removing primary disk in cluster volume group

[stranger_2_all]

Hi all,
currently, im having issue with one of the disks in my clustered server. Below are the disks configured on that clustered VG with RAID 5. The /dev/dsk/c13t2d4 is the primary disk however, it throws error of scsi read/write hence was replaced. However, it still persists after the replacement therefore suspecting that the cause is not the disk but something else.

--- Physical volumes ---
PV Name /dev/dsk/c13t2d4
PV Name /dev/dsk/c6t2d4 Alternate Link
PV Name /dev/dsk/c10t2d4 Alternate Link
PV Name /dev/dsk/c17t2d4 Alternate Link
PV Name /dev/dsk/c4t2d4 Alternate Link
PV Name /dev/dsk/c8t2d4 Alternate Link
PV Name /dev/dsk/c19t2d4 Alternate Link
PV Name /dev/dsk/c15t2d4 Alternate Link


Will my issue resolve by removing the primary disk on that clustered VG?

[Shibin_2]

You have only one disk and all other are alternate link to the one disk. If you remove the primary path, the disk will still be accessible through your alternate path.

What is your OS version? You said, there is SCSI read/write error. How did you identify it? Through STM ? What's your STM version running ?

[stranger_2_all]

You have only one disk and all other are alternate link to the one disk. If you remove the primary path, the disk will still be accessible through your alternate path. --> does it mean that it'll have no impact to out data if this primary disk be removed permanently in that VG?

What is your OS version? 11.23
You said, there is SCSI read/write error. -> error messages is like spam in our syslog.

[P Arumugavel]

If your disk has multiple PV-links, the disk is not removed from the volume group, until all the links to the volume group are removed. When a physical volume in the pv_path list is the primary link (in use) to a physical volume, removing the link forces LVM to switch to the alternate link to access the physical volume. When the pv_path removed is an alternate link to the device, only the link is removed, the volume group and physical volume are otherwise unchanged.

Rgds...

[stranger_2_all]

thanks all..im somewhat enlightened.

removing the primary disk path has no implication to the data residing on the disk. Im still a bit doubt to remove it as even next alternate path sometimes failed and it wouldnt help to resolve my issue.

last question. will there be any alternative or temporary solution from platform side as even our offline backup got affected.

[Torsten.]

The device /dev/dsk/c13t2d4 with so many alternate paths doesn't look like a single disk, but it is most likely a LUN provided by a hardware RAID, right?

Replacing a disk in a storage array should never affect a connected host.

[INH]

mentioned disk is from direct atatched storage or a LUN from SAN box?

output looks like LUN from SAN, replacing disk in SAN DAE having RAID5 should not have any impact at host side.

[stranger_2_all]

these are EVA SAN disk and configured EVA card. our storage team replaced the complaining disk but the error still persists (same error messages et al). our backend storage team still taking time to identify the root cause and im now thinking what would be the best alternative/temporary solution as even our offline backup (SAP/ORACLE) is affected.

[Viktor Balogh]

maybe you could present us with some lines of your syslog to let us see what happens on your system... ;)