Need help understanding the "timing" of tape jobs

[bhagen]

If I understand correctly, veeam will only backup files or Veeam backups.

When backing up a backup job to tape, if the parent backup job fires off, the tape job will stop. Optionally, we can check the "don't let backup job interrupt tape job" box, and the parent job will wait until the tape job is complete before running.

So, say I'm going to schedule a GFS rotation to tape. Every month, all my VBR jobs that are being backed up to tape are going to have to pause until the tape job is complete? The initial tape jobs will total about 115TB, give or take. With LTO 7 tapes at 5.5TB/tape, that's 21 tapes. (Not going with LTO8 until the lawsuit is settled.) I just did my first backup to tape test, and an 886.5GB job took 4.5Hr to complete. That equates to almost 5.5 days straight of backups to tape. During which time, I'm not getting regular nightly backups, since they are waiting on the tape job to finish.

Is this how it works for real? If so, how are people backing up huge amounts of backups to tape, without destroying their RPOs?

(Backing up files to tape is significantly faster, but VSS on the file(s) is attempted...so I'm assuming that if VBR tries to backup to a locked file, the backup will fail, so even that way RPOs are missed.)

So...would I instead do a GFS rotation to NAS, then backup the GFS backups to tape (since GFS will only run once a month)? I suppose that might be the way?

[tohhmas]

Yep, that's the way it works unfortunately.
Veeam Support's advice has been to split the tape job into smaller jobs so you can cycle through the source jobs allowing them to run.

Interesting thought on the Backup Copy -> Tape, it seems like it's supported: tape-f29/backup-copy-to-tape-t50802.html

[Dima P.]

Hi bhagen,

Is this how it works for real? If so, how are people backing up huge amounts of backups to tape, without destroying their RPOs?

Since Veeam B&R follows Disk to Disk to Tape approach the main idea would be to schedule tape job run at the different time after source backup job is completed. Another option would be to do the tape outs at the end of the work week when you do not run your primary backup jobs.

Optionally, we can check the "don't let backup job interrupt tape job" box, and the parent job will wait until the tape job is complete before running .... If so, how are people backing up huge amounts of backups to tape, without destroying their RPOs?

This checkbox is designed to save tape job from conflicts made by backup to disk job (for example vm backup running in forever forward incremental and performing merge or job with periodic fulls performing retention on retired backup chain).

So...would I instead do a GFS rotation to NAS, then backup the GFS backups to tape (since GFS will only run once a month)? I suppose that might be the way?

Yup, you can setup backup copy job and then use it as a source for backup to tape job.

[bhagen]

So...would I instead do a GFS rotation to NAS, then backup the GFS backups to tape (since GFS will only run once a month)? I suppose that might be the way?

Yup, you can setup backup copy job and then use it as a source for backup to tape job.

Ok, new information is that we'll only be using tapes for yearly archiving, keeping 5 years.

So in this scenario, it may be easier to avoid backup and tape backup conflicts.

Does this sound right:
a) Run a normal backup job (to disk - onsite nas)...32 restore points
b) Run a backup copy job (to disk - offsite nas)...32 restore points
c) Run a GFS backup copy job (to disk) off the backup copy job in step b...monthly (11 restore points), quarterly (3 restore points), yearly (5 restore points)
d) Run a backup to tape job of the yearly GFS backup copy job in step c (to tape, obviously) once a year (after the yearly GFS backup copy job completes)

Correct?

Now let's say I've run this scenario for 3 years, and I need files restored from a year ago. Lawsuit. Go figure.

Will I:
a) Restore those files directly from tape (step d, above)
b) Restore the GFS backup copy job (acquired in step d, above) from tape to disk, then restore the files from the job I just restored to disk

Hopefully step a...

Thoughts?

[aich365]

We regularly backup to tape and complete 2.3TB on LTO-06 on HP Autoloader in around 5 hours.
One of the things that caused us issues was using SMB3 with encryption turned on. This dropped throughput dramatically
This had been turned on by GPO
So if using SMB check connections

Get-SmbConnection
get-SmbServerConfiguration
Get-SmbClientConfiguration

Look for
EnableSecuritySignature : True
RequireSecuritySignature : True

Regards

[Dima P.]

Does this sound right

Sounds like a plan. However, if you are running Backup Copy GFS just for tape jobs - you can remove the middleman and use native GFS tape media pool functionality.

Now let's say I've run this scenario for 3 years, and I need files restored from a year ago. Lawsuit. Go figure.

You might want to keep some of the Veeam One tape reports next to the archived tape media (to get the proper information about 'where is what'). Check this Help Center - Veeam Backup Tape Reports.

Will I:
a) Restore those files directly from tape (step d, above)
b) Restore the GFS backup copy job (acquired in step d, above) from tape to disk, then restore the files from the job I just restored to disk

You have to restore backup file to a staging repository and then perform file level recovery from the backup file.