Today I had faced a very interesting issue. I was in a position where I need to move the disks from a “old” cluster to a new one. After several troubleshooting steps I found out that the disks has a “active” PR-3 reservation which blocks the access to the disks. This is a expected behavior in a shared disk model which supports PR-3 SCSI commands.
The magic here is :
cluster node %nodename% /clearpr:DISKID
to remove the persistent reservation 😉
More Details about “node cleanup” can be found here