Two competing jobs

[dspringer]

I need your swarm knowledge please, because I'm a little bit out of my depth right now.

To explain:
Recently, our backup jobs were edited by a service provider. The old ones had 2 different Backup2Disc jobs for each server group for day and week, with a lot of Synthetic Full Backups. Therefore they were very time consuming. Each of them had an own Tape job with a corresponding tape-pool.

After the rebuild, there is now only one job per server group with disk backup. These are Incrementell, where Saturday a full backup is created. As secondary target are 2 tape jobs for monthly and weekly backup. Behind this tape job is accordingly a suitable media pool for tapes.

My problem:
The weekly job for tape runs Sunday from 8am to Monday (round about 4:30pm) and currently consumes 6 LTO7 tapes (an upgrade is already under construction).
The monthly job for tape starts on the last Sunday of each month, starting at 1am. Consequently, the monthly job starts on the last Sunday, but is then cancelled as soon as the weekly job starts.

Question:
What would you guys do now? My thoughts would be:
- All tapes in a single pool and then just bring up own discipline, take the tapes intended as monthly backup from the tape robot and lock them
- To bind the job directly to the last day of the month is not possible.
- To set both tape jobs in dependence on each other is apparently not possible.

[HannesK]

Hello,
and welcome to the forums.

Therefore they were very time consuming

sounds like REFS / XFS should be used Or "active full" backup to avoid merges.

Did you think about a GFS media pool? https://helpcenter.veeam.com/docs/backu ... ml?ver=110

Best regards,
Hannes

[dspringer]

Hi There

I think this is fine now and all jobs vor over 40 vms are smooth now. My main problem are both tape-jobs.

[Dima P.]

Hello dspringer,

The weekly job for tape runs Sunday from 8am to Monday (round about 4:30pm) and currently consumes 6 LTO7 tapes (an upgrade is already under construction).
The monthly job for tape starts on the last Sunday of each month, starting at 1am. Consequently, the monthly job starts on the last Sunday, but is then cancelled as soon as the weekly job starts.

Is that a GFS media pool or simple media pool? With GFS media pool whenever you have two backups to different media sets scheduled on a same date tape out is performed only to the elder media set (so in your case that should be monthly).

[dspringer]

Damn, you're totally right! Since our LTO9 will hopefully be activated in 2 months at the latest, I will retire our normal media pools in that case and use a GFS right away.
Thank you for the tip! The variant I had until just no longer in the field of vision.

[Dima P.]

David,

Sounds great! Let us know how the migration to GFS goes and if you need any further assistance. Thank you!

[dspringer]

Small update from my side

Yesterday we had the first time the GFS with new hardware and LTO8 run. In reduced form we use here nevertheless only monthly and weekly backup. The weekly backup is set to run every Sunday and the monthly backup every first Monday.

This was the case on this weekend. Sunday started the tape job, described 3 tapes as "Weekly media". The day changed to Monday, another tape was needed and this one was then named "Monthly media".....

"Append backup files to incomplete tapes" if I read correctly refers to the next run in which the tape could be used?
I have seen that there is still some free space on the 4 tapes.
( 1.9 TB from 10,9 TB // 4,2 GB from 10,9 TB // 9,1 TB from 10,9 TB // 3,6 GB from 10,9 TB )

[Dima P.]

Hello David,

Check the append configuration in GFS media set properties , looks like the append option is enabled which means every new tape job run will try to reuse all the free space from previously started tapes within specific media set.

[dspringer]

Sorry, I totally forgot to get back to you on the subject. You were right and everything now fits without any problems.
The topic can be closed.