Enhancement Request for Tape Jobs and Tape Mounts

[FredMez]

I would like to suggest an enhancement to the tape job processing as it relates to tape mounts. If this has already been discussed, my apologies.

When a tape job is awaiting a tape mount, all other tape jobs queue up behind the awaiting job. The awaiting job should release the tape resource so that other jobs may process, and check for the tape availability whenever the tape resource is available.

For example, job1 wants a tape mount for a tape from media-pool1. There are no tapes in the library that are available from said media pool. Then job2 comes along and wants a tape from media-pool2, which are available. currently job2 waits because the tape drive is not available. why can't job2 process while job1 still waits for its tape. Then after job2 completes and the tape resource is available again, it can check to see if a tape is now available in the library for media-pool1.

The current configuration causes all tape jobs to queue up, creating a huge mess as backup data gets changed.

thank you,
Fred

[veremin]

So, there are two backup to tape jobs that have different backup jobs or repositories as their sources, the tape jobs are pointed to two different media pools that are created on top of different libraries, right? Thanks.

[pkelly_sts]

Useful suggestion. Personally I wouldn't assume that they're also in different libraries...

[veremin]

In that case the question would be - do those media pools have different tapes assigned to them, instead of sharing free media pool?

[FredMez]

job1 is a daily job that uses tapes from pool1 (which has a specific retention period). job2 is a weekly job that uses tapes from pool2 (which has a different retention period). If job2 doesn't have a tape available in the tape library for that pool, it locks the tape drive waiting for a tape mount. Meanwhile the daily job cannot use the tape drive.
To compound the issue, once the job starts it does not check the retentions of tape again. Meaning, that while the job is waiting for a tape in pool1, if a tape becomes expired in that pool, the awaiting job never sees it unless i manually mark it as free.
The tape management in Veeam is a good starting point, but it definitely needs maturing.

[veremin]

And just to clarify - how many drives does your library have? Thanks.

[FredMez]

it is a single tape drive.

[veremin]

When a tape job starts, it automatically locks a drive. The job releases the drive as soon as it's is finished (either successfully or not). Thus, what you currently see is an expected behaviour. The tape drive was locked, because the tape job was not finished.

Anyway, thank you for your feedback, we will think how the current implementation might be improved in future; appreciated.