Backup Jobs Stopping Tapes - v9

[riotarms]

Afternoon Veeamers,

I have the ye old problem of tape jobs being stopped by backup jobs even though:

1. Jobs are linked
2. Wait times are implemented

I thought they have improved tape // backup wait for linked job so the file back up doesnt barge its way in and kills the tapes.

Oddly enough it doesnt occur with ALL jobs only a few. Im curious to see what will happen if i recreate them (These jobs were carried over from 8 to 9) however just wanted to pick the brains trust.

Any thoughts?

[PTide]

Backup Job has a higher priority than Tape Job thus if Backup Job needs an access to a backup file that is being used by Tape Job then the Tape Job will be stopped. Could you please provide more details about your job configuration?

Thank you.

[riotarms]

We have 7 file servers that get backed up to file @ 12pm and 6pm everyday using standard incremental and CBT with optimum compression.

With all the file backups we have a secondary target enabled which is the tape library. This job is ordered to take the full backup only.

I would assume when the incremental merges with the full it proceeds to lock the full file however this is only occurring with 2 tape jobs. The other 5 have no problems with the backup job stopping and they have larger and more accessed files. All jobs are configured identically as well.

[veremin]

Which backup mode do those jobs use? Are you saying that some backup jobs do not stop running tape jobs that are copying full backups that are being transformed at the moment? Thanks.

[riotarms]

Hiya,

That is correct.

I have figured out that someone on my team has recreated the jobs therefore the file back ups have not proceeded to merge causing the fulls not to lock.

The jobs that work havent reached their retention points yet.

Besides scheduling what other options are there to make file backups wait for tapes?

I looked at scripts and i can code however im not sure what language or format is required for scripting. I assume PS?

[veremin]

I have figured out that someone on my team has recreated the jobs therefore the file back ups have not proceeded to merge causing the fulls not to lock.

It explains the situation indeed. The full backups that are not being transformed at the time of backup can be archived to tapes smoothly.

Besides scheduling what other options are there to make file backups wait for tapes?

Pre-job activity that checks whether there are any running tape jobs and postponing a backup job in case of positive answer should do the trick.

I looked at scripts and i can code however im not sure what language or format is required for scripting. I assume PS?

We have a PowerShell snap-in using which you can script the process described above. If you need a help with scripting, kindly, create a topic on our dedicated PS subforum.

Thanks.

[Christer]

Hi all

I have almost the exact same problem, but using a GFS tape job. In addition the tape jobs that gets stopped does not clean up after itself.

I have a normal backup job that backs up several fileservers, and then a linked GFS tape job that starts at midnight every week (not possible to change start hour as far as I can tell).
As I only have LTO5, the backup to tape would take more than 24 hours, so the tape job would always get stopped by the backup job. As far as I can tell it is not possible to implement wait times for the GFS tape jobs.

After the tape job has been running from midnight until 7PM, it gets stopped when the backup to disk job starts. It has then used up 3-4 tapes when it failes. As far as the Veeam GUI shows, there are no backups on these tapes, but all space has been used. If you catalog the tapes, you can see that there is a partial backup on the tapes.

I have made a support case for this (01756660), and uploaded logs and pictures to a FTP. The respons from support was:

Unfortunately, it is behavior by design now since backup job has higher priority than backup to tape and if backup job should run at the time when backup to tape is in progress, backup to tape will fail.
It is not considered as bug, but you can request a feature.

Feature request 1:
Veeam Backup and Replication should be able to clean up after itself when a tape job is stopped mid run, and return the tapes to the free pool. Partial full backups are useless.

Feature request 2:
I want a way to be able to take GFS tape backups in the GUI of large backup jobs without resorting to scripting. As it is now my GFS tape job would fail once a week, filling up 3-4 tapes, and doing this until my library ran out of tapes.

Feature request 3:
It should be possible to select when in a day the GFS tape job should run. Now all of them starts at 00:00.

Feature request 4:
It should be possible to implement wait times for GFS tape jobs

Regards, Christer

[veremin]

I got the your ideas.

For now your options are limited either to tweaking jobs' schedules so that jobs do not overlap or to usage of pre-job script that will put a source job on hold if any tape jobs are running at the same moment.

Thanks.

[riotarms]

No problems at all thanks for the assist team; all makes sense.

[Moebius]

I'm in the same situation as Christer and second all his requests.