This is a quick-and-dirty guide to recovering VxVM volumes and filesystems on a Solaris server. These instructions are not intended to be an in-depth troubleshooting guide for the Veritas Volume Manager. This is just something to try if your VxVM volumes did not come up. A possible scenario would be: you rebooted your Sun box and some or all VxVM-based filesystems are missing.

So, without further ado, here are the steps:

1. First things first: check the physical status of your disks and/or storage array. If you are using PowerPath with Symmetrix or CLARiiON SAN, use the powermt command:

   # powermt display
   Symmetrix logical device count=6
   CLARiiON logical device count=0
   ==============================================================================
   ----- Host Bus Adapters ---------  ------ I/O Paths -----  ------ Stats ------
   ### HW Path                        Summary   Total   Dead  IO/Sec Q-IOs Errors
   ==============================================================================
   3072 pci@1e/SUNW,qlc@2/fp@0         optimal       5      0       -     0      0
   3073 pci@1e/SUNW,qlc@3/fp@0         optimal       5      0       -     0      0

   If everything checks out, proceed to step 2 and if not – call your storage guru.

2. Use vxprint to see the status of your VxVM volumes:

   #vxprint
   v  V8           -            DISABLED ACTIVE   20971520 SELECT    -        fsgen
   pl V8-01        V8           DISABLED ACTIVE   20971520 CONCAT    -        RW
   sd doorsDG04-01 V8-01        doorsDG04 0       20971520 0         emcpower1 ENA

   In this example the diskgroup name is “doorsDG”, the volume name is “V8″, plex name is “V8-01″, and subdisk name is “doorsDG04-01″. For a more detailed status of a particular diskgroup use the following syntax:

   #vxprint -ht -g doorsDG

   A disabled volume may have plexes in several different states: DISABLED-RECOVER, DISABLED-STALE, and DISABLED-CLEAN.

3. If your plex is DISABLED-RECOVER, change it to STALE using vxmend:

   #vxmend -g doorsDG fix stale V8-01

4. If your plex is in DISABLED-STALE, change it to DISABLED-CLEAN:

   #vxmend -g doorsDG fix clean V8-01

5. Once all plexes in a volume are listed as DISABLED-CLEAN, you can try to start the volume:

   #vxvol -g doorsDG start V8

   To verify the status of your volume, run this command:

   #vxprint -ht -g doorsDG
   v  V8           fsgen        ENABLED  20971520 -        ACTIVE   -       -
   pl V8-01        V8           ENABLED  20971520 -        ACTIVE   -       -
   sd doorsDG04-01 V8-01        ENABLED  20971520 0        -        -       -

6. The next step would be to mount your filesystems. However, chances are that there is some disk corruption that needs to be fixed first. So, run fsck on the VxVM disks:

   #fsck -F vxfs -y /dev/vx/dsk/doorsDG/V8
   log replay in progress
   replay complete - marking super-block as CLEAN

   If everything comes up as CLEAN, then run mount -a or mountall and you are back in business.

Popularity: 6% [?]

Related posts:

1. Monitoring Veritas VM problems on Solaris
2. ReiserFS Filesystem Recovery
3. MySQL query cache