Video Screencast Help

vxdisk list

Created: 23 May 2013 • Updated: 28 May 2013 | 4 comments
This issue has been solved. See solution.

Hi people need some help vxdisk list on system 1 has 7dsk and system2 has 6dsk what could have happened

root@system1:~]# vxdisk list
DEVICE       TYPE            DISK         GROUP        STATUS
sda          auto:none       -            -            online invalid
sdb          auto:cdsdisk    -            -            online thinrclm
sdd          auto:cdsdisk    -            -            online thinrclm
sdf          auto:cdsdisk    -            -            online thinrclm
sdh          auto:cdsdisk    disk02       appsdg       online thinrclm
sdj          auto:cdsdisk    disk03       appsdg       online thinrclm
sdl          auto:cdsdisk    disk04       appsdg       online thinrclm
sdn          auto:cdsdisk    disk05       appsdg       online thinrclm
sdp          auto:cdsdisk    disk01       appsdg       online thinrclm
sdr          auto:cdsdisk    disk06       appsdg       online thinrclm
sdt          auto:cdsdisk    disk07       appsdg       online thinrclm
[root@system1:~]#

[root@system2:~]# vxdisk -o alldgs list
DEVICE       TYPE            DISK         GROUP        STATUS
sda          auto:none       -            -            online invalid
sdb          auto:cdsdisk    -            (vxfencoorddg) online thinrclm
sdd          auto:cdsdisk    -            (vxfencoorddg) online thinrclm
sdf          auto:cdsdisk    -            (vxfencoorddg) online thinrclm
sdh          auto:cdsdisk    -            (appsdg)     online thinrclm
sdj          auto:cdsdisk    -            (appsdg)     online thinrclm
sdl          auto:cdsdisk    -            (appsdg)     online thinrclm
sdn          auto:cdsdisk    -            (appsdg)     online thinrclm
sdp          auto:cdsdisk    -            (appsdg)     online thinrclm
sdr          auto:cdsdisk    -            (appsdg)     online thinrclm
sdt          auto:none       -            -            online invalid thinrclm
sdu          auto:none       -            -            online invalid thinrclm
sdw          auto:none       -            -            online invalid thinrclm

 Thanks in Advance

Mig31

Operating Systems:
Discussion Filed Under:

Comments 4 CommentsJump to latest comment

stinsong's picture

Hi mig31,

From the output I understand you are meaning missing one disk belongs to appsdg.

You can use vxdisk list to check which one is missing. And use OS command to list device tree if there is disk missing from OS level.

Typically it's OS device level issue. But remove disk.info and dmp node device to restart vxconfigd could be helpful.

================================

Deport all DGs.

# rm /etc/vx/disk.info

# rm /dev/vx/*dmp/*        

# vxconfigd -k -r reset

================================

SOLUTION
g_lee's picture

MIG31,

As vxfencoorddg exists, this implies VCS is being used.

If you have VCS, before performing the cleanup steps / resetting vxconfigd, you should freeze the service groups containing DiskGroup / Volume resources to avoid unexpected failovers (ie: so the cluster won't try to probe/failover while you are doing the reset/cleanup)

See the following technote for example/steps to freeze the service groups in cluster (if you are not using powerpath, ignore the ppath specific steps; the cluster parts are still valid):

How to perform a device tree cleanup with EMC storage, EMC Powerpath and Volume Manager in the configuration on a Solaris system

http://www.symantec.com/business/support/index?pag...

If this post has helped you, please vote or mark as solution

MIG31's picture

Thanks,g_lee for the info it was helpful the first solution worked thanks

Peter

g_lee's picture

Peter,

To be clear - both technotes ultimately resolve the device tree issue. stinsong's solution cleans up the disk/device tree issue only. If you are only using VxVM (not using VCS) then this isn't an issue.

The technote provides additional steps to perform if you are using VCS - so it does the steps in stinsong's post, but with additional steps before and after for cluster. If you are using VCS and you do not perform the additional steps, you can get unexpected failovers.

You might be able to get away with not performing the additional steps if you happen to be lucky/it doesn't happen to probe when you're cleaning up, but it's an extra precaution to avoid unnecessary outages.

regards,
Grace

If this post has helped you, please vote or mark as solution