Repopulate Repository from tape?

[kubimike]

Having issues with my HP controller, this morning my repository went offline because of an HP related hardware issue. My question is if my tapes contain full backups and no incrementals can I re-populate the repository in the event of a hardware failure ? I want to avoid rebuilding the repository with active fulls because of long they take to create during the work week.

[veremin]

My question is if my tapes contain full backups and no incrementals can I re-populate the repository in the event of a hardware failure ?

Yes, you can. Restore backup files first, then, map either existing jobs or jobs restored from configuration backup to those backup files. In this case initial job cycles will be incremental.

[kubimike]

@v.Eremin thank you. Is there documentation for this somewhere ?

[veremin]

This section describes how backups can be restored from tapes, this one - how backups can be mapped to data already present in a repository.

[pkelly_sts]

If you have a disk repository with say, 30 RPs, and all of those RPs are on tape via a copy to tape (not file to tape) job, if you then lose the disk repository for any reason, are you able to restore the latest tape-based backup to a newly created repository, then seed the original backup job to it?

If yes, then what gets written to tape again first after things pick up again, presumably a full copy of a full backup to tape?

The scenario I'm talking about is for a backup copy job at a remote site, hence the desire to be able to re-create a seed from the tape based copy of that backup.

I hope I'm explaining things clearly enough!

[veremin]

Hi, Paul,

I think your question is more or less similar to the one discussed above (even topic names look similarly). Thus, I've merged your post into this thread.

If anything is still not clear to you, let us know.

Thanks.

[pkelly_sts]

Ah yes, it does look very similar, thanks. I'll kick off a restore from tape & give it a go, thanks.

[veremin]

Kindly, let us know about the results of your findings and whether everything's gone as expected. Thanks.

[pkelly_sts]

Will do. As a side but related query, I've seeded backup copies many times but always on the basis of either copying the original full backup files or a purposely created local backup copy which then gets shipped to destination & switched to.
As part of our recent moves, I've effectively created a brand new backup job (lets' say Local-02) but of the same VMs as the original job (Local-01), and Local-01 was originaly copied to remote site & working for months (RemoteCopy-01).

Is it possible for me to create a new copy job (RemoteCopy-02) but of the new local job as source (Local-02) but map that to the pre-existing RemoteCopy-01 backup files? The backups are of the exact same VMs (except for one which had to be removed & re-added to vcenter so I'm aware that'll result in a full copy) but as the underlying VMs are exactly the same, should this plan work in theory? I'm guessing/wondering if the recreating digest process will effectively tidy things up?

Hopefully this makes sense!

[pkelly_sts]

Kindly, let us know about the results of your findings and whether everything's gone as expected. Thanks.

Hmm, restore from tape went ok, I could see it then automatically imported the restored backup into Veeams catalog, I mapped the copy job to it (the newly restored data) fine, but first time I ran the copy job it has kicked straight into an active full.

What I noticed immediately was that it didn't attempt to perform a health check (which is what's happening with another job going in the reverse direction which was also newly mapped to an existing backup set but which is currently processing a health-check as I'd expect it to.

[veremin]

Is it possible for me to create a new copy job (RemoteCopy-02) but of the new local job as source (Local-02) but map that to the pre-existing RemoteCopy-01 backup files?

Correct.

Hmm, restore from tape went ok, I could see it then automatically imported the restored backup into Veeams catalog, I mapped the copy job to it (the newly restored data) fine, but first time I ran the copy job it has kicked straight into an active full.

Given mapping went properly, the experienced behaviour looks unexpected and should be investigated within support ticket.

Thanks.