We have an issue whereby when our weekly backups run longer than usual the duplication to tape naturally also runs late (obviously expected) - now when this causes us a problem is when it then causes the following daily backup to queue (weekly and daily tapes are different media pools). The scenario is this;
- VM to Tape job detects both full and incremental backup files ('weekly' and 'daily')
- VM to Tape job locks the drive AND the weekly pool, and begins backing up
- File to Tape job incremental is triggered (time-based trigger)
- File to Tape job locks the daily pool, but is unable to continue due to no drive
- VM to Tape job completes the weekly duplication, and begins the daily duplication
- VM to Tape job is unable to continue due to the daily pool being locked by the File to Tape job, but doesn't release the lock on the drive, creating a deadlock - neither job proceeds. Ever.
Support (case # 00882109) haven't been overly helpful so far (suggested breaking up the media sets which isn't practical due to tape utilisation issues) - to me it seems like a logic bug in the software (why lock a media set when you're unable to lock a drive to use that media) - would be extremely happy if someone could point out that it's something that I'm doing wrong!
Cheers,
Ben.