I recently migrated some VMs from some host to another because I was doing storage maintenance on the hosts. I did not notice until a week or two after the migration, the VMs that were migrated either got their disks corrupted (asks to run fsck at boot time) or their filesystem mounted as read-only.
I’ve no idea why such behavior occur. I checked the hosts in Sunstone and found out that some hosts have zombie VMs and they have identical IDs to the VMs that experienced the issue.
My Datastore is using iSCSI with OCFS2 on each hosts to form a cluster.
Now my questions are:
Does anybody have any insights what caused the zombie VMs?
How to get rid of them?
Any advice how to fix the corrupted disks/read-only issue?