for new to AIX this could be a surprise – a volume group with active dump devices cannot be varied on. Well, one my ask – “why do I need to vary an active volume group – (presence on an active (dump) device (logical volume) indicates an active volume group, right)?
Let’s use my case of this morning – I had to build a few
lpars but could not proceed because the two VIO servers providing virtual FC and SCSI adapters reached the end of their slot ranges (
maximum virtual adapters). To activate the new maximum number a VIO server need to rebooted.
The two VIO servers provide disks to lpars rootvg’s which are mirrored by LVM. Rebooting one VIO server automatically spoils lpars
rootvg's because of the momentarily loss of connectivity with the disk provided by the VIO server being rebooted. To recover the missing disk and to restore the
rootvg to its fully operational state, one have to re-acquire the disk and to synchronize the mirror. These steps could be done executing a single command –
varyonvg. It does not matter that the volume group is already ON, after verifying access to the previously missing disk the mirrored volume group will be automatically synchronized. But before this may happen, any active dump logical volumes must first be temporarily disabled.
So what was my process? OK, first I booted the VIO server with the passive
SEA adapter. Next, I repeated the following steps on all partitions:
Turn off the primary dump device (your name might be different, use the
sysdumpdev command for answers).
# sysdumpdev -P -p /dev/sysdumpnull
Turn off the secondary dump device.
# sysdumpdev -P -s /dev/sysdumpnull
rootvg triggering an automatic sync of
stale physical partitions.
# varyonvg rootvg
After all partitions
rootvg's were fully synced, I logged into the second VIO sever, and failed its SEA (primary) adapter to the previously rebooted VIO server, reboot followed.
Again, all partitions went
stale and the
varyonvg rootvg command was called for and followed with re-activation of the dormant dump volumes.
# sysdumpdev -P -p /dev/dump0
# sysdumpdev -P -s /dev/dump1
Remember, in your case the names associated with the dump devices and their number may be different.