AIX

Hi,
One of my disk is in 'disk missing state'. It is a sharedVG and cluster nodes.
The errpt keeps reporting stale partition error.
lvs are in open/stale state.
In this sceanario is replacing the disk the best practice?
When i do a lsdev the disk is labelled as below.
hdisk3 Available Remote Physical Volume Client

Can you please tell me what does 'Remote Physical Volume Client' mean?

I've never had to do it myself as our clustered VGs are all SAN based, so disk failures are hidden away. I think you would do this as normal on the active node, then after each action, tell the passive node to learn the configuration. This search may prove useful



Regards,
Robin

You have asked in this thread a very similar question and the solution i suggested there can be used here too.

A "STALE" partition means it is mirrored but one of the mirror copies is not there any more. It is possible to remove the failed copy (most probably from the disk in question) from the LV and remirror it to another disk. Do not forget to remove the disk from the VG afterwards.

Also, do a "learning import" on the other cluster node to synchronize the new VG layout. For a learning import to work only one hdisk previously already known as to be part of the VG has to stay. (This is why you have to do 2 learning imports when you replace all disks in a VG.)

That the disk is "Available" means only that its electronics is still working and doing what it takes to communicate with the driver if ther is one. It can still be defective because its magnetic platter cannot be read by the (working) drive electronics.

I hope this helps.

bakunin