Hi,
During a storage migration project, we will do the following:
suppose volumegroup testvg containing 2 disks:
hdisk10 and hdisk11.
logical volumes are mirrored on both disks.
we want to umount the logical volumes, and then block access to hdisk11 (to use it in case sth. goes wrong). aix will no longer see hdisk11 and mark it missing (and stale in lvm terminology).
we will then add a copy on a new disk hdisk12. (so we would have a good copy on hdisk10, a stale copy on hdisk11 and again a good copy on hdisk12.
suppose during the synchronization of the pp's from hdisk10 to hdisk12 sth. goes wrong and hdisk10 becomes corrupt, breaks down or whatever.
the idea is then to use the 'stale' copy on hdisk11 like this: make hdisk10 unavailable to aix. the volumegroup will probably varyoff because there are no more valid disks (depends on # disks in vg and quorum settings). then we would like to force the volumegroup backonline (using the -force parameter and setting the MISSINGPV_VARYON parameter). in this case, the only visible disk would be hdisk11 (which was previously stale).
and now it comes: will we be able to do this ? will the contents on this disk still be considered stale since it will be the only copy available ?
thanks,
Tom.
During a storage migration project, we will do the following:
suppose volumegroup testvg containing 2 disks:
hdisk10 and hdisk11.
logical volumes are mirrored on both disks.
we want to umount the logical volumes, and then block access to hdisk11 (to use it in case sth. goes wrong). aix will no longer see hdisk11 and mark it missing (and stale in lvm terminology).
we will then add a copy on a new disk hdisk12. (so we would have a good copy on hdisk10, a stale copy on hdisk11 and again a good copy on hdisk12.
suppose during the synchronization of the pp's from hdisk10 to hdisk12 sth. goes wrong and hdisk10 becomes corrupt, breaks down or whatever.
the idea is then to use the 'stale' copy on hdisk11 like this: make hdisk10 unavailable to aix. the volumegroup will probably varyoff because there are no more valid disks (depends on # disks in vg and quorum settings). then we would like to force the volumegroup backonline (using the -force parameter and setting the MISSINGPV_VARYON parameter). in this case, the only visible disk would be hdisk11 (which was previously stale).
and now it comes: will we be able to do this ? will the contents on this disk still be considered stale since it will be the only copy available ?
thanks,
Tom.