The default size for this buffer is 65536 bytes (64KB).  
Increasing this buffer can provide for larger traces to be taken without loss for  
very heavily used volumes. Care should be taken not to increase this value above  
the value for the voliot_iobuf_limittunable value.  
voliot_max_open  
The maximum number of tracing channels that can be open simultaneously.  
Tracing channels are clone entry points into the tracing device driver. Each  
vxtraceprocess running on a system consumes a single trace channel.  
The default number of channels is 32. The allocation of each channel takes up  
approximately 20 bytes even when not in use.  
volpagemod_max_memsz  
The amount of memory, measured in kilobytes, that is allocated for caching  
FastResync and cache object metadata. This tunable has a default value of  
6144KB (6MB) of physical memory.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Performance monitoring and tuning 485  
Tuning VxVM  
Note: The memory allocated for this cache is exclusively dedicated to it. It is not  
available for other processes or applications.  
Setting the value of volpagemod_max_memszbelow 512KB fails if cache  
objects or volumes that have been prepared for instant snapshot operations are  
present on the system.  
If you do not use the FastResync or DRL features that are implemented using a  
version 20 DCO volume, the value of volpagemod_max_memszcan be set to 0.  
However, if you subsequently decide to enable these features, you can use the  
vxtunecommand to change the value to a more appropriate one:  
# vxtune volpagemod_max_memsz value  
where the new value is specified in kilobytes. Using the vxtunecommand to  
adjust the value of volpagemod_max_memsz does not persist across system  
reboots unless you also adjust the value that is configured in the /stand/  
systemfile.  
volraid_minpool_sz  
The initial amount of memory that is requested from the system by VxVM for  
RAID-5 operations. The maximum size of this memory pool is limited by the  
value of voliomem_maxpool_sz.  
The default value for this tunable is 16348 sectors (16MB).  
volraid_rsrtransmax  
The maximum number of transient reconstruct operations that can be  
performed in parallel for RAID-5. A transient reconstruct operation is one that  
occurs on a non-degraded RAID-5 volume that has not been predicted. Limiting  
the number of these operations that can occur simultaneously removes the  
possibility of flooding the system with many reconstruct operations, and so  
reduces the risk of causing memory starvation.  
The default number of transient reconstruct operations that can be performed  
in parallel is 1.  
Increasing this size improves the initial performance on the system when a  
failure first occurs and before a detach of a failing object is performed, but can  
lead to memory starvation.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
486 Performance monitoring and tuning  
Tuning VxVM  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Appendix  
A
Commands summary  
This appendix summarizes the usage and purpose of important commonly-used  
commands in Veritas Volume Manager (VxVM). References are included to  
longer descriptions in the remainder of this book.  
Most commands (excepting daemons, library commands and supporting scripts)  
are linked to the /usr/sbindirectory from the /opt/VRTS/bindirectory. It is  
recommended that you add the following directories to your PATHenvironment  
variable:  
■
If you are using the Bourne or Korn shell (shor ksh), use the commands:  
$ PATH=$PATH:/usr/sbin:/opt/VRTS/bin:/opt/VRTSvxfs/sbin:\  
/opt/VRTSdbed/bin:/opt/VRTSdb2ed/bin:/opt/VRTSsybed/bin:\  
/opt/VRTSob/bin  
$ MANPATH=/usr/share/man:/opt/VRTS/man:$MANPATH  
$ export PATH MANPATH  
■
If you are using a C shell (cshor tcsh), use the commands:  
% set path = ( $path /usr/sbin /opt/VRTSvxfs/sbin \  
/opt/VRTSdbed/bin /opt/VRTSdb2ed/bin /opt/VRTSsybed/bin \  
/opt/VRTSob/bin /opt/VRTS/bin )  
% setenv MANPATH /usr/share/man:/opt/VRTS/man:$MANPATH  
Note: If you have not installed database software, you can omit /opt/  
VRTSdbed/bin, /opt/VRTSdb2ed/binand /opt/VRTSsybed/bin.  
Similarly, /opt/VRTSvxfs/binis only required to access some VxFS  
commands.  
VxVM library commands and supporting scripts are located under the /usr/  
lib/vxvmdirectory hierarchy. You can include these directories in your path if  
you need to use them on a regular basis.  
For detailed information about an individual command, refer to the appropriate  
manual page in the 1M section. A list of manual pages is provided in “Online  
manual pages” on page 507. Commands and scripts that are provided to support  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
488 Commands summary  
other commands and scripts, and which are not intended for general use, are not  
located in /opt/VRTS/binand do not have manual pages.  
The following tables summarize the commonly-used commands:  
■
■
■
■
■
■
■
■
“Obtaining information about objects in VxVM” on page 488  
“Administering disks” on page 489  
“Creating and administering disk groups” on page 492  
“Creating and administering subdisks” on page 494  
“Creating and administering plexes” on page 496  
“Creating volumes” on page 498  
“Administering volumes” on page 501  
“Monitoring and controlling tasks” on page 505  
Table A-1  
Obtaining information about objects in VxVM  
Command  
Description  
vxdctl license  
List licensed features of VxVM.  
vxdisk [-g diskgroup] list [diskname]  
Lists disks under control of VxVM.  
See “Displaying disk information” on  
page 120.  
Example:  
# vxdisk -g mydg list  
vxdg list [diskgroup]  
Lists information about disk groups.  
See “Displaying disk group  
information” on page 169.  
Example:  
# vxdg list mydg  
vxdg -s list  
Lists information about shared disk  
groups.  
See “Listing shared disk groups” on  
page 421.  
Example:  
# vxdg -s list  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Commands summary 489  
Table A-1  
Command  
Obtaining information about objects in VxVM  
Description  
vxinfo [-g diskgroup] [volume ...]  
Displays information about the  
accessibility and usability of volumes.  
See “Listing Unstartable Volumes” in  
the Veritas Volume Manager  
Troubleshooting Guide.  
Example:  
# vxinfo -g mydg myvol1 \  
myvol2  
vxprint -hrt [-g diskgroup] [object]  
Prints single-line information about  
objects in VxVM.  
See “Displaying volume information”  
on page 264.  
Example:  
# vxprint -g mydg myvol1 \  
myvol2  
vxprint -st [-g diskgroup] [subdisk]  
Displays information about subdisks.  
See “Displaying subdisk information”  
on page 216.  
Example:  
# vxprint -st -g mydg  
vxprint -pt [-g diskgroup] [plex]  
Displays information about plexes.  
See “Displaying plex information” on  
page 224.  
Example:  
# vxprint -pt -g mydg  
Table A-2  
Administering disks  
Command  
Description  
vxdiskadm  
Administers disks in VxVM using a  
menu-based interface.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
490 Commands summary  
Table A-2  
Command  
Administering disks  
Description  
vxdiskadd [devicename ...]  
Adds a disk specified by device name.  
See “Using vxdiskadd to place a disk  
under control of VxVM” on page 101.  
Example:  
# vxdiskadd c0t1d0  
vxedit [-g diskgroup] rename olddisk \ Renames a disk under control of  
newdisk  
VxVM.  
See “Renaming a disk” on page 119.  
Example:  
# vxedit -g mydg rename \  
mydg03 mydg02  
vxedit [-g diskgroup] set \  
reserve=on|off diskname  
Sets aside/does not set aside a disk  
from use in a disk group.  
See “Reserving disks” on page 119.  
Examples:  
# vxedit -g mydg set \  
reserve=on mydg02  
# vxedit -g mydg set \  
reserve=off mydg02  
vxedit [-g diskgroup] set \  
nohotuse=on|off diskname  
Does not/does allow free space on a  
disk to be used for hot-relocation.  
See “Excluding a disk from hot-  
relocation use” on page 388.  
See “Making a disk available for hot-  
relocation use” on page 389.  
Examples:  
# vxedit -g mydg set \  
nohotuse=on mydg03  
# vxedit -g mydg set \  
nohotuse=off mydg03  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Commands summary 491  
Table A-2  
Command  
Administering disks  
Description  
vxedit [-g diskgroup] set \  
spare=on|off diskname  
Adds/removes a disk from the pool of  
hot-relocation spares.  
See “Marking a disk as a hot-  
relocation spare” on page 387.  
See “Removing a disk from use as a  
hot-relocation spare” on page 388.  
Examples:  
# vxedit -g mydg set \  
spare=on mydg04  
# vxedit -g mydg set \  
spare=off mydg04  
vxdisk offline devicename  
vxdg -g diskgroup rmdisk diskname  
vxdiskunsetup devicename  
Takes a disk offline.  
See “Taking a disk offline” on  
page 118.  
Example:  
# vxdisk offline c0t1d0  
Removes a disk from its disk group.  
See “Removing a disk from a disk  
group” on page 172.  
Example:  
# vxdg -g mydg rmdisk c0t2d0  
Removes a disk from control of VxVM.  
See “Removing a disk from a disk  
group” on page 172.  
Example:  
# vxdiskunsetup c0t3d0  
Download from Www.Somanuals.com. All Manuals Search And Download.  
492 Commands summary  
Table A-3  
Command  
Creating and administering disk groups  
Description  
vxdg [-s] init diskgroup \  
[diskname=]devicename  
Creates a disk group using a pre-  
initialized disk.  
See “Creating a disk group” on  
page 170.  
See “Creating a shared disk group” on  
page 422.  
Example:  
# vxdg init mydg \  
mydg01=c0t1d0  
vxsplitlines -g diskgroup  
Reports conflicting configuration  
information.  
See “Handling conflicting  
configuration copies” on page 190.  
Example:  
# vxsplitlines -g mydg  
vxdg [-n newname] deport diskgroup  
Deports a disk group and optionally  
renames it.  
See “Deporting a disk group” on  
page 173.  
Example:  
# vxdg -n newdg deport mydg  
vxdg [-n newname] import diskgroup  
Imports a disk group and optionally  
renames it.  
See “Importing a disk group” on  
page 174.  
Example:  
# vxdg -n newdg import mydg  
vxdg [-n newname] -s import diskgroup Imports a disk group as shared by a  
cluster, and optionally renames it.  
See “Importing disk groups as shared”  
on page 423.  
Example:  
# vxdg -n newsdg -s import \  
mysdg  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Commands summary 493  
Table A-3  
Command  
Creating and administering disk groups  
Description  
vxdg [-o expand] listmove sourcedg \  
Lists the objects potentially affected  
targetdg object ...  
by moving a disk group.  
See “Listing objects potentially  
affected by a move” on page 200.  
Example:  
# vxdg -o expand listmove \  
mydg newdg myvol1  
vxdg [-o expand] move sourcedg \  
Moves objects between disk groups.  
targetdg object ...  
See “Moving objects between disk  
groups” on page 203.  
Example:  
# vxdg -o expand move mydg \  
newdg myvol1  
vxdg [-o expand] split sourcedg \  
targetdg object ...  
Splits a disk group and moves the  
specified objects into the target disk  
group.  
See “Splitting disk groups” on  
page 205.  
Example:  
# vxdg -o expand split mydg \  
newdg myvol2 myvol3  
vxdg join sourcedg targetdg  
Joins two disk groups.  
See “Joining disk groups” on page 206.  
Example:  
# vxdg join newdg mydg  
vxdg -g diskgroup set \  
activation=ew|ro|sr|sw|off  
Sets the activation mode of a shared  
disk group in a cluster.  
See “Changing the activation mode on  
a shared disk group” on page 425.  
Example:  
# vxdg -g mysdg set \  
activation=sw  
Download from Www.Somanuals.com. All Manuals Search And Download.  
494 Commands summary  
Table A-3  
Command  
Creating and administering disk groups  
Description  
vxrecover -g diskgroup -sb  
Starts all volumes in an imported disk  
group.  
See “Moving disk groups between  
systems” on page 185.  
Example:  
# vxrecover -g mydg -sb  
vxdg destroy diskgroup  
Destroys a disk group and releases its  
disks.  
See “Destroying a disk group” on  
page 208.  
Example:  
# vxdg destroy mydg  
Table A-4  
Command  
Creating and administering subdisks  
Description  
vxmake [-g diskgroup] sd subdisk \  
Creates a subdisk.  
diskname,offset,length  
See “Creating subdisks” on page 215.  
Example:  
# vxmake -g mydg sd \  
mydg02-01 mydg02,0,8000  
vxsd [-g diskgroup] assoc plex \  
subdisk...  
Associates subdisks with an existing  
plex.  
See “Associating subdisks with  
plexes” on page 218.  
Example:  
# vxsd -g mydg assoc home-1  
mydg02-01 \ mydg02-00  
mydg02-01  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Commands summary 495  
Table A-4  
Command  
Creating and administering subdisks  
Description  
vxsd [-g diskgroup] assoc plex \  
subdisk1:0 ... subdiskM:N-1  
Adds subdisks to the ends of the  
columns in a striped or RAID-5  
volume.  
See “Associating subdisks with  
plexes” on page 218.  
Example:  
# vxsd -g mydg assoc \  
vol01-01 mydg10-01:0 \  
mydg11-01:1 mydg12-01:2  
vxsd [-g diskgroup] mv oldsubdisk \  
newsubdisk ...  
Replaces a subdisk.  
See “Moving subdisks” on page 217.  
Example:  
# vxsd -g mydg mv mydg01-01 \  
mydg02-01  
vxsd [-g diskgroup] -s size split \  
subdisk sd1 sd2  
Splits a subdisk in two.  
See “Splitting subdisks” on page 217.  
Example:  
# vxsd -g mydg -s 1000m \  
split mydg03-02 mydg03-02 \  
mydg03-03  
vxsd [-g diskgroup] join sd1 sd2 ... \  
subdisk  
Joins two or more subdisks.  
See “Joining subdisks” on page 218.  
Example:  
# vxsd -g mydg join \  
mydg03-02 mydg03-03 \  
mydg03-02  
vxassist [-g diskgroup] move \  
volume !olddisk newdisk  
Relocates subdisks in a volume  
between disks.  
See “Moving and unrelocating  
subdisks using vxassist” on page 392.  
Example:  
# vxassist -g mydg move \  
myvol !mydg02 mydg05  
Download from Www.Somanuals.com. All Manuals Search And Download.  
496 Commands summary  
Table A-4  
Command  
Creating and administering subdisks  
Description  
vxunreloc [-g diskgroup] original_disk  
Relocates subdisks to their original  
disks.  
See “Moving and unrelocating  
subdisks using vxunreloc” on  
page 392.  
Example:  
# vxunreloc -g mydg mydg01  
vxsd [-g diskgroup] dis subdisk  
Dissociates a subdisk from a plex.  
See “Dissociating subdisks from  
plexes” on page 221.  
Example:  
# vxsd -g mydg dis mydg02-01  
vxedit [-g diskgroup] rm subdisk  
Removes a subdisk.  
See “Removing subdisks” on page 221.  
Example:  
# vxedit -g mydg rm mydg02-01  
vxsd [-g diskgroup] -o rm dis subdisk  
Dissociates and removes a subdisk  
from a plex.  
See “Dissociating subdisks from  
plexes” on page 221.  
Example:  
# vxsd -g mydg -o rm dis \  
mydg02-01  
Table A-5  
Command  
Creating and administering plexes  
Description  
vxmake [-g diskgroup] plex plex \  
sd=subdisk1[,subdisk2,...]  
Creates a concatenated plex.  
See “Creating plexes” on page 223.  
Example:  
# vxmake -g mydg plex \  
vol01-02 \  
sd=mydg02-01,mydg02-02  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Commands summary 497  
Table A-5  
Command  
Creating and administering plexes  
Description  
vxmake [-g diskgroup] plex plex \  
layout=stripe|raid5stwidth=W \  
ncolumn=N sd=subdisk1[,subdisk2,...]  
Creates a striped or RAID-5 plex.  
See “Creating a striped plex” on  
page 224.  
Example:  
# vxmake -g mydg plex pl-01 \  
layout=stripe stwidth=32 \  
ncolumn=2 \  
sd=mydg01-01,mydg02-01  
vxplex [-g diskgroup] att volume plex  
Attaches a plex to an existing volume.  
See “Attaching and associating  
plexes” on page 229.  
See “Reattaching plexes” on page 231.  
Example:  
# vxplex -g mydg att vol01 \  
vol01-02  
vxplex [-g diskgroup] det plex  
vxmend [-g diskgroup] off plex  
Detaches a plex.  
See “Detaching plexes” on page 231.  
Example:  
# vxplex -g mydg det vol01-02  
Takes a plex offline for maintenance.  
See “Taking plexes offline” on  
page 230.  
Example:  
# vxmend -g mydg off vol02-02  
vxmend [-g diskgroup] on plex  
Re-enables a plex for use.  
See “Reattaching plexes” on page 231.  
Example:  
# vxmend -g mydg on vol02-02  
vxplex [-g diskgroup] mv oldplex newplex Replaces a plex.  
See “Moving plexes” on page 232.  
Example:  
# vxplex -g mydg mv \  
vol02-02 vol02-03  
Download from Www.Somanuals.com. All Manuals Search And Download.  
498 Commands summary  
Table A-5  
Command  
Creating and administering plexes  
Description  
vxplex [-g diskgroup] cp volume newplex  
vxmend [-g diskgroup] fix clean plex  
vxplex [-g diskgroup] -o rm dis plex  
Copies a volume onto a plex.  
See “Copying volumes to plexes” on  
page 233.  
Example:  
# vxplex -g mydg cp vol02 \  
vol03-01  
Sets the state of a plex in an  
unstartable volume to CLEAN.  
See “Reattaching plexes” on page 231.  
Example:  
# vxmend -g mydg fix clean \  
vol02-02  
Dissociates and removes a plex from a  
volume.  
See “Dissociating and removing  
plexes” on page 233.  
Example:  
# vxplex -g mydg -o rm dis \  
vol03-01  
Table A-6  
Command  
Creating volumes  
Description  
vxassist [-g diskgroup] maxsize \  
layout=layout [attributes]  
Displays the maximum size of volume  
that can be created.  
See “Discovering the maximum size of  
a volume” on page 242.  
Example:  
# vxassist -g mydg maxsize \  
layout=raid5 nlog=2  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Commands summary 499  
Table A-6  
Command  
Creating volumes  
Description  
vxassist -b [-g diskgroup] make \  
Creates a volume.  
volume length [layout=layout ] [attributes]  
See “Creating a volume on any disk”  
on page 243.  
See “Creating a volume on specific  
disks” on page 244.  
Example:  
# vxassist -b -g mydg make \  
myvol 20g layout=concat \  
mydg01 mydg02  
vxassist -b [-g diskgroup] make \  
volume length layout=mirror \  
[nmirror=N] [attributes]  
Creates a mirrored volume.  
See “Creating a mirrored volume” on  
page 249.  
Example:  
# vxassist -b -g mydg make \  
mymvol 20g layout=mirror \  
nmirror=2  
vxassist -b [-g diskgroup] make \  
volume length layout=layout \  
exclusive=on [attributes]  
Creates a volume that may be opened  
exclusively by a single node in a  
cluster.  
See “Creating volumes with exclusive  
open access by a node” on page 426.  
Example:  
# vxassist -b -g mysdg make \  
mysmvol 20g layout=mirror \  
exclusive=on  
vxassist -b [-g diskgroup] make \  
Creates a striped or RAID-5 volume.  
volume length layout={stripe|raid5} \ See “Creating a striped volume” on  
[stripeunit=W] [ncol=N] [attributes]  
page 253.  
See“Creating a RAID-5 volume” on  
page 256.  
Example:  
# vxassist -b -g mydg make \  
mysvol 20g layout=stripe \  
stripeunit=32 ncol=4  
Download from Www.Somanuals.com. All Manuals Search And Download.  
500 Commands summary  
Table A-6  
Command  
Creating volumes  
Description  
vxassist -b [-g diskgroup] make \  
volume length layout=mirror \  
mirror=ctlr [attributes]  
Creates a volume with mirrored data  
plexes on separate controllers.  
See “Mirroring across targets,  
controllers or enclosures” on  
page 255.  
Example:  
# vxassist -b -g mydg make \  
mymcvol 20g layout=mirror \  
mirror=ctlr  
vxmake -b [-g diskgroup] -Uusage_type \ Creates a volume from existing plexes.  
vol volume [len=length] plex=plex,...  
See “Creating a volume using vxmake”  
on page 258.  
Example:  
# vxmake -g mydg -Uraid5 \  
vol r5vol \  
plex=raidplex,raidlog1,\  
raidlog2  
vxvol [-g diskgroup] start volume  
Initializes and starts a volume for use.  
See “Initializing and starting a  
volume” on page 260.  
See “Starting a volume” on page 271.  
Example:  
# vxvol -g mydg start r5vol  
vxvol [-g diskgroup] init zero volume  
Initializes and zeros out a volume for  
use.  
See “Initializing and starting a  
volume” on page 260.  
Example:  
# vxvol -g mydg init zero \  
myvol  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Commands summary 501  
Table A-7  
Command  
Administering volumes  
Description  
vxassist [-g diskgroup] mirror volume \ Adds a mirror to a volume.  
[attributes]  
See “Adding a mirror to a volume” on  
page 271.  
Example:  
# vxassist -g mydg mirror \  
myvol mydg10  
vxassist [-g diskgroup] remove \  
mirror volume [attributes]  
Removes a mirror from a volume.  
See “Removing a mirror” on page 273.  
Example:  
# vxassist -g mydg remove \  
mirror myvol !mydg11  
vxassist [-g diskgroup] \  
{growto|growby} volume length  
Grows a volume to a specified size or  
by a specified amount.  
See “Resizing volumes using vxassist”  
on page 286.  
Example:  
# vxassist -g mydg growby \  
myvol 10g  
vxassist [-g diskgroup] \  
{shrinkto|shrinkby} volume length  
Shrinks a volume to a specified size or  
by a specified amount.  
See “Resizing volumes using vxassist”  
on page 286.  
Example:  
# vxassist -g mydg shrinkto \  
myvol 20g  
vxresize -b -F vxfs [-g diskgroup] \  
volume length diskname ...  
Resizes a volume and the underlying  
Veritas File System.  
See “Resizing volumes using vxresize”  
on page 285.  
Example:  
# vxassist -b -F vxfs \  
-g mydg myvol 20g mydg10 \  
mydg11  
Download from Www.Somanuals.com. All Manuals Search And Download.  
502 Commands summary  
Table A-7  
Command  
Administering volumes  
Description  
vxsnap [-g diskgroup] prepare volume \  
[drl=on|sequential|off]  
Prepares a volume for instant  
snapshots and for DRL logging.  
See “Preparing a volume for DRL and  
instant snapshots” on page 275.  
Example:  
# vxsnap -g mydg prepare \  
myvol drl=on  
vxsnap [-g diskgroup] make \  
source=volume/newvol=snapvol\  
[/nmirror=number]  
Takes a full-sized instant snapshot of  
a volume by breaking off plexes of the  
original volume.  
See “Creating instant snapshots” on  
page 319.  
Example:  
# vxsnap -g mydg make \  
source=myvol/\  
newvol=mysnpvol/\  
nmiror=2  
vxsnap [-g diskgroup] make \  
source=volume/snapvol=snapvol  
Takes a full-sized instant snapshot of  
a volume using a prepared empty  
volume.  
See “Creating a volume for use as a  
full-sized instant or linked break-off  
snapshot” on page 323.  
See “Creating instant snapshots” on  
page 319.  
Example:  
# vxsnap -g mydg make \  
source=myvol/snapvol=snpvol  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Commands summary 503  
Table A-7  
Command  
Administering volumes  
Description  
vxmake [-g diskgroup] cache \  
cache_object cachevolname=volume \  
[regionsize=size]  
Creates a cache object for use by  
space-optimized instant snapshots.  
See “Creating a shared cache object”  
on page 322.  
A cache volume must have already  
been created, as shown in this  
example:  
# vxassist -g mydg make \  
cvol 1g layout=mirror \  
init=active mydg16 mydg17  
# vxmake -g mydg cache cobj \  
cachevolname=cvol  
vxsnap [-g diskgroup] make \  
source=volume/newvol=snapvol\  
/cache=cache_object  
Takes a space-optimized instant  
snapshot of a volume.  
See “Creating instant snapshots” on  
page 319.  
Example:  
# vxsnap -g mydg make \  
source=myvol/\  
newvol=mysosvol/\  
cache=cobj  
vxsnap [-g diskgroup] refresh snapshot  
Refreshes a snapshot from its original  
volume.  
See “Refreshing an instant snapshot”  
on page 337.  
Example:  
# vxsnap -g mydg refresh \  
mysnpvol  
vxsnap [-g diskgroup] dis snapshot  
Turns a snapshot into an independent  
volume.  
See “Dissociating an instant  
snapshot” on page 340.  
Example:  
# vxsnap -g mydg dis mysnpvol  
Download from Www.Somanuals.com. All Manuals Search And Download.  
504 Commands summary  
Table A-7  
Command  
Administering volumes  
Description  
vxsnap [-g diskgroup] unprepare volume Removes support for instant  
snapshots and DRL logging from a  
volume.  
See “Removing support for DRL and  
instant snapshots from a volume” on  
page 279.  
Example:  
# vxsnap -g mydg unprepare \  
myvol  
vxassist [-g diskgroup] relayout \  
volume [layout=layout] [relayout_options]  
Performs online relayout of a volume.  
See “Performing online relayout” on  
page 294.  
Example:  
# vxassist -g mydg relayout \  
vol2 layout=stripe  
vxassist [-g diskgroup] relayout \  
volume layout=raid5 stripeunit=W \  
ncol=N  
Relays out a volume as a RAID-5  
volume with stripe width W and N  
columns.  
See “Performing online relayout” on  
page 294.  
Example:  
# vxassist -g mydg relayout \  
vol3 layout=raid5 \  
stripeunit=16 ncol=4  
vxrelayout [-g diskgroup] -o bg \  
reverse volume  
Reverses the direction of a paused  
volume relayout.  
See “Controlling the progress of a  
relayout” on page 299.  
Example:  
# vxrelayout -g mydg -o bg \  
reverse vol3  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Commands summary 505  
Table A-7  
Command  
Administering volumes  
Description  
vxassist [-g diskgroup] convert \  
volume [layout=layout] [convert_options]  
Converts between a layered volume  
and a non-layered volume layout.  
See “Converting between layered and  
non-layered volumes” on page 300.  
Example:  
# vxassist -g mydg convert \  
vol3 layout=stripe-mirror  
vxassist [-g diskgroup] remove \  
volume volume  
Removes a volume.  
See “Removing a volume” on  
page 290.  
Example:  
# vxassist -g mydg remove \  
myvol  
Table A-8  
Command  
Monitoring and controlling tasks  
Description  
command [-g diskgroup] -t tasktag \  
[options] [arguments]  
Specifies a task tag to a VxVM  
command.  
See “Specifying task tags” on  
page 267.  
Example:  
# vxrecover -g mydg \  
-t mytask -b mydg05  
vxtask [-h] [-g diskgroup] list  
Lists tasks running on a system.  
See “Using the vxtask command” on  
page 269.  
Example:  
# vxtask -h -g mydg list  
vxtask monitor task  
Monitors the progress of a task.  
See “Using the vxtask command” on  
page 269.  
Example:  
# vxtask monitor mytask  
Download from Www.Somanuals.com. All Manuals Search And Download.  
506 Commands summary  
Table A-8  
Command  
Monitoring and controlling tasks  
Description  
vxtask pause task  
Suspends operation of a task.  
See “Using the vxtask command” on  
page 269.  
Example:  
# vxtask pause mytask  
vxtask -p [-g diskgroup] list  
Lists all paused tasks.  
See “Using the vxtask command” on  
page 269.  
Example:  
# vxtask -p -g mydg list  
vxtask resume task  
Resumes a paused task.  
See “Using the vxtask command” on  
page 269.  
Example:  
# vxtask resume mytask  
vxtask abort task  
Cancels a task and attempts to reverse  
its effects.  
See “Using the vxtask command” on  
page 269.  
Example:  
# vxtask abort mytask  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Commands summary 507  
Online manual pages  
Online manual pages  
Manual pages are organized into three sections:  
■
■
■
Section 1M — administrative commands  
Section 4 — file formats  
Section 7 — device driver interfaces  
Section 1M — administrative commands  
Manual pages in section 1M describe commands that are used to administer  
Veritas Volume Manager.  
Table A-9  
Section 1M manual pages  
Description  
Name  
dgcfgbackup  
dgcfgdaemon  
dgcfgrestore  
vgrestore  
Create or update VxVM volume group configuration backup file.  
Start the VxVM configuration backup daemon.  
Display or restore VxVM disk group configuration from backup.  
Restore a VxVM disk group back to an LVM volume group.  
Start Veritas Volume Manager from recovery media.  
vx_emerg_start  
vxassist  
Create, relayout, convert, mirror, backup, grow, shrink, delete,  
and move volumes.  
vxbootsetup  
Set up system boot information on a Veritas Volume Manager  
disk.  
vxbrk_rootmir  
Break off a mirror of a VxVM root disk to create a separate root  
disk generation.  
vxcache  
Administer the cache object for space-optimized snapshots.  
Resize cache volumes when required.  
vxcached  
vxcdsconvert  
vxchg_rootid  
Make disks and disk groups portable between systems.  
Set up VxVM root disk that has been cloned or copied from  
another operational VxVM root disk.  
vxclustadm  
Start, stop, and reconfigure a cluster.  
Administer command logging.  
vxcmdlog  
vxconfigbackup  
vxconfigbackupd  
Back up disk group configuration.  
Disk group configuration backup daemon.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
508 Commands summary  
Online manual pages  
Table A-9  
Section 1M manual pages  
Name  
Description  
vxconfigd  
Veritas Volume Manager configuration daemon  
Restore disk group configuration.  
vxconfigrestore  
vxcp_lvmroot  
Copy LVM root disk onto new Veritas Volume Manager root  
disk.  
vxdarestore  
vxdco  
Restore simple or nopriv disk access records.  
Perform operations on version 0 DCO objects and DCO volumes.  
Control the volume configuration daemon.  
vxdctl  
vxddladm  
Device Discovery Layer subsystem administration.  
vxdestroy_lvmroot Remove LVM root disk and associated LVM volume group.  
vxdg  
Manage Veritas Volume Manager disk groups.  
Define and manage Veritas Volume Manager disks.  
Add one or more disks for use with Veritas Volume Manager.  
Menu-driven Veritas Volume Manager disk administration.  
Configure a disk for use with Veritas Volume Manager.  
Deconfigure a disk from use with Veritas Volume Manager.  
DMP subsystem administration.  
vxdisk  
vxdiskadd  
vxdiskadm  
vxdisksetup  
vxdiskunsetup  
vxdmpadm  
vxdmpinq  
vxedit  
Display SCSI inquiry data.  
Create, remove, and modify Veritas Volume Manager records.  
Evacuate all volumes from a disk.  
vxevac  
vximportdg  
Import a disk group into the Veritas Volume Manager  
configuration.  
vxinfo  
Print accessibility and usability of volumes.  
vxinstall  
vxintro  
vxiod  
Menu-driven Veritas Volume Manager initial configuration.  
Introduction to the Veritas Volume Manager utilities.  
Start, stop, and report on Veritas Volume Manager kernel I/O  
threads.  
vxmake  
Create Veritas Volume Manager configuration records.  
Display memory statistics for Veritas Volume Manager.  
vxmemstat  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Commands summary 509  
Online manual pages  
Table A-9  
Section 1M manual pages  
Name  
Description  
vxmend  
vxmirror  
vxnotify  
vxpfto  
vxplex  
vxpool  
vxprint  
Mend simple problems in configuration records.  
Mirror volumes on a disk or control default mirroring.  
Display Veritas Volume Manager configuration events.  
Set Powerfail Timeout (pfto).  
Perform Veritas Volume Manager operations on plexes.  
Create and administer ISP storage pools.  
Display records from the Veritas Volume Manager  
configuration.  
vxr5check  
vxreattach  
vxrecover  
vxrelayout  
vxrelocd  
Verify RAID-5 volume parity.  
Reattach disk drives that have become accessible again.  
Perform volume recovery operations.  
Convert online storage from one layout to another.  
Monitor Veritas Volume Manager for failure events and relocate  
failed subdisks.  
vxres_lvmroot  
vxresize  
vxrootmir  
vxsd  
Restore LVM root disk from Veritas Volume Manager root disk.  
Change the length of a volume containing a file system.  
Create a mirror of a Veritas Volume Manager root disk.  
Perform Veritas Volume Manager operations on subdisks.  
Storage Expert rules.  
vxse  
vxsnap  
Enable DRL on a volume, and create and administer instant  
snapshots.  
vxsplitlines  
vxstat  
Show disks with conflicting configuration copies in a cluster.  
Veritas Volume Manager statistics management utility.  
List and administer Veritas Volume Manager tasks.  
Install and administer ISP volume templates and template sets.  
Trace operations on volumes.  
vxtask  
vxtemplate  
vxtrace  
vxtranslog  
Administer transaction logging.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
510 Commands summary  
Online manual pages  
Table A-9  
Section 1M manual pages  
Name  
Description  
vxtune  
Adjust Veritas Volume Replicator and Veritas Volume Manager  
tunables.  
vxunreloc  
Move a hot-relocated subdisk back to its original disk.  
Create and administer ISP user templates.  
vxusertemplate  
vxvmboot  
Prepare Veritas Volume Manager volume as a root, boot,  
primary swap or dump volume.  
vxvmconvert  
vxvol  
Convert LVM volume groups to VxVM disk groups.  
Perform Veritas Volume Manager operations on volumes.  
vxvoladm  
Create and administer ISP application volumes on allocated  
storage.  
vxvoladmtask  
vxvset  
Administer ISP tasks.  
Create and administer volume sets.  
Section 4 — file formats  
Manual pages in section 4 describe the format of files that are used by Veritas  
Volume Manager.  
Table A-10  
Section 4 manual pages  
Name  
Description  
vol_pattern  
vxmake  
Disk group search specifications.  
vxmakedescription file.  
Section 7 — device driver interfaces  
Manual pages in section 7 describe the interfaces to Veritas Volume Manager  
devices.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Appendix  
B
Configuring Veritas  
Volume Manager  
This appendix provides guidelines for setting up efficient storage management  
after installing the Veritas Volume Manager software.  
This chapter describes:  
■
■
■
■
■
■
■
■
Setup tasks after installation  
Adding unsupported disk arrays as JBODs  
Adding foreign devices  
Adding disks to disk groups  
Guidelines for configuring storage  
Controlling VxVM’s view of multipathed devices  
Configuring cluster support  
Reconfiguration tasks  
Setup tasks after installation  
The setup sequence listed below is a typical example. Your system requirements  
may differ.  
Initial Setup Tasks  
■
■
Create disk groups by placing disks under Veritas Volume Manager control.  
If you intend to use the Intelligent Storage Provisioning (ISP) feature, create  
storage pools within the disk groups.  
■
■
Create volumes in the disk groups.  
Configure file systems on the volumes.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
512 Configuring Veritas Volume Manager  
Adding unsupported disk arrays as JBODs  
Optional Setup Tasks  
■
Place the rootdisk under VxVM control and mirror it to create an alternate  
boot disk.  
■
■
■
Designate hot-relocation spare disks in each disk group.  
Add mirrors to volumes.  
Configure DRL and FastResync on volumes.  
Maintenance Tasks  
■
■
■
Resize volumes and file systems.  
Add more disks, create new disk groups, and create new volumes.  
Create and maintain snapshots.  
Adding unsupported disk arrays as JBODs  
After installation, add any disk arrays that are unsupported by Symantec to the  
DISKS (JBOD) category as described in “Administering the Device Discovery  
Layer” on page 85.  
Adding foreign devices  
The device discovery feature of VxVM can discover some devices that are  
controlled by third-party drivers, such as for EMC PowerPath. For these devices  
it may be preferable to use the multipathing capability that is provided by the  
third-party drivers rather than using the Dynamic Multipathing (DMP) feature.  
Provided that a suitable array support library is available, DMP can co-exist with  
such drivers. Other foreign devices, for which a compatible ASL does not exist,  
can be made available to Veritas Volume Manager as simple disks by using the  
vxddladmaddforeigncommand. This also has the effect of bypassing DMP.  
Refer to “Administering the Device Discovery Layer” on page 85 for more  
information.  
Adding disks to disk groups  
To place disks in disk groups, use VEA or the vxdiskadmprogram after  
completing the installation. Refer to “Adding a disk to VxVM” on page 97 and  
the VEA online help for information on how to add your disks to new disk  
groups.  
See the Veritas Storage Foundation Intelligent Storage Provisioning  
Administrator’s Guide for information about creating storage pools within disk  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Configuring Veritas Volume Manager 513  
Guidelines for configuring storage  
groups. Storage pools are only required if you intend using the ISP feature of  
VxVM.  
Guidelines for configuring storage  
A disk failure can cause loss of data on the failed disk and loss of access to your  
system. Loss of access is due to the failure of a key disk used for system  
operations. Veritas Volume Manager can protect your system from these  
problems.  
To maintain system availability, data important to running and booting your  
system must be mirrored. The data must be preserved so it can be used in case of  
failure.  
The following are suggestions for protecting your system and data:  
■
Perform regular backups to protect your data. Backups are necessary if all  
copies of a volume are lost or corrupted. Power surges can damage several  
(or all) disks on your system. Also, typing a command in error can remove  
critical files or damage a file system directly. Performing regular backups  
ensures that lost or corrupted data is available to be retrieved.  
■
Place the disk containing the rootfile system (the root or boot disk) under  
Veritas Volume Manager control. Mirror the root disk so that an alternate  
root disk exists for booting purposes. By mirroring disks critical to booting,  
you ensure that no single disk failure leaves your system unbootable and  
unusable. For more information, see “Rootability” on page 102.  
■
■
■
■
Use mirroring to protect data against loss from a disk failure. See “Mirroring  
guidelines” on page 514 for details.  
Use the DRL feature to speed up recovery of mirrored volumes after a  
system crash. See “Dirty region logging guidelines” on page 515 for details.  
Use striping to improve the I/O performance of volumes. See “Striping  
guidelines” on page 515 for details.  
Make sure enough disks are available for a combined striped and mirrored  
configuration. At least two disks are required for the striped plex, and one or  
more additional disks are needed for the mirror.  
■
■
When combining striping and mirroring, never place subdisks from one plex  
on the same physical disk as subdisks from the other plex.  
Use logging to prevent corruption of recovery data in RAID-5 volumes. Make  
sure that each RAID-5 volume has at least one log plex. See “RAID-5  
guidelines” on page 516 for details.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
514 Configuring Veritas Volume Manager  
Guidelines for configuring storage  
■
Leave the Veritas Volume Manager hot-relocation feature enabled. See “Hot-  
relocation guidelines” on page 516 for details.  
Mirroring guidelines  
Refer to the following guidelines when using mirroring.  
■
Do not place subdisks from different plexes of a mirrored volume on the  
same physical disk. This action compromises the availability benefits of  
mirroring and degrades performance. Using the vxassistor vxdiskadm  
commands precludes this from happening.  
■
To provide optimum performance improvements through the use of  
mirroring, at least 70 percent of physical I/O operations should be read  
operations. A higher percentage of read operations results in even better  
performance. Mirroring may not provide a performance increase or may  
even result in a performance decrease in a write-intensive workload  
environment.  
Note: The operating system implements a file system cache. Read requests can  
frequently be satisfied from the cache. This can cause the read/write ratio for  
physical I/O operations through the file system to be biased toward writing  
(when compared to the read/write ratio at the application level).  
■
Where possible, use disks attached to different controllers when mirroring  
or striping. Most disk controllers support overlapped seeks. This allows  
seeks to begin on two disks at once. Do not configure two plexes of the same  
volume on disks that are attached to a controller that does not support  
overlapped seeks. This is important for older controllers or SCSI disks that  
do not cache on the drive. It is less important for modern SCSI disks and  
controllers. Mirroring across controllers allows the system to survive a  
failure of one of the controllers. Another controller can continue to provide  
data from a mirror.  
■
A plex exhibits superior performance when striped or concatenated across  
multiple disks, or when located on a much faster device. Set the read policy  
to prefer the faster plex. By default, a volume with one striped plex is  
configured to prefer reading from the striped plex.  
For more information, see “Mirroring (RAID-1)” on page 42.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Configuring Veritas Volume Manager 515  
Guidelines for configuring storage  
Dirty region logging guidelines  
Dirty region logging (DRL) can speed up recovery of mirrored volumes following  
a system crash. When DRL is enabled, Veritas Volume Manager keeps track of  
the regions within a volume that have changed as a result of writes to a plex.  
Note: Using Dirty Region Logging can impact system performance in a write-  
intensive environment.  
For more information, see “Dirty region logging” on page 60.  
Striping guidelines  
Refer to the following guidelines when using striping.  
■
Do not place more than one column of a striped plex on the same physical  
disk.  
■
Calculate stripe-unit sizes carefully. In general, a moderate stripe-unit size  
(for example, 64 kilobytes, which is also the default used by vxassist) is  
recommended.  
■
If it is not feasible to set the stripe-unit size to the track size, and you do not  
know the application I/O pattern, use the default stripe-unit size.  
Note: Many modern disk drives have variable geometry. This means that the  
track size differs between cylinders, so that outer disk tracks have more sectors  
than inner tracks. It is therefore not always appropriate to use the track size as  
the stripe-unit size. For these drives, use a moderate stripe-unit size (such as 64  
kilobytes), unless you know the I/O pattern of the application.  
■
■
■
Volumes with small stripe-unit sizes can exhibit poor sequential I/O latency  
if the disks do not have synchronized spindles. Generally, striping over disks  
without synchronized spindles yields better performance when used with  
larger stripe-unit sizes and multi-threaded, or largely asynchronous,  
random I/O streams.  
Typically, the greater the number of physical disks in the stripe, the greater  
the improvement in I/O performance; however, this reduces the effective  
mean time between failures of the volume. If this is an issue, combine  
striping with mirroring to combine high-performance with improved  
reliability.  
If only one plex of a mirrored volume is striped, set the policy of the volume  
to preferfor the striped plex. (The default read policy, select, does this  
automatically.)  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
516 Configuring Veritas Volume Manager  
Guidelines for configuring storage  
■
■
■
If more than one plex of a mirrored volume is striped, configure the same  
stripe-unit size for each striped plex.  
Where possible, distribute the subdisks of a striped volume across drives  
connected to different controllers and buses.  
Avoid the use of controllers that do not support overlapped seeks. (Such  
controllers are rare.)  
The vxassistcommand automatically applies and enforces many of these  
rules when it allocates space for striped plexes in a volume.  
For more information, see “Striping (RAID-0)” on page 38.  
RAID-5 guidelines  
Refer to the following guidelines when using RAID-5.  
In general, the guidelines for mirroring and striping together also apply to  
RAID-5. The following guidelines should also be observed with RAID-5:  
■
Only one RAID-5 plex can exist per RAID-5 volume (but there can be  
multiple log plexes).  
■
The RAID-5 plex must be derived from at least three subdisks on three or  
more physical disks. If any log plexes exist, they must belong to disks other  
than those used for the RAID-5 plex.  
■
■
RAID-5 logs can be mirrored and striped.  
If the volume length is not explicitly specified, it is set to the length of any  
RAID-5 plex associated with the volume; otherwise, it is set to zero. If you  
specify the volume length, it must be a multiple of the stripe-unit size of the  
associated RAID-5 plex, if any.  
■
If the log length is not explicitly specified, it is set to the length of the  
smallest RAID-5 log plex that is associated, if any. If no RAID-5 log plexes  
are associated, it is set to zero.  
■
■
Sparse RAID-5 log plexes are not valid.  
RAID-5 volumes are not supported for sharing in a cluster.  
For more information, see “RAID-5 (striping with parity)” on page 45.  
Hot-relocation guidelines  
Hot-relocation automatically restores redundancy and access to mirrored and  
RAID-5 volumes when a disk fails. This is done by relocating the affected  
subdisks to disks designated as spares and/or free space in the same disk group.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Configuring Veritas Volume Manager 517  
Guidelines for configuring storage  
The hot-relocation feature is enabled by default. The associated daemon,  
vxrelocd, is automatically started during system startup.  
Refer to the following guidelines when using hot-relocation.  
■
The hot-relocation feature is enabled by default. Although it is possible to  
disable hot-relocation, it is advisable to leave it enabled. It will notify you of  
the nature of the failure, attempt to relocate any affected subdisks that are  
redundant, and initiate recovery procedures.  
■
Although hot-relocation does not require you to designate disks as spares,  
designate at least one disk as a spare within each disk group. This gives you  
some control over which disks are used for relocation. If no spares exist,  
Veritas Volume Manager uses any available free space within the disk group.  
When free space is used for relocation purposes, it is possible to have  
performance degradation after the relocation.  
■
■
■
After hot-relocation occurs, designate one or more additional disks as spares  
to augment the spare space. Some of the original spare space may be  
occupied by relocated subdisks.  
If a given disk group spans multiple controllers and has more than one spare  
disk, set up the spare disks on different controllers (in case one of the  
controllers fails).  
For a mirrored volume, configure the disk group so that there is at least one  
disk that does not already contain a mirror of the volume. This disk should  
either be a spare disk with some available space or a regular disk with some  
free space and the disk is not excluded from hot-relocation use.  
■
For a mirrored and striped volume, configure the disk group so that at least  
one disk does not already contain one of the mirrors of the volume or  
another subdisk in the striped plex. This disk should either be a spare disk  
with some available space or a regular disk with some free space and the  
disk is not excluded from hot-relocation use.  
■
For a RAID-5 volume, configure the disk group so that at least one disk does  
not already contain the RAID-5 plex (or one of its log plexes) of the volume.  
This disk should either be a spare disk with some available space or a  
regular disk with some free space and the disk is not excluded from hot-  
relocation use.  
■
■
If a mirrored volume has a DRL log subdisk as part of its data plex, you  
cannot relocate the data plex. Instead, place log subdisks in log plexes that  
contain no data.  
Hot-relocation does not guarantee to preserve the original performance  
characteristics or data layout. Examine the locations of newly-relocated  
Download from Www.Somanuals.com. All Manuals Search And Download.  
518 Configuring Veritas Volume Manager  
Controlling VxVM’s view of multipathed devices  
subdisks to determine whether they should be relocated to more suitable  
disks to regain the original performance benefits.  
■
Although it is possible to build Veritas Volume Manager objects on spare  
disks (using vxmakeor the VEA interface), it is recommended that you use  
spare disks for hot-relocation only.  
See “Administering hot-relocation” on page 379 for more information.  
Accessing volume devices  
As soon as a volume has been created and initialized, it is available for use as a  
virtual disk partition by the operating system for the creation of a file system, or  
by application programs such as relational databases and other data  
management software.  
Creating a volume in a disk group sets up block and character (raw) device files  
that can be used to access the volume:  
/dev/vx/dsk/diskgroup/volume  
/dev/vx/rdsk/diskgroup/volume  
block device file for volume  
character device file for volume  
The pathnames include a directory named for the disk group. Use the  
appropriate device node to create, mount and repair file systems, and to lay out  
databases that require raw partitions.  
Controlling VxVM’s view of multipathed devices  
To control how a device is treated by the Dynamic Multipathing (DMP) feature of  
VxVM, use the vxdiskadmcommand as described in “Disabling and enabling  
multipathing for specific devices” on page 133.  
Configuring cluster support  
The Veritas Volume Manager software includes an optional cluster feature that  
enables it to be used in a cluster environment. The cluster functionality in  
Veritas Volume Manager allows multiple hosts to simultaneously access and  
manage a set of disks under Veritas Volume Manager control. A cluster is a set of  
hosts sharing a set of disks; each host is referred to as a node in the cluster.  
Note: The Veritas Volume Manager cluster feature requires a license, which can  
be obtained from your Customer Support channel.  
For information about enabling cluster functionality in Veritas Volume  
Manager, refer to the Veritas Storage Solutions Getting Started Guide.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Configuring Veritas Volume Manager 519  
Configuring cluster support  
Configuring shared disk groups  
This section describes how to configure shared disks in a cluster. If you are  
installing Veritas Volume Manager for the first time or adding disks to an  
existing cluster, you need to configure new shared disks.  
If you are setting up Veritas Volume Manager for the first time, configure the  
shared disks using the following procedure:  
1
2
Start the cluster on one node only to prevent access by other nodes.  
On one node, run the vxdiskadmprogram and choose option 1 to initialize  
new disks. When asked to add these disks to a disk group, choose noneto  
leave the disks for future use.  
3
4
On other nodes in the cluster, run vxdctlenableto see the newly initialized  
disks.  
From the master node, create disk groups on the shared disks. To determine  
if a node is a master or slave, run the command vxdctl -c mode.  
Use the vxdgcommand or VEA to create disk groups. If you use the vxdg  
command, specify the -soption to create shared disk groups.  
5
6
From the master node only, use vxassistor VEA to create volumes in the  
disk groups.  
Note: RAID-5 volumes are not supported for sharing in a cluster.  
If the cluster is only running with one node, bring up the other cluster  
nodes. Enter the vxdglistcommand on each node to display the shared  
disk groups.  
Converting existing VxVM disk groups to shared disk groups  
To convert existing disk groups to shared disk groups:  
1
2
Start the cluster on one node only to prevent access by other nodes.  
Configure the disk groups using the following procedure.  
To list all disk groups, use the following command:  
# vxdg list  
To deport the disk groups that are to be shared, use the following command:  
# vxdg deport diskgroup  
To import disk groups to be shared, use the following command:  
# vxdg -s import diskgroup  
This procedure marks the disks in the shared disk groups as shared and  
stamps them with the ID of the cluster, enabling other nodes to recognize  
the shared disks.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
520 Configuring Veritas Volume Manager  
Reconfiguration tasks  
If dirty region logs exist, ensure they are active. If not, replace them with  
larger ones.  
To display the shared flag for all the shared disk groups, use the following  
command:  
# vxdg list  
The disk groups are now ready to be shared.  
3
Bring up the other cluster nodes. Enter the vxdglistcommand on each  
node to display the shared disk groups. This command displays the same list  
of shared disk groups displayed earlier.  
For information on converting disk groups in a Veritas Storage Foundation  
Cluster File System environment, see the Veritas Storage Foundation Cluster File  
System Installation Guide.  
Reconfiguration tasks  
The following sections describe tasks that allow you to make changes to the  
configuration that you specified during installation.  
Changing the name of the default disk group  
If you use the Veritas installer to install the Veritas Volume Manager software,  
you can enter the name of the default disk group. This disk group will be used by  
commands if you do not specify the -g option, or the VXVM_DEFAULTDG  
environment variable is not set. If required, you can use the vxdctldefaultdg  
command to change the default disk group. See “Displaying and specifying the  
system-wide default disk group” on page 168 for details.  
Enabling or disabling enclosure-based naming  
If you use the Veritas installer to install the Veritas Volume Manager software,  
you can choose whether the displayed disk access names are based on device  
names or on names that you assign to disk enclosures. If required, you can use  
the vxdiskadmor vxddladmcommands to change the naming convention that is  
used. See “Changing the disk-naming scheme” on page 91 for more information.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Glossary  
Active/Active disk arrays  
This type of multipathed disk array allows you to access a disk in the disk array through all  
the paths to the disk simultaneously, without any performance degradation.  
Active/Passive disk arrays  
This type of multipathed disk array allows one path to a disk to be designated as primary  
and used to access the disk at any time. Using a path other than the designated active path  
results in severe performance degradation in some disk arrays. Also see path, primary  
path, and secondary path.  
associate  
The process of establishing a relationship between VxVM objects; for example, a subdisk  
that has been created and defined as having a starting point within a plex is referred to as  
being associated with that plex.  
associated plex  
A plex associated with a volume.  
associated subdisk  
A subdisk associated with a plex.  
atomic operation  
An operation that either succeeds completely or fails and leaves everything as it was before  
the operation was started. If the operation succeeds, all aspects of the operation take effect  
at once and the intermediate states of change are invisible. If any aspect of the operation  
fails, then the operation aborts without leaving partial changes.  
In a cluster, an atomic operation takes place either on all nodes or not at all.  
attached  
A state in which a VxVM object is both associated with another object and enabled for use.  
block  
The minimum unit of data transfer to or from a disk or array.  
boot disk  
A disk that is used for the purpose of booting a system.  
boot disk group  
A private disk group that contains the disks from which the system may be booted.  
bootdg  
A reserved disk group name that is an alias for the name of the boot disk group.  
clean node shutdown  
The ability of a node to leave a cluster gracefully when all access to shared volumes has  
ceased.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
522 Glossary  
cluster  
A set of hosts (each termed a node) that share a set of disks.  
cluster manager  
An externally-provided daemon that runs on each node in a cluster. The cluster managers  
on each node communicate with each other and inform VxVM of changes in cluster  
membership.  
cluster-shareable disk group  
A disk group in which access to the disks is shared by multiple hosts (also referred to as a  
shared disk group). Also see private disk group.  
column  
A set of one or more subdisks within a striped plex. Striping is achieved by allocating data  
alternately and evenly across the columns within a plex.  
concatenation  
A layout style characterized by subdisks that are arranged sequentially and contiguously.  
configuration copy  
A single copy of a configuration database.  
configuration database  
A set of records containing detailed information on existing VxVM objects (such as disk  
and volume attributes).  
data change object (DCO)  
A VxVM object that is used to manage information about the FastResync maps in the DCO  
volume. Both a DCO object and a DCO volume must be associated with a volume to  
implement Persistent FastResync on that volume.  
data stripe  
This represents the usable data portion of a stripe and is equal to the stripe minus the  
parity region.  
DCO volume  
A special volume that is used to hold Persistent FastResync change maps, and dirty region  
logs (see dirty region logging).  
detached  
A state in which a VxVM object is associated with another object, but not enabled for use.  
device name  
The device name or address used to access a physical disk, such as c0t0d0. The c#t#d#  
syntax identifies the controller, target address, and disk. In a SAN environment, it is more  
convenient to use enclosure-based naming, which forms the device name by concatenating  
the name of the enclosure (such as enc0) with the disk’s number within the enclosure,  
separated by an underscore (for example, enc0_2). The term disk access name can also be  
used to refer to a device name.  
dirty region logging  
The method by which the VxVM monitors and logs modifications to a plex as a bitmap of  
changed regions. For a volumes with a new-style DCO volume, the dirty region log (DRL) is  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Glossary 523  
maintained in the DCO volume. Otherwise, the DRL is allocated to an associated subdisk  
called a log subdisk.  
disabled path  
A path to a disk that is not available for I/O. A path can be disabled due to real hardware  
failures or if the user has used the vxdmpadmdisablecommand on that controller.  
disk  
A collection of read/write data blocks that are indexed and can be accessed fairly quickly.  
Each disk has a universally unique identifier.  
disk access name  
An alternative term for a device name.  
disk access records  
Configuration records used to specify the access path to particular disks. Each disk access  
record contains a name, a type, and possibly some type-specific information, which is used  
by VxVM in deciding how to access and manipulate the disk that is defined by the disk  
access record.  
disk array  
A collection of disks logically arranged into an object. Arrays tend to provide benefits such  
as redundancy or improved performance. Also see disk enclosure and JBOD.  
disk array serial number  
This is the serial number of the disk array. It is usually printed on the disk array cabinet or  
can be obtained by issuing a vendor- specific SCSI command to the disks on the disk array.  
This number is used by the DMP subsystem to uniquely identify a disk array.  
disk controller  
In the multipathing subsystem of VxVM, the controller (host bus adapter or HBA) or disk  
array connected to the host, which the operating system represents as the parent node of a  
disk.  
disk enclosure  
An intelligent disk array that usually has a backplane with a built-in Fibre Channel loop,  
and which permits hot-swapping of disks.  
disk group  
A collection of disks that share a common configuration. A disk group configuration is a set  
of records containing detailed information on existing VxVM objects (such as disk and  
volume attributes) and their relationships. Each disk group has an administrator-assigned  
name and an internally defined unique ID. The disk group names bootdg(an alias for the  
boot disk group), defaultdg(an alias for the default disk group) and nodg(represents no  
disk group) are reserved.  
disk group ID  
A unique identifier used to identify a disk group.  
disk ID  
A universally unique identifier that is given to each disk and can be used to identify the  
disk, even if it is moved.  
disk media name  
Download from Www.Somanuals.com. All Manuals Search And Download.  
524 Glossary  
An alternative term for a disk name.  
disk media record  
A configuration record that identifies a particular disk, by disk ID, and gives that disk a  
logical (or administrative) name.  
disk name  
A logical or administrative name chosen for a disk that is under the control of VxVM, such  
as disk03. The term disk media name is also used to refer to a disk name.  
dissociate  
The process by which any link that exists between two VxVM objects is removed. For  
example, dissociating a subdisk from a plex removes the subdisk from the plex and adds  
the subdisk to the free space pool.  
dissociated plex  
A plex dissociated from a volume.  
dissociated subdisk  
A subdisk dissociated from a plex.  
distributed lock manager  
A lock manager that runs on different systems in a cluster, and ensures consistent access  
to distributed resources.  
enabled path  
A path to a disk that is available for I/O.  
encapsulation  
A process that converts existing partitions on a specified disk to volumes. Encapsulation  
does not apply to HP-UX.  
enclosure  
See disk enclosure.  
enclosure-based naming  
See device name.  
fabric mode disk  
A disk device that is accessible on a Storage Area Network (SAN) via a Fibre Channel  
switch.  
FastResync  
A fast resynchronization feature that is used to perform quick and efficient  
resynchronization of stale mirrors, and to increase the efficiency of the snapshot  
mechanism. Also see Persistent FastResync and Non-Persistent FastResync.  
Fibre Channel  
A collective name for the fiber optic technology that is commonly used to set up a Storage  
Area Network (SAN).  
file system  
A collection of files organized together into a structure. The UNIX file system is a  
hierarchical structure consisting of directories and files.  
free space  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Glossary 525  
An area of a disk under VxVM control that is not allocated to any subdisk or reserved for  
use by any other VxVM object.  
free subdisk  
A subdisk that is not associated with any plex and has an empty putil[0]field.  
hostid  
A string that identifies a host to VxVM. The hostid for a host is stored in its volboot file, and  
is used in defining ownership of disks and disk groups.  
hot-relocation  
A technique of automatically restoring redundancy and access to mirrored and RAID-5  
volumes when a disk fails. This is done by relocating the affected subdisks to disks  
designated as spares and/or free space in the same disk group.  
hot-swap  
Refers to devices that can be removed from, or inserted into, a system without first turning  
off the power supply to the system.  
initiating node  
The node on which the system administrator is running a utility that requests a change to  
VxVM objects. This node initiates a volume reconfiguration.  
JBOD  
The common name for an unintelligent disk array which may, or may not, support the hot-  
swapping of disks. The name is derived from “just a bunch of disks.”  
log plex  
A plex used to store a RAID-5 log. The term log plex may also be used to refer to a Dirty  
Region Logging plex.  
log subdisk  
A subdisk that is used to store a dirty region log.  
master node  
A node that is designated by the software to coordinate certain VxVM operations in a  
cluster. Any node is capable of being the master node.  
mastering node  
The node to which a disk is attached. This is also known as a disk owner.  
mirror  
A duplicate copy of a volume and the data therein (in the form of an ordered collection of  
subdisks). Each mirror consists of one plex of the volume with which the mirror is  
associated.  
mirroring  
A layout technique that mirrors the contents of a volume onto multiple plexes. Each plex  
duplicates the data stored on the volume, but the plexes themselves may have different  
layouts.  
multipathing  
Download from Www.Somanuals.com. All Manuals Search And Download.  
526 Glossary  
Where there are multiple physical access paths to a disk connected to a system, the disk is  
called multipathed. Any software residing on the host, (for example, the DMP driver) that  
hides this fact from the user is said to provide multipathing functionality.  
node  
One of the hosts in a cluster.  
node abort  
A situation where a node leaves a cluster (on an emergency basis) without attempting to  
stop ongoing operations.  
node join  
The process through which a node joins a cluster and gains access to shared disks.  
Non-Persistent FastResync  
A form of FastResync that cannot preserve its maps across reboots of the system because it  
stores its change map in memory.  
object  
An entity that is defined to and recognized internally by VxVM. The VxVM objects are:  
volume, plex, subdisk, disk, and disk group. There are actually two types of disk objects—  
one for the physical aspect of the disk and the other for the logical aspect.  
parity  
A calculated value that can be used to reconstruct data after a failure. While data is being  
written to a RAID-5 volume, parity is also calculated by performing an exclusive OR (XOR)  
procedure on data. The resulting parity is then written to the volume. If a portion of a  
RAID-5 volume fails, the data that was on that portion of the failed volume can be  
recreated from the remaining data and the parity.  
parity stripe unit  
A RAID-5 volume storage region that contains parity information. The data contained in  
the parity stripe unit can be used to help reconstruct regions of a RAID-5 volume that are  
missing because of I/O or disk failures.  
partition  
The standard division of a physical disk device, as supported directly by the operating  
system and disk drives.  
path  
When a disk is connected to a host, the path to the disk consists of the HBA (Host Bus  
Adapter) on the host, the SCSI or fibre cable connector and the controller on the disk or  
disk array. These components constitute a path to a disk. A failure on any of these results  
in DMP trying to shift all I/O for that disk onto the remaining (alternate) paths. Also see  
Active/Passive disk arrays, primary path and secondary path.  
pathgroup  
In the case of disks which are not multipathed by vxdmp, VxVM will see each path as a disk.  
In such cases, all paths to the disk can be grouped. This way only one of the paths from the  
group is made visible to VxVM.  
Persistent FastResync  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Glossary 527  
A form of FastResync that can preserve its maps across reboots of the system by storing its  
change map in a DCO volume on disk. Also see data change object (DCO).  
persistent state logging  
A logging type that ensures that only active mirrors are used for recovery purposes and  
prevents failed mirrors from being selected for recovery. This is also known as kernel  
logging.  
physical disk  
The underlying storage device, which may or may not be under VxVM control.  
plex  
A plex is a logical grouping of subdisks that creates an area of disk space independent of  
physical disk size or other restrictions. Mirroring is set up by creating multiple data plexes  
for a single volume. Each data plex in a mirrored volume contains an identical copy of the  
volume data. Plexes may also be created to represent concatenated, striped and RAID-5  
volume layouts, and to store volume logs.  
primary path  
In Active/Passive disk arrays, a disk can be bound to one particular controller on the disk  
array or owned by a controller. The disk can then be accessed using the path through this  
particular controller. Also see path and secondary path.  
private disk group  
A disk group in which the disks are accessed by only one specific host in a cluster. Also see  
shared disk group.  
private region  
A region of a physical disk used to store private, structured VxVM information. The private  
region contains a disk header, a table of contents, and a configuration database. The table  
of contents maps the contents of the disk. The disk header contains a disk ID. All data in  
the private region is duplicated for extra reliability.  
public region  
A region of a physical disk managed by VxVM that contains available space and is used for  
allocating subdisks.  
RAID  
A Redundant Array of Independent Disks (RAID) is a disk array set up with part of the  
combined storage capacity used for storing duplicate information about the data stored in  
that array. This makes it possible to regenerate the data if a disk failure occurs.  
read-writeback mode  
A recovery mode in which each read operation recovers plex consistency for the region  
covered by the read. Plex consistency is recovered by reading data from blocks of one plex  
and writing the data to all other writable plexes.  
root configuration  
The configuration database for the root disk group. This is special in that it always  
contains records for other disk groups, which are used for backup purposes only. It also  
contains disk records that define all disk devices on the system.  
root disk  
Download from Www.Somanuals.com. All Manuals Search And Download.  
528 Glossary  
The disk containing the root file system. This disk may be under VxVM control.  
root file system  
The initial file system mounted as part of the UNIX kernel startup sequence.  
root partition  
The disk region on which the root file system resides.  
root volume  
The VxVM volume that contains the root file system, if such a volume is designated by the  
system configuration.  
rootability  
The ability to place the rootfile system and the swapdevice under VxVM control. The  
resulting volumes can then be mirrored to provide redundancy and allow recovery in the  
event of disk failure.  
secondary path  
In Active/Passive disk arrays, the paths to a disk other than the primary path are called  
secondary paths. A disk is supposed to be accessed only through the primary path until it  
fails, after which ownership of the disk is transferred to one of the secondary paths. Also  
see path and primary path.  
sector  
A unit of size, which can vary between systems. Sector size is set per device (hard drive,  
CD-ROM, and so on). Although all devices within a system are usually configured to the  
same sector size for interoperability, this is not always the case. A sector is commonly 1024  
bytes.  
shared disk group  
A disk group in which access to the disks is shared by multiple hosts (also referred to as a  
cluster-shareable disk group). Also see private disk group.  
shared volume  
A volume that belongs to a shared disk group and is open on more than one node of a  
cluster at the same time.  
shared VM disk  
A VM disk that belongs to a shared disk group in a cluster.  
slave node  
A node that is not designated as the master node of a cluster.  
slice  
The standard division of a logical disk device. The terms partition and slice are sometimes  
used synonymously.  
snapshot  
A point-in-time copy of a volume (volume snapshot) or a file system (file system snapshot).  
spanning  
A layout technique that permits a volume (and its file system or database) that is too large  
to fit on a single disk to be configured across multiple physical disks.  
sparse plex  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Glossary 529  
A plex that is not as long as the volume or that has holes (regions of the plex that do not  
have a backing subdisk).  
Storage Area Network (SAN)  
A networking paradigm that provides easily reconfigurable connectivity between any  
subset of computers, disk storage and interconnecting hardware such as switches, hubs  
and bridges.  
stripe  
A set of stripe units that occupy the same positions across a series of columns.  
stripe size  
The sum of the stripe unit sizes comprising a single stripe across all columns being striped.  
stripe unit  
Equally-sized areas that are allocated alternately on the subdisks (within columns) of each  
striped plex. In an array, this is a set of logically contiguous blocks that exist on each disk  
before allocations are made from the next disk in the array. A stripe unit may also be  
referred to as a stripe element.  
stripe unit size  
The size of each stripe unit. The default stripe unit size is 64KB. The stripe unit size is  
sometimes also referred to as the stripe width.  
striping  
A layout technique that spreads data across several physical disks using stripes. The data  
is allocated alternately to the stripes within the subdisks of each plex.  
subdisk  
A consecutive set of contiguous disk blocks that form a logical disk segment. Subdisks can  
be associated with plexes to form volumes.  
swap area  
A disk region used to hold copies of memory pages swapped out by the system pager  
process.  
swap volume  
A VxVM volume that is configured for use as a swap area.  
transaction  
A set of configuration changes that succeed or fail as a group, rather than individually.  
Transactions are used internally to maintain consistent configurations.  
volboot file  
A small file that is used to locate copies of the boot disk group configuration. The file may  
list disks that contain configuration copies in standard locations, and can also contain  
direct pointers to configuration copy locations. The volbootfile is stored in a system-  
dependent location.  
VM disk  
A disk that is both under VxVM control and assigned to a disk group. VM disks are  
sometimes referred to as VxVM disks or simply disks.  
volume  
Download from Www.Somanuals.com. All Manuals Search And Download.  
530 Glossary  
A virtual disk, representing an addressable range of disk blocks used by applications such  
as file systems or databases. A volume is a collection of from one to 32 plexes.  
volume configuration device  
The volume configuration device (/dev/vx/config) is the interface through which all  
configuration changes to the volume device driver are performed.  
volume device driver  
The driver that forms the virtual disk drive between the application and the physical  
device driver level. The volume device driver is accessed through a virtual disk device node  
whose character device nodes appear in /dev/vx/rdsk, and whose block device nodes  
appear in /dev/vx/dsk.  
volume event log  
The device interface (/dev/vx/event) through which volume driver events are reported  
to utilities.  
vxconfigd  
The VxVM configuration daemon, which is responsible for making changes to the VxVM  
configuration. This daemon must be running before VxVM operations can be performed.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Index  
allsites attribute 436  
APM  
configuring 162  
application volumes 32  
array policy module (APM)  
configuring 162  
Symbols  
/dev/vx/dmp directory 126  
/dev/vx/rdmp directory 126  
/etc/default/vxassist file 241, 390  
/etc/default/vxdg defaults file 403  
/etc/default/vxdg file 171  
/etc/default/vxdisk file 81, 97  
/etc/default/vxse file 448  
/etc/fstab file 290  
/etc/volboot file 212  
/etc/vx/darecs file 212  
/etc/vx/disk.info file 93  
/etc/vx/dmppolicy.info file 148  
/etc/vx/volboot file 186  
array ports  
disabling for DMP 153  
displaying information about 142  
enabling for DMP 154  
array support library (ASL) 83  
ASL  
array support library 83  
Asymmetric Active/Active disk arrays 126  
ATTACHING state 311  
attributes  
/sbin/init.d/vxvm-recover file 395  
active 146  
autogrow 322, 325  
autogrowby 322  
cache 325  
cachesize 325  
comment 222, 234  
dcolen 69, 251, 357  
dgalign_checking 243  
displaying for rules 447  
drl 252, 281  
fastresync 251, 252, 293  
for specifying storage 244  
for Storage Expert 447  
hasdcolog 293  
A
A/A disk arrays 126  
A/A-A disk arrays 126  
A/P disk arrays 125  
A/P-C disk arrays 126  
A/PF disk arrays 126  
A/PF-C disk arrays 126  
A/PG disk arrays 126  
A/PG-C disk arrays 126  
access port 125  
activation modes for shared disk groups 402  
ACTIVE  
plex state 225  
highwatermark 322  
init 260  
len 222  
volume state 265  
active path attribute 146  
ACTIVE state 311  
listing for rules 447  
loglen 253  
logtype 253  
maxautogrow 322  
maxdev 189  
Active/Active disk arrays 126  
Active/Passive disk arrays 125  
adaptive load-balancing 148  
adaptiveminq load-balancing 148  
adding disks 101  
mirdg 332  
agile devices 77  
mirvol 331  
name 221, 234  
alignment constraints 242  
allocation  
ncachemirror 325  
site-based 432  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
532 Index  
ndcomirror 251, 252, 357  
ndcomirs 275, 321  
newvol 330  
cache objects  
creating 322  
enabling 323  
nmirror 330  
listing snapshots in 345  
nomanual 146  
nopreferred 146  
plex 234  
caches  
creating 322  
deleting 347  
preferred priority 146  
primary 147  
finding out snapshots configured on 347  
growing 347  
putil 222, 234  
listing snapshots in 345  
removing 347  
resizing 347  
shrinking 347  
stopping 347  
secondary 147  
sequential DRL 252  
setting for paths 146  
setting for rules 448  
snapvol 327, 332  
source 327, 332  
standby 147  
subdisk 221  
syncing 319, 344  
tutil 222, 234  
used by space-optimized instant  
snapshots 309  
cachesize attribute 325  
Campus Cluster feature  
administering 431  
campus clusters  
auto disk type 81  
autogrow  
tuning 346  
autogrow attribute 322, 325  
autogrowby attribute 322  
autotrespass mode 125  
administering 431  
serial split brain condition in 190  
cascade instant snapshots 312  
cascaded snapshot hierarchies  
creating 337  
categories  
disks 83  
CDS  
B
alignment constraints 242  
compatible disk groups 171  
disk format 81  
backups  
created using snapshots 319  
creating for volumes 303  
creating using instant snapshots 319  
creating using third-mirror snapshots 348  
for multiple volumes 333, 352  
implementing online 371  
of disk group configuration 213  
balanced path policy 148  
base minor number 187  
blocks on disks 29  
cds attribute 171  
cdsdisk format 81  
check_all policy 160  
check_alternate policy 160  
check_disabled policy 160  
check_periodic policy 160  
checkpoint interval 479  
CLEAN  
plex state 225  
volume state 265  
clone_disk flag 177  
cloned disks 176  
boot disk group 167  
bootdg 167  
BROKEN state 311  
cluster functionality  
enabling 518  
shared disks 519  
cluster protocol version  
checking 428  
C
c# 20, 78  
c#t#d# 78  
c#t#d# based naming scheme 78  
cache attribute 325  
upgrading 428  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Index 533  
clusters  
activating disk groups 403  
vxstat 428  
cluster-shareable disk groups in clusters 401  
columns  
activating shared disk groups 425  
activation modes for shared disk groups 402  
benefits 397  
checking cluster protocol version 427  
cluster-shareable disk groups 401  
configuration 410  
changing number of 298  
checking number in volume 454  
in striping 38  
mirroring in striped-mirror volumes 255  
comment  
configuring exclusive open of volume by  
node 426  
plex attribute 234  
subdisk attribute 222  
connectivity policies 404  
concatenated volumes 35, 236  
concatenated-mirror volumes  
converting shared disk groups to private 424  
creating shared disk groups 422  
designating shareable disk groups 401  
detach policies 404  
converting to mirrored-concatenated 300  
creating 249  
defined 45  
determining if disks are shared 421  
forcibly adding disks to disk groups 423  
forcibly importing disk groups 423  
importing disk groups as shared 423  
initialization 410  
recovery 237  
concatenation 35  
condition flags for plexes 228  
configuration backup and restoration 213  
configuration changes  
monitoring using vxnotify 213  
configuration copies for disk group 474  
configuration database  
checking number of copies 451  
checking size of 451  
introduced 398  
joining disk groups in 424  
limitations of shared disk groups 409  
listing shared disk groups 421  
maximum number of nodes in 397  
moving objects between disk groups 424  
node abort 417  
copy size 166  
in private region 80  
node shutdown 416  
nodes 399  
listing disks with 178  
metadata 177  
operation of vxconfigd in 414  
operation of VxVM in 398  
private disk groups 401  
reducing size of 195  
configuring  
shared disks 519  
private networks 399  
protection against simultaneous writes 402  
reconfiguration of 410  
connectivity policies 404  
setting for disk groups 425  
controllers  
resolving disk status in 404  
setting disk connectivity policies in 425  
setting failure policies in 426  
shared disk groups 401  
checking for disabled 454  
disabling for DMP 153  
disabling in DMP 136  
displaying information about 141  
enabling for DMP 154  
mirroring across 247, 255  
number 20  
shared objects 402  
splitting disk groups in 424  
upgrading cluster protocol version 428  
use of DMP in 132  
use of MC/ServiceGuard with VxVM 410  
vol_fmr_logsz tunable 480  
volume reconfiguration 413  
vxclustadm 411  
specifying to vxassist 244  
upgrading firmware 154  
controllers, mirroring guidelines 514  
converting disks 91  
copymaps 69  
vxdctl 420  
copy-on-write  
vxrecover 428  
used by instant snapshots 307  
Download from Www.Somanuals.com. All Manuals Search And Download.  
534 Index  
crash dumps  
using VxVM volumes for 107  
Cross-platform Data Sharing (CDS)  
alignment constraints 242  
disk format 81  
implementing 374  
default disk group 167  
defaultdg 167, 168  
defaults  
for vxdisk 81, 97  
CVM  
description file with vxmake 259  
detach policy  
cluster functionality of VxVM 397  
global 405  
local 406  
DETACHED  
plex kernel state 229  
volume kernel state 267  
device discovery  
D
d# 20, 78  
data change object  
DCO 69  
data redundancy 42, 43, 46  
data volume configuration 62  
database replay logs and sequential DRL 61  
databases  
introduced 22  
partial 82  
Device Discovery Layer 85  
Device Discovery Layer (DDL) 22, 85  
device files to access volumes 262, 518  
device names 20, 77  
configuring persistent 93  
device nodes  
controlling access for volume sets 366  
displaying access for volume sets 366  
enabling access for volume sets 365  
for volume sets 364  
Device path not valid 117  
devices  
resilvering 62  
resynchronizing 62  
DCO  
adding to RAID-5 volumes 277  
adding version 0 DCOs to volumes 356  
adding version 20 DCOs to volumes 275  
calculating plex size for version 20 70  
considerations for disk layout 200  
creating volumes with version 0 DCOs  
attached 250  
creating volumes with version 20 DCOs  
attached 252  
data change object 69  
adding foreign 89  
agile 77  
fabric 82  
metadevices 78  
determining version of 277  
dissociating version 0 DCOs from volumes 358  
effect on disk group split and join 200  
log plexes 71  
pathname 78  
dgalign_checking attribute 243  
dgfailpolicy attribute 409  
dirty bits in DRL 60  
dirty flags set on volumes 59  
dirty region logging. See DRL  
dirty regions 482  
log volume 69  
moving log plexes 277, 358  
reattaching version 0 DCOs to volumes 359  
removing version 0 DCOs from volumes 358  
specifying storage for version 0 plexes 357  
specifying storage for version 20 plexes 276  
used with DRL 60  
version 0 69  
version 20 69  
versioning 68  
disable failure policy 407  
DISABLED  
plex kernel state 229  
volume kernel state 267  
disabled paths 138  
disk access records  
stored in /etc/vx/darecs 212  
disk arrays  
dcolen attribute 69, 251, 357  
DCOSNP  
plex state 225  
A/A 126  
A/A-A 126  
A/P 125  
DDL 22  
Device Discovery Layer 85  
decision support  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Index 535  
A/P-C 126  
A/PF 126  
defaults file for shared 403  
defined 28  
A/PF-C 126  
deporting 173  
A/PG 126  
A/PG-C 126  
designating as shareable 401  
destroying 208  
Active/Active 126  
Active/Passive 125  
determining the default disk group 168  
disabling 207  
adding disks to DISKS category 87  
adding vendor-supplied support package 84  
Asymmetric Active/Active 126  
defined 21  
excluding support for 86  
listing excluded 86  
listing supported 85  
listing supported disks in DISKS category 87  
multipathed 22  
displaying boot disk group 168  
displaying default disk group 168  
displaying free space in 170  
displaying information about 169  
displaying version of 211  
effect of size on private region 166  
elimination of rootdg 165  
failure policy 407  
features supported by version 210  
forcing import of 187  
free space in 384  
impact of number of configuration copies on  
performance 473  
re-including support for 86  
removing disks from DISKS category 89  
removing vendor-supplied support package 84  
disk drives  
variable geometry 515  
disk duplexing 42, 255  
disk groups  
importing 174  
importing as shared 423  
importing forcibly 423  
activating shared 425  
activation in clusters 403  
adding disks to 171  
importing with cloned disks 176  
joining 197, 206  
joining in clusters 424  
avoiding conflicting minor numbers on  
import 187  
boot disk group 167  
layout of DCO plexes 200  
limitations of move, split, and join 199  
listing objects affected by a move 200  
listing shared 421  
bootdg 167  
checking for non-imported 452  
checking initialized disks 452  
checking number of configuration copies  
in 452  
checking on disk config size 451  
checking size of configuration database 451  
checking version number 451  
clearing locks on disks 186  
cluster-shareable 401  
compatible with CDS 171  
configuration backup and restoration 213  
configuring site consistency on 435  
configuring site-based allocation on 434  
converting to private 424  
creating 170  
making site consistent 439  
moving between systems 185  
moving disks between 184, 203  
moving licensed EMC disks between 203  
moving objects between 196, 203  
moving objects in clusters 424  
names reserved by system 167  
nodg 167  
number of spare disks 454  
private in clusters 401  
recovering destroyed 208  
recovery from failed reconfiguration 198  
removing disks from 172  
renaming 183  
reorganizing 195  
creating shared 422  
creating with old version number 212  
default disk group 167  
reserving minor numbers 187  
restarting moved volumes 204, 205, 207  
root 28  
defaultdg 167  
rootdg 28, 165  
Download from Www.Somanuals.com. All Manuals Search And Download.  
536 Index  
serial split brain condition 190  
disk access records file 212  
disk arrays 21  
displaying information 120, 121  
displaying information about 120, 169  
displaying spare 386  
dynamic LUN expansion 108  
EFI 81  
enabled path 138  
setting connectivity policies in clusters 425  
setting default disk group 168  
setting failure policies in clusters 426  
setting number of configuration copies 474  
shared in clusters 401  
specifying to commands 167  
splitting 196, 205  
splitting in clusters 424  
Storage Expert rules 451  
upgrading version of 208, 211  
version 208, 210  
enabling 117  
enclosures 23  
excluding free space from hot-relocation  
use 388  
disk media names 28, 77  
disk names 77  
failure handled by hot-relocation 380  
formatting 96  
configuring persistent 93  
disk sparing  
handling duplicated identifiers 175  
hot-relocation 379  
Storage Expert rules 454  
disk## 29, 78  
HP format 81  
initializing 90, 97  
disk##-## 29  
installing 96  
diskdetpolicy attribute 409  
diskgroup## 77  
disks 83  
invoking discovery of 84  
layout of DCO plexes 200  
listing tags on 177  
adding 101  
listing those supported in JBODs 87  
making available for hot-relocation 387  
making free space available for hot-relocation  
use 389  
marking as spare 387  
media name 77  
adding to disk groups 171  
adding to disk groups forcibly 423  
adding to DISKS category 87  
array support library 83  
auto-configured 81  
categories 83  
metadevices 78  
CDS format 81  
mirroring volumes on 272  
moving between disk groups 184, 203  
moving disk groups between systems 185  
moving volumes from 290  
names 77  
changing default layout attributes 97  
changing naming scheme 91  
checking for failed 454  
checking initialized disks not in disk group 452  
checking number of configuration copies in  
disk group 452  
naming schemes 78  
nopriv 80  
checking proportion spare in disk group 454  
clearing locks on 186  
cloned 176  
number 20  
obtaining performance statistics 470  
OTHER_DISKS category 83  
partial failure messages 383  
postponing replacement 112  
primary path 138  
putting under control of VxVM 90  
reinitializing 101  
releasing from disk groups 208  
removing 110, 112  
complete failure messages 384  
configuring newly added 82  
configuring persistent names 93  
converting 91  
default initialization values 97  
determining failed 383  
determining if shared 421  
Device Discovery Layer 85  
disabled path 138  
removing from disk groups 172  
removing from DISKS category 89  
removing from pool of hot-relocation  
discovery of by VxVM 83  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Index 537  
spares 388  
removing from VxVM control 112, 172  
removing tags from 178  
removing with subdisks 111, 112  
renaming 119  
replacing 112  
displaying information about paths 137  
displaying LUN group for a node 140  
displaying paths controlled by DMP node 140  
displaying paths for a controller 141  
displaying paths for an array port 141  
displaying recoveryoption values 159  
displaying status of DMP error handling  
thread 162  
replacing removed 115  
reserving for special purposes 119  
resolving status in clusters 404  
scanning for 82  
displaying status of DMP path restoration  
thread 161  
secondary path 138  
displaying TPD information 143  
dynamic multipathing 125  
enabling array ports 154  
setting connectivity policies in clusters 425  
setting failure policies in clusters 426  
setting tags on 177  
enabling controllers 154  
simple 80  
spare 384  
enabling multipathing 134  
enabling paths 154  
specifying to vxassist 244  
stripe unit size 515  
tagging with site name 434  
taking offline 118  
enclosure-based naming 127  
gathering I/O statistics 144  
in a clustered environment 132  
load balancing 129  
UDID flag 175  
logging levels 476  
unique identifier 175  
metanodes 126  
unreserving 120  
upgrading contoller firmware 154  
VM 28  
migrating to or from native multipathing 130  
nodes 126  
path aging 476  
writing a new identifier to 176  
DISKS category 83  
path failover mechanism 128  
path-switch tunable 476  
adding disks 87  
listing supported disks 87  
removing disks 89  
renaming an enclosure 155  
restore policy 160  
scheduling I/O on secondary paths 151  
setting the DMP restore polling interval 160  
stopping the DMP restore daemon 161  
vxdmpadm 139  
DMP  
check_all restore policy 160  
check_alternate restore policy 160  
check_disabled restore policy 160  
check_periodic restore policy 160  
coexistence with native multipathing 130  
configuring DMP path restoration policies 160  
configuring I/O throttling 157  
configuring response to I/O errors 156  
disabling array ports 153  
disabling controllers 153  
disabling multipathing 133  
disabling paths 153  
displaying DMP database information 137  
displaying DMP node for a path 139  
displaying DMP node for an enclosure 140  
displaying information about array ports 142  
displaying information about controllers 141  
displaying information about enclosures 142  
dmp_cache_open tunable 475  
dmp_daemon_count tunable 475  
dmp_delayq_interval tunable 475  
dmp_failed_io_threshold tunable 475  
dmp_fast_recovery tunable 475  
dmp_health_time tunable 476  
dmp_log_level tunable 476  
dmp_path_age tunable 476  
dmp_pathswitch_blks_shift tunable 476  
dmp_probe_idle_lun tunable 477  
dmp_queue_depth tunable 477  
dmp_restore_cycles tunable 477  
dmp_restore_interval tunable 477  
dmp_restore_policy tunable 478  
dmp_retry_count tunable 478  
dmp_retry_timeout tunable 478  
Download from Www.Somanuals.com. All Manuals Search And Download.  
538 Index  
dmp_scsi_timeout tunable 478  
dmp_stat_interval tunable 479  
DRL  
issues with simple disks 94  
mirroring across 255  
setting attributes of paths 146  
adding log subdisks 220  
adding logs to mirrored volumes 281  
checking existence of 450  
checking existence of mirror 450  
creating volumes with DRL enabled 252, 253  
determining if active 278  
determining if enabled 278  
dirty bits 60  
error messages  
Association count is incorrect 419  
Association not resolved 419  
Cannot auto-import group 419  
Configuration records are inconsistent 419  
Disk for disk group not found 187  
Disk group has no valid configuration  
copies 186, 419  
dirty region logging 60  
disabling 278  
Disk group version doesn’t support  
feature 209  
enabling on volumes 275  
hot-relocation limitations 381  
log subdisks 61  
maximum number of dirty regions 482  
minimum number of sectors 482  
recovery map in version 20 DCO 69  
re-enabling 278  
removing logs from mirrored volumes 282  
removing support for 279  
sequential 61  
Disk is in use by another host 186  
Disk is used by one or more subdisks 172  
Disk not moving, but subdisks on it are 200  
Duplicate record in configuration 419  
import failed 186  
No valid disk found containing disk group 186  
tmpsize too small to perform this relayout 55  
Volume has different organization in each  
mirror 286  
vxdg listmove failed 200  
use of DCO with 60  
errord daemon 128  
drl attribute 252, 281  
exclusive-write mode 402  
DRL guidelines 515  
exclusivewrite mode 402  
dumps  
explicit failover mode 126  
using VxVM volumes for 107  
duplexing 42, 255  
Extensible Firmware Interface (EFI) disks 81  
dynamic LUN expansion 108  
F
fabric devices 82  
FAILFAST flag 128  
E
EFI disks 81  
failover 397, 398  
EMC arrays  
failover mode 125  
moving disks between disk groups 203  
EMPTY  
failure handled by hot-relocation 380  
failure in RAID-5 handled by hot-relocation 380  
failure policies 407  
plex state 226  
volume state 265  
ENABLED  
setting for disk groups 426  
FastResync  
plex kernel state 229  
volume kernel state 267  
enabled paths, displaying 138  
enclosure-based naming 23, 79, 91  
displayed by vxprint 94  
DMP 127  
checking if enabled on volumes 293  
disabling on volumes 293  
effect of growing volume on 73  
enabling on new volumes 251  
enabling on volumes 292  
limitations 74  
enclosures 23  
Non-Persistent 67  
discovering disk access names in 94  
displaying information about 142  
issues with nopriv disks 94  
Persistent 68, 70  
size of bitmap 479  
snapshot enhancements 305  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Index 539  
use with snapshots 66  
fastresync attribute 251, 252, 293  
file systems  
limitations 381  
making free space on disks available for use  
by 389  
growing using vxresize 285  
shrinking using vxresize 285  
unmounting 290  
marking disks as spare 387  
modifying behavior of 395  
notifying users other than root 395  
operation of 379  
fire drill  
defined 432  
testing 440  
firmware  
upgrading 154  
FMR. See FastResync  
foreign devices  
partial failure messages 383  
preventing from running 395  
reducing performance impact of recovery 395  
removing disks from spare pool 388  
Storage Expert rules 454  
subdisk relocation 385  
adding 89  
subdisk relocation messages 390  
unrelocating subdisks 390  
formatting disks 96  
free space in disk groups 384  
fullinst snapshot type 343  
full-sized instant snapshots 307  
creating 327  
unrelocating subdisks using vxassist 392  
unrelocating subdisks using vxdiskadm 391  
unrelocating subdisks using vxunreloc 392  
use of free space in disk groups 384  
use of spare disks 384  
creating volumes for use as 323  
use of spare disks and free space 384  
using only spare disks for 390  
vxrelocd 380  
G
GAB 410  
HP disk format 81  
hpdisk format 81  
HP-UX multipathing 77  
global detach policy 405  
Group Membership and Atomic Broadcast  
(GAB) 410  
guidelines  
I
DRL 515  
mirroring 514  
I/O  
RAID-5 516  
gathering statistics for DMP 144  
kernel threads 19  
scheduling on secondary paths 151  
throttling 128  
use of statistics in performance tuning 469  
using traces for performance tuning 472  
I/O operations  
H
hasdcolog attribute 293  
HFS file systems  
resizing 285  
highwatermark attribute 322  
host failures 442  
hostnames  
maximum size of 480  
I/O policy  
displaying 147  
example 151  
specifying 147  
checking 455  
hot-relocation  
complete failure messages 384  
configuration summary 385  
daemon 380  
I/O throttling 157  
identifiers for tasks 267  
idle LUNs 477  
implicit failover mode 125  
init attribute 260  
initialization  
defined 75  
detecting disk failure 380  
detecting plex failure 380  
detecting RAID-5 subdisk failure 380  
excluding free space on disks from use by 388  
default attributes 97  
of disks 90, 97  
Download from Www.Somanuals.com. All Manuals Search And Download.  
540 Index  
initialization of disks 90  
instant snapshots  
backing up multiple volumes 333  
changing for disks 97  
layouts  
changing default used by vxassist 243  
left-symmetric 48  
cascaded 312  
creating backups 319  
creating for volume sets 334  
creating full-sized 327  
creating space-optimized 324  
creating volumes for use as full-sized 323  
displaying information about 342  
dissociating 340  
specifying default 243  
types of volume 236  
leave failure policy 407  
left-symmetric layout 48  
len subdisk attribute 222  
LIF area 102  
LIF LABEL record 102  
link objects 311  
full-sized 307  
improving performance of  
synchronization 345  
linked break-off snapshots 311  
creating 331  
reattaching 338  
refreshing 337  
linked third-mirror snapshots  
reattaching 339  
removing 341  
load balancing 126  
removing support for 279  
restoring volumes using 340  
space-optimized 309  
splitting hierarchies 341  
synchronizing 344  
across nodes in a cluster 398  
displaying policy for 147  
specifying policy for 147  
load-balancing  
specifying policy for 147  
local detach policy 406  
lock clearing on disks 186  
LOG plex state 226  
Intelligent Storage Provisioning (ISP) 32  
intent logging 304  
INVALID volume state 266  
ioctl calls 480, 481  
log subdisks 515  
IOFAIL plex condition 228  
IOFAIL plex state 226  
ISP volumes 32  
associating with plexes 220  
DRL 61  
logdisk 251, 256, 257  
logical units 125  
loglen attribute 253  
logs  
J
JBODs  
adding DRL log 281  
adding disks to DISKS category 87  
listing supported disks 87  
removing disks from DISKS category 89  
adding for RAID-5 283  
adding sequential DRL logs 281  
adding to volumes 274  
checking for disabled 452  
checking for multiple RAID-5 logs on same  
disk 449  
K
kernel states  
for plexes 229  
volumes 266  
RAID-5 50, 59  
removing DRL log 282  
removing for RAID-5 284  
removing sequential DRL logs 282  
resizing using vxvol 288  
specifying number for RAID-5 256  
usage with volumes 237  
logtype attribute 253  
LUN 125  
L
layered volumes  
converting to non-layered 300  
defined 51, 237  
striped-mirror 43  
layout attributes  
LUN expansion 108  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Index 541  
LUN group failover 126  
LUN groups  
DRL 60  
FastResync 60  
displaying details of 140  
LUNs  
FR 60  
logging 60  
idle 477  
performance 464  
removing DRL logs 282  
removing sequential DRL logs 282  
snapshots 66  
M
maps  
mirrored-concatenated volumes  
converting to concatenated-mirror 300  
creating 249  
defined 43  
mirrored-stripe volumes  
benefits of 42  
adding to volumes 274  
usage with volumes 237  
master node  
defined 400  
discovering 420  
maxautogrow attribute 322  
maxdev attribute 189  
MC/ServiceGuard  
use with VxVM in clusters 410  
memory  
checking configuration 453  
converting to striped-mirror 300  
creating 254  
defined 236  
performance 465  
granularity of allocation by VxVM 482  
maximum size of pool for VxVM 483  
minimum size of pool for VxVM 485  
persistence of FastResync in 67  
messages  
complete disk failure 384  
hot-relocation of subdisks 390  
partial disk failure 383  
metadata 177  
metadevices 78  
metanodes  
DMP 126  
migrating between DMP and native  
multipathing 130  
mirroring  
defined 42  
guidelines 514  
mirroring controllers 514  
mirroring plus striping 43  
mirrors  
adding to volumes 271  
boot disk 103  
creating of VxVM root disk 104  
creating snapshot 349  
defined 33  
removing from volumes 273  
specifying number of 249  
mirvol attribute 331  
mirvol snapshot type 343  
multipathing  
minimum queue load balancing policy 150  
minor numbers 187  
mirbrk snapshot type 343  
mirdg attribute 332  
disabling 133  
displaying information about 137  
enabling 134  
mirrored volumes  
adding DRL logs 281  
native to HP-UX 77  
Multi-Volume Support 361  
adding sequential DRL logs 281  
changing read policies for 289  
checking existence of mirrored DRL 450  
checking existence of without DRL 450  
configuring VxVM to create by default 272  
creating 249  
creating across controllers 247, 255  
creating across enclosures 255  
creating across targets 245  
defined 236  
N
names  
changing for disk groups 183  
defining for snapshot volumes 352  
device 20, 77  
disk 77  
disk media 28, 77  
plex 31  
dirty region logging 60  
Download from Www.Somanuals.com. All Manuals Search And Download.  
542 Index  
plex attribute 234  
renaming disks 119  
subdisk 29  
subdisk attribute 221  
VM disk 29  
volume 31  
changing region size 299  
changing speed of 299  
changing stripe unit size 298  
combining with conversion 300  
controlling progress of 299  
defined 54  
naming scheme  
destination layouts 294  
failure recovery 58  
how it works 54  
changing for disks 91  
changing for TPD enclosures 94  
for disk devices 78  
limitations 57  
native multipathing 77, 130  
ncachemirror attribute 325  
ndcomirror attribute 251, 252, 357  
ndcomirs attribute 275, 321  
NEEDSYNC volume state 266  
newvol attribute 330  
nmirror attribute 329, 330  
NODAREC plex condition 228  
nodes  
monitoring tasks for 299  
pausing 299  
performing 294  
resuming 299  
reversing direction of 300  
specifying non-default 298  
specifying plexes 298  
specifying task tags for 298  
temporary area 54  
DMP 126  
in clusters 399  
transformation characteristics 58  
transformations and volume length 58  
types of transformation 295  
viewing status of 299  
maximum number in a cluster 397  
node abort in clusters 417  
requesting status of 420  
shutdown in clusters 416  
use of vxclustadm to control cluster  
functionality 411  
online status 120  
ordered allocation 245, 251, 257  
OTHER_DISKS category 83  
overlapped seeks 514  
NODEVICE plex condition 228  
nodg 167  
P
nomanual path attribute 146  
non-autotrespass mode 126  
non-layered volume conversion 300  
Non-Persistent FastResync 67  
nopreferred path attribute 146  
nopriv disk type 80  
parity in RAID-5 46  
partial device discovery 82  
partition size  
displaying the value of 147  
specifying 148  
path aging 476  
path failover in DMP 128  
pathgroups  
nopriv disks  
issues with enclosures 94  
creating 134  
paths  
O
objects  
disabling for DMP 153  
enabling for DMP 154  
setting attributes of 146  
performance  
physical 20  
virtual 26  
off-host processing 369, 397  
OFFLINE plex state 226  
online backups  
implementing 371  
online invalid status 120  
online relayout  
analyzing data 469  
benefits of using VxVM 463  
changing values of tunables 474  
combining mirroring and striping 465  
effect of read policies 466  
examining ratio of reads to writes 471  
changing number of columns 298  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Index 543  
hot spots identified by I/O traces 472  
impact of number of disk group configuration  
copies 473  
spares 388  
removing with subdisks 111, 112  
replacing 112  
improving for instant snapshot  
synchronization 345  
load balancing in DMP 129  
mirrored volumes 464  
replacing removed 115  
reserving for special purposes 119  
spare 384  
taking offline 118  
unreserving 120  
physical objects 20  
ping-pong effect 132  
plex attribute 330  
plex conditions  
monitoring 467  
moving volumes to improve 470  
obtaining statistics for disks 470  
obtaining statistics for volumes 468  
RAID-5 volumes 465  
setting priorities 467  
IOFAIL 228  
striped volumes 464  
NODAREC 228  
striping to improve 471  
NODEVICE 228  
tracing volume operations 468  
tuning large systems 473  
tuning VxVM 472  
RECOVER 228  
REMOVED 228  
plex kernel states  
DETACHED 229  
using I/O statistics 469  
persistent device name database 93  
persistent device naming 93  
Persistent FastResync 68, 69, 70  
physical disks  
DISABLED 229  
ENABLED 229  
plex states  
ACTIVE 225  
adding to disk groups 171  
clearing locks on 186  
CLEAN 225  
DCOSNP 225  
complete failure messages 384  
determining failed 383  
EMPTY 226  
IOFAIL 226  
displaying information 120  
displaying information about 120, 169  
displaying spare 386  
LOG 226  
OFFLINE 226  
SNAPATT 226  
enabling 117  
SNAPDIS 226  
excluding free space from hot-relocation  
use 388  
SNAPDONE 226  
SNAPTMP 227  
failure handled by hot-relocation 380  
initializing 90  
STALE 227  
TEMP 227  
installing 96  
TEMPRM 227  
making available for hot-relocation 387  
making free space available for hot-relocation  
use 389  
TEMPRMSD 228  
plexes  
adding to snapshots 353  
associating log subdisks with 220  
associating subdisks with 218  
associating with volumes 229  
attaching to volumes 229  
changing attributes 234  
changing read policies for 289  
checking for detached 452  
checking for disabled 452  
comment attribute 234  
complete failure messages 384  
marking as spare 387  
moving between disk groups 184, 203  
moving disk groups between systems 185  
moving volumes from 290  
partial failure messages 383  
postponing replacement 112  
releasing from disk groups 208  
removing 110, 112  
removing from disk groups 172  
removing from pool of hot-relocation  
Download from Www.Somanuals.com. All Manuals Search And Download.  
544 Index  
condition flags 228  
converting to snapshot 351  
copying 233  
private region  
checking size of configuration database 451  
configuration database 80  
defined 80  
creating 223  
creating striped 224  
effect of large disk groups on 166  
defined 30  
public region 80  
detaching from volumes temporarily 231  
disconnecting from volumes 230  
displaying information about 224  
dissociating from volumes 233  
dissociating subdisks from 221  
failure in hot-relocation 380  
kernel states 229  
limit on number per volume 467  
maximum number of subdisks 481  
maximum number per volume 31  
mirrors 33  
moving 232, 277, 358  
name attribute 234  
names 31  
partial failure messages 383  
putil attribute 234  
putting online 231, 270  
reattaching 231  
recovering after correctable hardware  
failure 383  
removing 233  
removing from volumes 273  
sparse 57, 219, 229, 232  
specifying for online relayout 298  
states 224  
striped 38  
taking offline 230, 270  
tutil attribute 234  
putil  
plex attribute 234  
subdisk attribute 222  
R
RAID-0 38  
RAID-0+1 42  
RAID-1 42  
RAID-1+0 43  
RAID-5  
adding logs 283  
adding subdisks to plexes 219  
checking existence of log 450  
checking log is mirrored 450  
checking size of log 450  
guidelines 516  
hot-relocation limitations 381  
logs 50, 59  
parity 46  
removing logs 284  
specifying number of logs 256  
subdisk failure handled by hot-relocation 380  
volumes 46  
RAID-5 volumes  
adding DCOs to 277  
adding logs 283  
changing number of columns 298  
changing stripe unit size 298  
checking existence of RAID-5 log 450  
checking number of columns 453  
checking RAID-5 log is mirrored 450  
checking size of RAID-5 log 450  
creating 256  
defined 236  
performance 465  
removing logs 284  
types 31  
polling interval for DMP restore 160  
prefer read policy 289  
preferred plex  
performance of read policy 466  
read policy 289  
preferred priority path attribute 146  
primary path 125, 138  
primary path attribute 147  
priority load balancing 150  
private disk groups  
converting from shared 424  
in clusters 401  
private network  
raw device nodes  
controlling access for volume sets 366  
displaying access for volume sets 366  
enabling access for volume sets 365  
for volume sets 364  
read policies  
changing 289  
in clusters 399  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Index 545  
performance of 466  
prefer 289  
round 289  
select 289  
siteread 289, 433, 434, 436  
split 289  
limitations 381  
partial failure messages 383  
Remote Mirror feature  
administering 431  
remote mirrors  
administering 431  
REMOVED plex condition 228  
removing disks 112  
removing physical disks 110  
replacing disks 112  
replay logs and sequential DRL 61  
REPLAY volume state 266  
resilvering  
read-only mode 402  
readonly mode 402  
RECOVER plex condition 228  
recovery  
checkpoint interval 479  
I/O delay 479  
preventing on restarting volumes 271  
recovery accelerator 62  
recovery time  
Storage Expert rules 449  
redo log configuration 63  
redundancy  
checking for volumes 452  
of data on mirrors 236  
of data on RAID-5 236  
redundant-loop access 25  
region 80  
databases 62  
restoration of disk group configuration 213  
restore policy  
check_all 160  
check_alternate 160  
check_disabled 160  
check_periodic 160  
restored daemon 128  
restrictions  
VxVM-bootable volumes 103  
resyncfromoriginal snapback 317  
resyncfromreplica snapback 317  
resynchronization  
checkpoint interval 479  
I/O delay 479  
of volumes 59  
resynchronizing  
databases 62  
root disk  
creating mirrors 104  
root disk group 28, 165  
root disks  
regionsize attribute 275, 321, 322  
reinitialization of disks 101  
relayout  
changing number of columns 298  
changing region size 299  
changing speed of 299  
changing stripe unit size 298  
combining with conversion 300  
controlling progress of 299  
limitations 57  
monitoring tasks for 299  
online 54  
pausing 299  
performing online 294  
resuming 299  
reversing direction of 300  
specifying non-default 298  
specifying plexes 298  
specifying task tags for 298  
storage 54  
creating LVM from VxVM 106  
creating VxVM 104  
removing LVM 105  
root mirrors  
checking 454  
root volumes  
booting 104  
rootability 102  
transformation characteristics 58  
types of transformation 295  
viewing status of 299  
relocation  
checking 454  
rootdg 28  
round read policy 289  
round-robin  
automatic 379  
complete failure messages 384  
load balancing 150  
performance of read policy 466  
Download from Www.Somanuals.com. All Manuals Search And Download.  
546 Index  
read policy 289  
rules  
attributes 458  
secondary path attribute 147  
secondary path display 138  
select read policy 289  
sequential DRL  
checking attribute values 447  
checking disk group configuration copies 451  
checking disk group version number 451  
checking for full disk group configuration  
database 451  
checking for initialized disks 452  
checking for mirrored DRL 450  
checking for multiple RAID-5 logs on a  
disk 449  
defined 61  
maximum number of dirty regions 482  
sequential DRL attribute 252  
serial split brain condition 432  
correcting 194  
in campus clusters 190  
in disk groups 190  
shared disk groups  
checking for non-imported disk groups 452  
checking for non-mirrored RAID-5 log 450  
checking for RAID-5 log 450  
checking hardware 454  
activating 425  
activation modes 402  
converting to private 424  
creating 422  
checking hot-relocation 454  
checking mirrored volumes without a DRL 450  
checking mirrored-stripe volumes 453  
checking number of configuration copies in  
disk group 452  
checking number of RAID-5 columns 453  
checking number of spare disks 454  
checking number of stripe columns 454  
checking on disk config size 451  
checking plex and volume states 452  
checking RAID-5 log size 450  
checking rootability 454  
importing 423  
in clusters 401  
limitations of 409  
listing 421  
shared disks, configuring 519  
shared-read mode 402  
sharedread mode 402  
shared-write mode 402  
sharedwrite mode 402  
simple disk type 80  
simple disks  
issues with enclosures 94  
single active path policy 150  
Site Awareness license 434  
site consistency  
checking stripe unit size 453  
checking system name 455  
checking volume redundancy 452  
definitions 456, 457  
configuring 435  
finding information about 447  
for checking hardware 454  
defined 432  
site failure  
for checking rootability 454  
for checking system name 455  
for disk groups 451  
simulating 440  
site failures  
host failures 442  
for disk sparing 454  
for recovery time 449  
for striped mirror volumes 453  
listing attributes 447  
result types 448  
loss of connectivity 442  
recovery from 440, 443  
scenarios and recovery procedures 441  
storage failures 442  
site reattachment daemon 440  
site storage class 436  
site-based allocation  
configuring for disk groups 434  
configuring for volumes 436  
defined 432  
running 447  
setting values of attributes 448  
S
scandisks  
vxdisk subcommand 82  
secondary path 125  
site-based consistency  
configuring on existing disk groups 439  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Index 547  
siteconsistent attribute 435  
siteread read policy 289, 433, 434, 436  
sites  
creating volumes for use as full-sized  
instant 323  
defining names for 352  
reattaching 440  
displaying information about 355  
size units 216  
slave nodes  
displaying information about instant 342  
dissociating instant 340  
defined 400  
SmartSync 62  
emulation of third-mirror 310  
finding out those configured on a cache 347  
full-sized instant 65, 307  
hierarchy of 312  
improving performance of  
synchronization 345  
disabling on shared disk groups 482  
enabling on shared disk groups 482  
snap objects 72  
snap volume naming 317  
snapabort 305  
linked break-off 311  
SNAPATT plex state 226  
snapback  
listing for a cache 345  
merging with original volumes 352  
of volumes 63  
defined 306  
merging snapshot volumes 352  
resyncfromoriginal 317  
resyncfromreplica 317, 353  
snapclear  
on multiple volumes 317  
reattaching instant 338  
reattaching linked third-mirror 339  
refreshing instant 337  
creating independent volumes 354  
SNAPDIS plex state 226  
SNAPDONE plex state 226  
snapmir snapshot type 343  
snapshot hierarchies  
creating 337  
splitting 341  
snapshot mirrors  
removing 350  
removing instant 341  
removing linked snapshots from volumes 337  
removing mirrors from volumes 336  
restoring from instant 340  
resynchronization on snapback 317  
resynchronizing volumes from 353  
space-optimized instant 65, 309  
synchronizing instant 344  
third-mirror 64  
adding to volumes 336  
removing from volumes 336  
snapshots  
use of copy-on-write mechanism 307  
adding mirrors to volumes 336  
adding plexes to 353  
and FastResync 66  
backing up multiple volumes 333, 352  
backing up volumes online using 319  
cascaded 312  
comparison of features 65  
converting plexes to 351  
creating a hierarchy of 337  
creating backups using third-mirror 348  
creating for volume sets 334  
creating full-sized instant 327  
creating independent volumes 354  
creating instant 319  
creating linked break-off 331  
creating snapshots of 313  
creating space-optimized instant 324  
creating third-mirror break-off 329  
snapstart 305  
SNAPTMP plex state 227  
snapvol attribute 327, 332  
snapwait 330, 332  
source attribute 327, 332  
spaceopt snapshot type 343  
space-optimized instant snapshots 309  
creating 324  
spanned volumes 35  
spanning 35  
spare disks  
checking proportion in disk group 454  
displaying 386  
marking disks as 387  
used for hot-relocation 384  
sparse plexes 57, 219, 229, 232  
split read policy 289  
STALE plex state 227  
Download from Www.Somanuals.com. All Manuals Search And Download.  
548 Index  
standby path attribute 147  
states  
for plexes 224  
creating 253  
defined 236  
failure of 38  
of link objects 311  
volume 265  
statistics gathering 128  
storage  
ordered allocation of 245, 251, 257  
storage attributes and volume layout 244  
storage cache  
performance 464  
specifying non-default number of columns 254  
specifying non-default stripe unit size 254  
Storage Expert rules 453  
striped-mirror volumes  
benefits of 43  
converting to mirrored-stripe 300  
creating 254  
defined 237  
used by space-optimized instant  
snapshots 309  
Storage Expert  
mirroring columns 255  
mirroring subdisks 255  
performance 465  
check keyword 447  
checking default values of rule attributes 447  
command-line syntax 446  
diagnosing configuration issues 449  
info keyword 447  
trigger point for mirroring 255  
stripe-mirror-col-split-trigger-pt 255  
striping 38  
introduced 445  
list keyword 447  
listing rule attributes 447  
obtaining a description of a rule 447  
requirements 446  
striping guidelines 515  
striping plus mirroring 42  
subdisk names 29  
subdisks  
associating log subdisks 220  
associating with plexes 218  
associating with RAID-5 plexes 219  
associating with striped plexes 219  
blocks 29  
rule attributes 458  
rule definitions 456, 457  
rule result types 448  
rules 446  
rules engine 446  
run keyword 447  
running a rule 447  
setting values of rule attributes 448  
vxse 445  
changing attributes 221  
comment attribute 222  
complete failure messages 384  
copying contents of 217  
creating 215  
storage failures 442  
storage processor 125  
storage relayout 54  
stripe columns 38  
stripe unit size recommendations 515  
stripe units  
defined 29  
determining failed 383  
displaying information about 216  
dissociating from plexes 221  
dividing 217  
DRL log 61  
changing size 298  
checking size 453  
defined 38  
hot-relocation 75, 379, 385  
hot-relocation messages 390  
joining 218  
striped plexes  
len attribute 222  
adding subdisks 219  
defined 38  
striped volumes  
listing original disks after hot-relocation 394  
maximum number per plex 481  
mirroring in striped-mirror volumes 255  
moving after hot-relocation 390  
moving contents of 217  
name attribute 221  
changing number of columns 298  
changing stripe unit size 298  
checking number of columns 454  
checking stripe unit size 453  
partial failure messages 383  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Index 549  
physical disk placement 513  
putil attribute 222  
changing state of 268, 269  
identifiers 267  
RAID-5 failure of 380  
listing 268  
RAID-5 plex, configuring 516  
removing from VxVM 221  
restrictions on moving 217  
specifying different offsets for  
unrelocation 393  
managing 268  
modifying parameters of 269  
monitoring 268  
monitoring online relayout 299  
pausing 269  
splitting 217  
resuming 269  
tutil attribute 222  
specifying tags 267  
unrelocating after hot-relocation 390  
unrelocating to different disks 393  
unrelocating using vxassist 392  
unrelocating using vxdiskadm 391  
unrelocating using vxunreloc 392  
specifying tags on online relayout  
operation 298  
tags 267  
TEMP plex state 227  
temporary area used by online relayout 54  
TEMPRM plex state 227  
TEMPRMSD plex state 228  
third-mirror  
swap space  
using VxVM volumes to increase 107  
SYNC volume state 266  
synchronization  
snapshots 64  
controlling for instant snapshots 344  
improving performance of 345  
syncing attribute 319, 344  
syncpause 344  
third-mirror break-off snapshots  
creating 329  
third-party driver (TPD) 84  
throttling 128  
syncresume 344  
TPD  
syncstart 344  
syncstop 344  
syncwait 344  
system names  
displaying path information 143  
support for coexistence 84  
tpdmode attribute 94  
trigger point in striped-mirror volumes 255  
tunables  
checking 454, 455  
changing values of 474  
dmp_cache_open 475  
dmp_daemon_count 475  
dmp_delayq_interval 475  
dmp_failed_io_threshold 475  
dmp_fast_recovery 475  
dmp_health_time 476  
dmp_log_level 476  
T
t# 20, 78  
tags  
for tasks 267  
listing for disks 177  
removing from disks 178  
removing from volumes 288  
renaming 288  
dmp_path_age 476  
dmp_pathswitch_blks_shift 476  
dmp_probe_idle_lun 477  
dmp_queue_depth 477  
dmp_restore_cycles 477  
dmp_restore_interval 477  
dmp_restore_policy 478  
dmp_retry_count 478  
dmp_retry_timeout 478  
dmp_scsi_timeout 478  
dmp_stat_interval 479  
vol_checkpt_default 479  
setting on disks 177  
setting on volumes 257, 288  
specifying for online relayout tasks 298  
specifying for tasks 267  
target IDs  
number 20  
specifying to vxassist 244  
target mirroring 245, 255  
task monitor in VxVM 267  
tasks  
aborting 268  
Download from Www.Somanuals.com. All Manuals Search And Download.  
550 Index  
vol_default_iodelay 479  
vol_fmr_logsz 68, 479  
vol_max_vol 480  
checking for disk group 451  
disk group 208  
displaying for disk group 211  
upgrading 208  
vol_maxio 480  
vol_maxioctl 480  
virtual objects 26  
VM disks  
defined 28  
vol_maxparallelio 481  
vol_maxspecialio 481  
vol_subdisk_num 481  
volcvm_smartsync 482  
voldrl_max_drtregs 482  
voldrl_max_seq_dirty 61, 482  
voldrl_min_regionsz 482  
voliomem_chunk_size 482  
voliomem_maxpool_sz 483  
voliot_errbuf_dflt 483  
voliot_iobuf_default 483  
voliot_iobuf_limit 483  
voliot_iobuf_max 484  
voliot_max_open 484  
volpagemod_max_memsz 484  
volraid_minpool_size 485  
volraid_rsrtransmax 485  
determining if shared 421  
displaying spare 386  
excluding free space from hot-relocation  
use 388  
initializing 90  
making free space available for hot-relocation  
use 389  
marking as spare 387  
mirroring volumes on 272  
moving volumes from 290  
names 29  
postponing replacement 112  
removing from pool of hot-relocation  
spares 388  
renaming 119  
tutil  
vol## 31  
plex attribute 234  
vol##-## 31  
subdisk attribute 222  
vol_checkpt_default tunable 479  
vol_default_iodelay tunable 479  
vol_fmr_logsz tunable 68, 479  
vol_max_vol tunable 480  
vol_maxio tunable 480  
vol_maxioctl tunable 480  
vol_maxparallelio tunable 481  
vol_maxspecialio tunable 481  
vol_subdisk_num tunable 481  
volbrk snapshot type 343  
U
UDID flag 175  
udid_mismatch flag 176  
units of size 216  
use_all_paths attribute 151  
V
volcvm_smartsync tunable 482  
voldrl_max_drtregs tunable 482  
voldrl_max_seq_dirty tunable 61, 482  
voldrl_min_regionsz tunable 482  
voliomem_chunk_size tunable 482  
voliomem_maxpool_sz tunable 483  
voliot_errbuf_dflt tunable 483  
voliot_iobuf_default tunable 483  
voliot_iobuf_limit tunable 483  
voliot_iobuf_max tunable 484  
voliot_max_open tunable 484  
volpagemod_max_memsz tunable 484  
volraid_minpool_size tunable 485  
volraid_rsrtransmax tunable 485  
volume kernel states  
V-5-1-2536 286  
V-5-1-2829 209  
V-5-1-552 172  
V-5-1-569 419  
V-5-1-587 186  
V-5-2-3091 200  
V-5-2-369 172  
V-5-2-4292 200  
version 0  
of DCOs 69  
version 20  
of DCOs 69  
versioning  
of DCOs 68  
versions  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Index 551  
DETACHED 267  
DISABLED 267  
ENABLED 267  
changing stripe unit size 298  
character device files 262, 518  
checking for disabled 452  
volume length, RAID-5 guidelines 516  
volume resynchronization 59  
volume sets  
checking for stopped 452  
checking if FastResync is enabled 293  
checking redundancy of 452  
combining mirroring and striping for  
performance 465  
adding volumes to 362  
administering 361  
controlling access to raw device nodes 366  
creating 362  
combining online relayout and conversion 300  
concatenated 35, 236  
creating instant snapshots of 334  
displaying access to raw device nodes 366  
enabling access to raw device nodes 365  
listing details of 363  
raw device nodes 364  
removing volumes from 364  
starting 363  
concatenated-mirror 45, 237  
configuring exclusive open by cluster node 426  
configuring site consistency on 435  
configuring site-based allocation on 436  
converting between layered and non-  
layered 300  
converting concatenated-mirror to mirrored-  
concatenated 300  
stopping 363  
volume states  
converting mirrored-concatenated to  
concatenated-mirror 300  
ACTIVE 265  
CLEAN 265  
EMPTY 265  
converting mirrored-stripe to striped-  
mirror 300  
INVALID 266  
NEEDSYNC 266  
converting striped-mirror to mirrored-  
stripe 300  
REPLAY 266  
creating 238  
SYNC 266  
volumes  
creating concatenated-mirror 249  
creating for use as full-sized instant  
snapshots 323  
creating from snapshots 354  
creating mirrored 249  
creating mirrored-concatenated 249  
creating mirrored-stripe 254  
creating RAID-5 256  
accessing device files 262, 518  
adding DRL logs 281  
adding logs and maps to 274  
adding mirrors 271  
adding RAID-5 logs 283  
adding sequential DRL logs 281  
adding snapshot mirrors to 336  
adding subdisks to plexes of 219  
adding to volume sets 362  
adding version 0 DCOs to 356  
adding version 20 DCOs to 275  
advanced approach to creating 238  
assisted approach to creating 239  
associating plexes with 229  
attaching plexes to 229  
backing up 303  
creating snapshots 350  
creating striped 253  
creating striped-mirror 254  
creating using vxmake 258  
creating using vxmake description file 259  
creating with version 0 DCOs attached 250  
creating with version 20 DCOs attached 252  
defined 31  
detaching plexes from temporarily 231  
disabling FastResync 293  
backing up online using snapshots 319  
block device files 262, 518  
booting VxVM-rootable 104  
changing layout online 294  
changing number of columns 298  
changing read policies for mirrored 289  
disconnecting plexes 230  
displaying information 264  
displaying information about snapshots 355  
dissociating plexes from 233  
dissociating version 0 DCOs from 358  
DRL 515  
Download from Www.Somanuals.com. All Manuals Search And Download.  
552 Index  
effect of growing on FastResync maps 73  
recovering after correctable hardware  
failure 383  
removing 290  
enabling FastResync on 292  
enabling FastResync on new 251  
excluding storage from use by vxassist 244  
finding maximum size of 242  
finding out maximum possible growth of 284  
flagged as dirty 59  
initializing contents to zero 261  
initializing using vxassist 260  
initializing using vxvol 261  
kernel states 266  
layered 43, 51, 237  
limit on number of plexes 31  
limitations 31  
making immediately available for use 260  
maximum number of 480  
maximum number of data plexes 467  
merging snapshots 352  
removing DRL logs 282  
removing from /etc/fstab 290  
removing linked snapshots from 337  
removing mirrors from 273  
removing plexes from 273  
removing RAID-5 logs 284  
removing sequential DRL logs 282  
removing snapshot mirrors from 336  
removing support for DRL and instant  
snapshots 279  
removing version 0 DCOs from 358  
resizing 284  
resizing using vxassist 286  
resizing using vxresize 285  
resizing using vxvol 287  
restarting moved 204, 205, 207  
restoring from instant snapshots 340  
restrictions on VxVM-bootable 103  
resynchronizing from snapshots 353  
snapshots 63  
mirrored 42, 236  
mirrored-concatenated 43  
mirrored-stripe 42, 236  
mirroring across controllers 247, 255  
mirroring across targets 245, 255  
mirroring all 272  
spanned 35  
mirroring on disks 272  
specifying default layout 243  
specifying non-default number of columns 254  
specifying non-default relayout 298  
specifying non-default stripe unit size 254  
specifying storage for version 0 DCO  
plexes 357  
mirroring VxVM-rootable 103  
moving from VM disks 290  
moving to improve performance 470  
names 31  
naming snap 317  
obtaining performance statistics 468  
performance of mirrored 464  
performance of RAID-5 465  
performance of striped 464  
performing online relayout 294  
placing in maintenance mode 270  
preparing for DRL and instant snapshot  
operations 275  
specifying storage for version 20 DCO  
plexes 276  
specifying use of storage to vxassist 244  
starting 271  
starting using vxassist 260  
starting using vxvol 261  
states 265  
stopping 270  
preventing recovery on restarting 271  
RAID-0 38  
stopping activity on 290  
striped 38, 236  
RAID-0+1 42  
striped-mirror 43, 237  
RAID-1 42  
RAID-1+0 43  
RAID-10 43  
striping to improve performance 471  
taking multiple snapshots 317  
tracing operations 468  
RAID-5 46, 236  
raw device files 262, 518  
reattaching plexes 231  
trigger point for mirroring in striped-  
mirror 255  
types of layout 236  
reattaching version 0 DCOs to 359  
reconfiguration in clusters 413  
upgrading to use new features 279  
using logs and maps with 237  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Index 553  
zeroing out contents of 261  
vxassist  
relaying out volumes online 294  
removing DCOs from volumes 280  
removing DRL logs 282  
adding a log subdisk 220  
adding a RAID-5 log 283  
removing mirrors 273  
adding DCOs to volumes 357  
removing plexes 273  
adding DRL logs 281  
removing RAID-5 logs 284  
removing tags from volumes 288  
removing version 0 DCOs from volumes 358  
removing volumes 290  
adding mirrors to volumes 230, 271  
adding sequential DRL logs 282  
advantages of using 239  
changing number of columns 298  
changing stripe unit size 298  
command usage 240  
replacing tags set on volumes 288  
reserving disks 119  
resizing volumes 286  
configuring exclusive access to a volume 426  
configuring site consistency on volumes 435  
converting between layered and non-layered  
volumes 300  
resynchronizing volumes from snapshots 353  
setting default values 241  
setting tags on volumes 257, 288, 289  
snapabort 305  
creating cache volumes 322  
snapback 306  
creating concatenated-mirror volumes 249  
creating mirrored volumes 249  
creating mirrored-concatenated volumes 249  
creating mirrored-stripe volumes 254  
creating RAID-5 volumes 256  
creating snapshots 348  
creating striped volumes 253  
creating striped-mirror volumes 254  
creating volumes 239  
snapshot 306  
snapstart 305  
specifing number of mirrors 249  
specifying number of RAID-5 logs 256  
specifying ordered allocation of storage 245  
specifying plexes for online relayout 298  
specifying storage attributes 244  
specifying storage for version 0 DCO  
plexes 357  
creating volumes for use as full-sized instant  
snapshots 324  
creating volumes with DRL enabled 253  
creating volumes with version 0 DCOs  
attached 251  
creating volumes with version 20 DCOs  
attached 252  
defaults file 241  
specifying tags for online relayout tasks 298  
taking snapshots of multiple volumes 352  
unrelocating subdisks after hot-relocation 392  
vxcache  
listing snapshots in a cache 345  
resizing caches 347  
starting cache objects 323  
stopping a cache 347  
defining layout on specified storage 244  
discovering maximum volume size 242  
displaying information about snapshots 355  
dissociating snapshots from volumes 354  
excluding storage from use 244  
finding out how much volumes can grow 284  
listing tags set on volumes 257, 288  
merging snapshots with volumes 353  
mirroring across controllers 247, 255  
mirroring across enclosures 255  
mirroring across targets 245, 247  
moving DCO log plexes 277  
tuning cache autogrow 346  
vxcached  
tuning 346  
vxclustadm 411  
vxconfigd  
managing with vxdctl 212  
monitoring configuration changes 213  
operation in clusters 414  
vxcp_lvm_root  
used to create VxVM root disk 104  
used to create VxVM root disk mirrors 104  
vxdarestore  
used to handle simple/nopriv disk failures 95  
vxdco  
dissociating version 0 DCOs from volumes 358  
moving DCO plexes 358  
moving subdisks after hot-relocation 392  
moving volumes 470  
Download from Www.Somanuals.com. All Manuals Search And Download.  
554 Index  
reattaching version 0 DCOs to volumes 359  
removing version 0 DCOs from volumes 358  
vxdctl  
forcing import of disk groups 187  
importing a disk group containing cloned  
disks 177  
checking cluster protocol version 427  
managing vxconfigd 212  
importing cloned disks 178  
importing disk groups 175  
setting a site tag 434  
importing shared disk groups 423  
joining disk groups 206  
listing disks with configuration database  
copies 178  
setting default disk group 168  
upgrading cluster protocol version 428  
usage in clusters 420  
vxdctl enable  
listing objects affected by move 200  
listing shared disk groups 421  
listing spare disks 386  
moving disk groups between systems 185  
moving disks between disk groups 184  
moving objects between disk groups 203  
obtaining copy size of configuration  
database 166  
configuring new disks 82  
invoking device discovery 84  
used to configure new disks 82  
vxddladm  
adding disks to DISKS category 87  
adding foreign devices 89  
changing naming scheme 92  
listing excluded disk arrays 86  
listing supported disk arrays 85  
listing supported disks in DISKS category 87  
removing disks from DISKS category 89  
used to exclude support for disk arrays 86  
used to re-include support for disk arrays 86  
vxdestroy_lvmroot  
placing a configuration database on cloned  
disks 177  
reattaching a site 440  
recovering destroyed disk groups 208  
removing disks from disk groups 172  
renaming disk groups 183  
setting base minor number 188  
setting disk connectivity policy in a cluster 425  
setting disk group policies 409  
setting failure policy in a cluster 426  
setting maximum number of devices 189  
simulating site failure 440  
used to remove LVM root disks 105  
vxdg  
changing activation mode on shared disk  
groups 425  
clearing locks on disks 186  
configuring site consistency for a disk  
group 435  
splitting disk groups 205  
upgrading disk group version 211  
configuring site-based allocation for a disk  
group 434  
controllingl CDS compatibility of new disk  
groups 171  
converting shared disk groups to private 424  
correcting serial split brain condition 195  
creating disk groups 171  
creating disk groups with old version  
number 212  
vxdisk  
clearing locks on disks 186  
defaults file 81, 97  
determining if disks are shared 421  
discovering disk access names 94  
displaying information about disks 169  
displaying multipathing information 137  
listing disks 120  
listing spare disks 386  
creating shared disk groups 422  
deporting disk groups 174  
destroying disk groups 208  
listing tags on disks 177  
notifying dynamic LUN expansion 108  
placing a configuration database on a cloned  
disk 177  
disabling a disk group 207  
displaying boot disk group 168  
displaying default disk group 168  
displaying disk group version 211  
displaying free space in disk groups 170  
displaying information about disk groups 169  
removing tags from disks 178  
scanning disk devices 82  
setting a site name 434  
setting tags on disks 177  
updating the disk identifier 176  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Index 555  
vxdisk scandisks  
rescanning devices 82  
scanning devices 82  
vxdiskadd  
vxdiskunsetup  
removing disks from VxVM control 112, 172  
vxdmpadm  
changing TPD naming scheme 94  
adding disks to disk groups 171  
creating disk groups 171  
placing disks under VxVM control 101  
configuring an APM 162  
configuring I/O throttling 157  
configuring response to I/O errors 156  
disabling controllers in DMP 136  
disabling I/O in DMP 153  
discovering disk access names 94  
displaying APM information 162  
displaying DMP database information 137  
displaying DMP node for a path 140  
displaying DMP node for an enclosure 140  
displaying I/O error recovery settings 159  
displaying I/O policy 147  
vxdiskadm  
Add or initialize one or more disks 97, 171  
adding disks 97  
adding disks to disk groups 171  
Change/display the default disk layout 97  
changing disk-naming scheme 91  
changing the disk-naming scheme 91  
creating disk groups 171  
deporting disk groups 173  
Disable (offline) a disk device 118  
Enable (online) a disk device 117  
Enable access to (import) a disk group 174  
Exclude a disk from hot-relocation use 389  
excluding free space on disks from hot-  
relocation use 389  
displaying I/O throttling settings 159  
displaying information about controllers 141  
displaying information about enclosures 142  
displaying partition size 147  
displaying paths controlled by DMP node 140  
displaying status of DMP error handling  
thread 162  
importing disk groups 174  
initializing disks 97  
List disk information 121  
displaying status of DMP restoration  
thread 161  
listing spare disks 386  
Make a disk available for hot-relocation  
use 389  
displaying TPD information 143  
enabling I/O in DMP 154  
gathering I/O statistics 144  
making free space on disks available for hot-  
relocation use 389  
listing information about array ports 142  
removing an APM 163  
Mark a disk as a spare for a disk group 387  
marking disks as spare 387  
renaming enclosures 155  
setting I/O policy 150, 151  
Mirror volumes on a disk 272  
mirroring volumes 272  
setting path attributes 146  
setting restore polling interval 160  
specifying DMP path restoration policy 160  
stopping DMP restore daemon 161  
Move volumes from a disk 291  
moving disk groups between systems 187  
moving disks between disk groups 185  
moving subdisks after hot-relocation 391  
moving subdisks from disks 172  
moving volumes from VM disks 291  
Remove a disk 110, 172  
vxedit  
changing plex attributes 234  
changing subdisk attributes 221, 222  
configuring number of configuration copies for  
a disk group 474  
Remove a disk for replacement 113  
Remove access to (deport) a disk group 173  
removing disks from pool of hot-relocation  
spares 388  
excluding free space on disks from hot-  
relocation use 388  
making free space on disks available for hot-  
relocation use 389  
Replace a failed or removed disk 115  
Turn off the spare flag on a disk 388  
Unrelocate subdisks back to a disk 391  
unrelocating subdisks after hot-relocation 391  
marking disks as spare 387  
removing a cache 347  
removing disks from pool of hot-relocation  
spares 388  
Download from Www.Somanuals.com. All Manuals Search And Download.  
556 Index  
removing instant snapshots 341  
enclosure-based disk names 94  
identifying RAID-5 log plexes 284  
listing spare disks 386  
used with enclosure-based disk names 94  
verifying if volumes are prepared for instant  
snapshots 321  
removing plexes 234  
removing snapshots from a cache 347  
removing subdisks from VxVM 221  
removing volumes 290  
renaming disks 119  
reserving disks 119  
viewing base minor number 188  
VxFS file system resizing 285  
vxiod I/O kernel threads 19  
vxmake  
vxrecover  
preventing recovery 271  
recovering plexes 383  
associating plexes with volumes 229  
associating subdisks with new plexes 218  
creating cache objects 322  
creating plexes 223, 271  
creating striped plexes 224  
creating subdisks 215  
restarting moved volumes 204, 205, 207  
restarting volumes 271  
vxrelayout  
resuming online relayout 299  
reversing direction of online relayout 300  
viewing status of online relayout 299  
creating volumes 258  
vxrelocd  
using description file with 259  
vxmend  
hot-relocation daemon 380  
modifying behavior of 395  
putting plexes online 270  
re-enabling plexes 231  
notifying users other than root 395  
operation of 381  
taking plexes offline 230, 270  
vxmirror  
preventing from running 395  
reducing performance impact of recovery 395  
configuring VxVM default behavior 272  
mirroring volumes 272  
vxnotify  
vxres_lvmroot  
used to create LVM root disks 106  
vxresize  
monitoring configuration changes 213  
vxplex  
growing volumes and file systems 285  
limitations 285  
adding RAID-5 logs 283  
shrinking volumes and file systems 285  
attaching plexes to volumes 229, 271  
converting plexes to snapshots 351  
copying plexes 233  
vxrootmir  
used to create VxVM root disk mirrors 105  
vxsd  
detaching plexes temporarily 231  
dissociating and removing plexes 233  
dissociating plexes from volumes 234  
moving plexes 232  
adding log subdisks 220  
adding subdisks to RAID-5 plexes 219  
adding subdisks to striped plexes 219  
associating subdisks with existing plexes 219  
dissociating subdisks 221  
reattaching plexes 231  
removing mirrors 273  
filling in sparse plexes 219  
removing plexes 273  
joining subdisks 218  
removing RAID-5 logs 284  
vxprint  
moving subdisk contents 217  
removing subdisks from VxVM 221  
splitting subdisks 217  
checking if FastResync is enabled 293  
determining if DRL is enabled 278  
displaying DCO information 276, 357  
displaying plex information 224  
displaying snapshots configured on a  
cache 347  
vxse  
Storage Expert 445  
vxse_dc_failures  
rule to check for hardware failures 454  
vxse_dg1  
rule to check for full disk group configuration  
database 451  
displaying subdisk information 216  
displaying volume information 264  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Index 557  
vxse_dg2  
rule to check disk group configuration  
copies 451  
vxse_dg3  
rule to check on disk config size 451  
vxse_dg4  
rule to check disk group version number 451  
vxse_dg5  
administering instant snapshots 307  
backing up multiple volumes 333  
controlling instant snapshot  
synchronization 344  
creating a cascaded snapshot hierarchy 337  
creating full-sized instant snapshots 327, 332  
creating linked break-off snapshot  
volumes 331  
rule to check number of configuration copies in  
disk group 452  
creating space-optimized instant  
snapshots 325  
vxse_dg6  
rule to check for non-imported disk groups 452  
vxse_disk  
rule to check for initialized disks 452  
vxse_disklog  
displaying information about instant  
snapshots 342  
dissociating instant snapshots 340  
preparing volumes for DRL and instant  
snapshots operations 275  
rule to check for multiple RAID-5 logs on a  
disk 449  
vxse_drl1  
preparing volumes for instant snapshots 321  
reattaching instant snapshots 338  
reattaching linked third-mirror snapshots 339  
refreshing instant snapshots 337  
removing a snapshot mirror from a  
volume 336, 337  
rule to check for mirrored volumes without a  
DRL 450  
vxse_drl2  
rule to check for mirrored DRL 450  
vxse_host  
rule to check system name 455  
vxse_mirstripe  
rule to check mirrored-stripe volumes 453  
vxse_raid5  
rule to check number of RAID-5 columns 453  
vxse_raid5log1  
rule to check for RAID-5 log 450  
vxse_raid5log2  
rule to check RAID-5 log size 450  
vxse_raid5log3  
rule to check for non-mirrored RAID-5 log 450  
vxse_redundancy  
rule to check volume redundancy 452  
vxse_rootmir  
rule to check rootability 454  
vxse_spares  
rule to check number of spare disks 454  
vxse_stripes1  
rule to check stripe unit size 453  
vxse_stripes2  
rule to check number of stripe columns 454  
vxse_volplex  
removing support for DRL and instant  
snapshots 279  
restore 308  
restoring volumes 340  
splitting snapshot hierarchies 341  
vxsplitlines  
diagnosing serial split brain condition 194  
vxstat  
determining which disks have failed 383  
obtaining disk performance statistics 470  
obtaining volume performance statistics 468  
usage with clusters 428  
zeroing counters 469  
vxtask  
aborting tasks 269  
listing tasks 269  
monitoring online relayout 299  
monitoring tasks 269  
pausing online relayout 299  
resuming online relayout 299  
resuming tasks 269  
vxtrace  
tracing volume operations 468  
vxtune  
setting volpagemod_max_memsz 485  
vxunreloc  
rule to check plex and volume states 452  
vxsited daemon 440  
vxsnap  
listing original disks of hot-relocated  
subdisks 394  
adding snapshot mirrors to volumes 336  
Download from Www.Somanuals.com. All Manuals Search And Download.  
558 Index  
moving subdisks after hot-relocation 392  
restarting after errors 394  
specifying different offsets for unrelocated  
subdisks 393  
setting read policy 289  
starting volumes 261, 271  
stopping volumes 270, 290  
zeroing out volumes 261  
unrelocating subdisks after hot-relocation 392  
unrelocating subdisks to different disks 393  
VxVM  
vxvset  
adding volumes to volume sets 362  
controlling access to raw device nodes 366  
creating volume sets 362  
creating volume sets with raw device  
access 365  
listing details of volume sets 363  
removing volumes from volume sets 364  
starting volume sets 363  
benefits to performance 463  
cluster functionality (CVM) 397  
configuration daemon 212  
configuring disk devices 82  
configuring to create mirrored volumes 272  
dependency on operating system 19  
disk discovery 83  
stopping volume sets 363  
granularity of memory allocation by 482  
limitations of shared disk groups 409  
maximum number of data plexes per  
volume 467  
maximum number of subdisks per plex 481  
maximum number of volumes 480  
maximum size of memory pool 483  
minimum size of memory pool 485  
objects in 26  
W
warning messages  
Specified region-size is larger than the limit on  
the system 320  
Z
zero  
operation in clusters 398  
performance tuning 472  
setting volume contents to 261  
removing disks from 172  
removing disks from control of 112  
shared objects in cluster 402  
size units 216  
task monitor 267  
types of volume layout 236  
upgrading 208  
upgrading disk group version 211  
use with MC/ServiceGuard in clusters 410  
VXVM_DEFAULTDG environment variable 167  
VxVM-rootable volumes  
mirroring 103  
vxvol  
configuring exclusive access to a volume 426  
configuring site consistency on volumes 436  
disabling DRL 278  
disabling FastResync 293  
enabling FastResync 292  
initializing volumes 261  
putting volumes in maintenance mode 270  
re-enabling DRL 278  
resizing logs 288  
resizing volumes 287  
restarting moved volumes 204, 205, 207  
Download from Www.Somanuals.com. All Manuals Search And Download.  

Graco Baby Accessories PD226470A User Manual
Graco Water Pump 332317C User Manual
Hans Grohe Plumbing Product 06672XX0 User Manual
Hasbro Robotics 81273 User Manual
Havis Shields Work Light KR DL User Manual
Hotpoint Refrigerator HL161AI User Manual
HP Hewlett Packard Laptop x1100 User Manual
HP Hewlett Packard Personal Computer BL465C G7 User Manual
HP Hewlett Packard Welding System XLWW User Manual
IBM Computer Monitor 31P6240 User Manual