Linux SCSI reservation

[TJ_16]

We are running RedHat AS 2.1 in a clustered environment using VCS on Proliant DL580 G2's. We have Emulex LP9002L HBA cards attached to SAN and shared tape.
We are seeing a problem where if a NetBackup recovery of a database is being done on one node and the other node is rebooted, the recovery fails. This fails because as the NBU daemon starts up it does something to the shared tape drives and the recovery fails.

We believe this has something to do with SCSI reservation, but are unable to find a solution. Anyone out there have/had a similar problem and a solution?

[Steven E. Protter]

I would suggest openning a call with Veritas. This is probably a problem with the daemon settings, versus SCSI.

I reacall reading something on their website that your symptom reminds me of. Sorry I can't be more specific.

SEP

[TJ_16]

Funny you should mention that as we do have a Veritas ticket opened and several Veritas folks working on it... they are saying OS issue, we are saying Veritas issue.

Just wanted to see if any users out there have seen this.

Thanks for the reply.