We recently migrated a 5 server NCS cluster to a new SAN using NSS mirroring. Everything went swimmingly. After the migration, because of a scheduled power outage, we had to down the entire cluster.

When the first server came back up, it activated all the NSS Pools for all the cluster resources. This would prevent any of the resources from coming online on any but the first server that booted. The volumes didn't mount, but the pool is activated.

On the remaining servers in the cluster, each of the servers attempts to activate every NSS pool. This adds about five minutes to each server boot process.

Activating pool "NCSPOOL1"...
** Pool layout v43.02
** Pool ownership detection (this may take 30 seconds to complete).

Aug 26, 2008 4:48:54 pm NSS<COMN>-3.26-689: comnPool.c[407]
Could not change pool NCSPOOL1 to the ACTIVE state.
Status=21621 zlssMSAP.c[1220].
Use 'NSS /ErrorCode=21621' to obtain more information.

I've tried NSS /POOLAUTODEACTIVATE=ALL, as well as individually specifying each of the poolnames in the /POOLAUTOACTIVATE command, to no avail.

The servers are NW6.5 SP7.

I was of the understanding that NSS honors cluster resources and doesn't auto activate the pool or mount the volume.

Any thoughts on this?