Suspend prevents new jobs from selecting the tape. Although I’ve not seen this myself it makes sense to me that a suspend done after
the tape was assigned would have no effect on that existing assignment. To have a backup release a tape the best way to do it would be to suspend the backup job then suspend the tape. This is only an option if you’re using checkpoints on the backup so
it can be resumed (and pick a different tape on the resume).
There are ways to modify the database to clear resource assignments but doing that while a tape is already mounted and writing would
be a rather idea.
From:
veritas-bu-bounces AT mailman.eng.auburn DOT edu [mailto:veritas-bu-bounces AT mailman.eng.auburn DOT edu]
On Behalf Of Mark Hickey
Sent: Wednesday, October 05, 2011 10:48 AM
To: veritas-bu AT mailman.eng.auburn DOT edu
Subject: [Veritas-bu] Question about SUSPEND behavior
Hi folks,
My customer has implemented an offsite strategy that calls for tapes from the previous night’s backups to be suspended after the backup cycle so that they can be completely
replicated offsite from the production VTL. A job is run every day at 14:30 pm that suspends the tapes written in the past 24 hours, and a report is published every day at 17:30 pm that lists the tapes that from the backup cycle and whether they are synch-ed
to the DR location or not.
Yesterday we noticed that a tape that was listed as synched on the report from 10/03/2011 was not synched on 10/04. When we checked the logs we found the following:
1.
tape was used for backup during the 09/30 - 10/01 backup cycle, with the job ending sometime around 23:00 on 10/01.
2.
tape was picked up by the suspend job on 10/01 because it was written in the past 24 hours
3.
tape was reported as fully synched to the DR VTL on 10/02 and on 10/03
4.
tape was selected for a backup late on 10/03, and was finally given a tape drive early on 10/04 and started writing (this is, one would assume, why it was no longer synched)
So it appears that the tape was still writing when the suspend script ran. I guess we expected that the tape would go to SUSPEND mode when the write finished, but 4. above seems
to belie that expectation.
Can anyone comment on the expected behavior if an attempt is made to suspend a tape when it is writing?
Thanks,
Mark
Mark Hickey
Principal Technical Consultant - Business Continuity
Hitachi Data Systems
North Weymouth,
MA
All phones: 781-817-3055