Prevent tape job from interrupted by source backup jobs

[Troubleshooter_0815]

Hi Community,

we have configured a simple "weekly backup to tape" job.
The Target ist a Single LTO6 SAS Drive in a a HPE Library. The job takes about 48 hrs to complete, caused by the datasize.
The performance is expected, so it depends only on the datasize of about 12TB, written to a set of tapes.

For sure, if the tape backup is running, it's not possible to start a daily source job, because the Tape Job will stop, as expected.
If we change the priority of the tape job, the daily job will wait for the 48 hrs, also expected.

So is there any workaround to run the daily job without interrupt the "weekly tape job". It's enough if we could write the last full backups of Friday evening to tape.

Maybe it could be a solution to stage the daily jobs in a dedicated repository (yes double space for the fulls is required) and choose the stage repo as a source for the tape backup?
If I tried this, I struggled with the scheduling, because the only scheduling fpr copy jobs is every X days or continously, but I only want to run this copy job i.e. on friday evening and I want to pick the fulls after the copy job by the tape job. Did I missed some settings here? I'm sure I'm on the wrong page somehow.

Maybe there are other thoughts on this?

Thanks for any suggestion...

[HannesK]

Hello,
and welcome to the forums.

Can you tell us which backup mode (forward with active / synthetic full?) you are using?

It's enough if we could write the last full backups of Friday evening to tape.

so you selected "only full backups" to tape? Meaning you don't have checked the incremental pool: https://helpcenter.veeam.com/docs/backu ... ml?ver=100

By the way, 48h for 12TB is only 73MByte/s... that's much slower than LTO6 can do.

Best regards,
Hannes

[Troubleshooter_0815]

Thanks for suggest, we using only Reverse Incremental Jobs and also we enabled the "Do not backup incremental" option.
Ok, the 48 hrs are only a number, lets say the job is running for more than one day, that's also too slow to match the RPO requirements, because the source jobs have to wait for too long.

Yes, we also thought about Tape increments and a seperate media pool, but this is a kind of unhandy for us, so we prefer one media set with all the FULL's of Friday, that we can take offsite. Let's say, they are a bit scared about the FULL Tapes will be offsite for months and maybe worst case the only tapes with FULLs are unreadable.

BTW, it's absolutely clear Veeam can't perform any kind of magic, only one 6 Gbit/s SAS Drive is too slow to do the job as expected or required.
I thought there would be a solution to seperate the source jobs from the tape jobs "anyhow" with a workaround or else.

[HannesK]

Hello,
Reverse Incremental is the key point... it's impossible to run tape and backup job in that case at the same time. That's because the backup file changes during backup.

My suggestion would be to switch to forward incremental with synthetic full on REFS / XFS. https://www.veeam.com/blog/advanced-ref ... suite.html

Best regards,
Hannes