Seeding Catalyst-based repo with tape?

[Vejete]

Sites A and B are totally separate Veeam and storage infrastructure. I wanted to use tape to seed VBKs from Site A to Site B then map a backup copy job to them at Site A. I also want the Veeam server at Site B to see the restore points and be able to restore from them at Site B.

I thought this would work, but it did not:
Run backup to tape at Site A. (Source repo was StoreOnce Catalyst.)
Ship tapes to Site B.
Import tapes at Site B.
Restore from tape to repo at Site B. (Target repo is StoreOnce Catalyst.)

Restore finished successfully many hours later. I would expect to see those restore points listed in Disk (Imported) on the Site B Veeam server. They are not there. I checked the Catalyst store, and some 70-ish TB got restored into separate Catalyst items. But there is no corresponding .vbm file. I presume that's why they are not seen even after both rescanning the repo and even removing and adding it back with "import" box ticked.

The WAN link between Sites A and B is microscopic and wholly unfit for the initial transfer of backup copy data. That's primarily why we were trying to seed with tape. Secondly, if Site A is totally lost, we want the surviving Veeam infrastructure at Site B to be able to see Site A's copy target repo and restore from it.

I'll admit to being a bit thick at times. Am I missing something? It shouldn't be this hard to seed a remote copy with tape or make those restore points usable by a foreign Veeam server.

[HannesK]

Hello,

The WAN link between Sites A and B is microscopic and wholly unfit for the initial transfer of backup copy data.

I don't see a way of seeding with tapes. The documented way to seed a backup copy job is described in KB1856.

Best regards,
Hannes

[Vejete]

When we restored the 100+ VMs from tape to a repo at Site B, it created Catalyst items for each .vbk but no .vbm metadata item was created. So we restored a single VM from the media set to a repo at Site B. One .vbk and one .vbm - good. We then restored six (6) VMs from the media set to a repo at site B. Six .vbk's and one .vbm - good. Then the same with twelve - good. We tried an even larger batch overnight - results TBD.

I have no answer as to why restoring all .vbk's in the media set did not also trigger the creation of a .vbm metadata file. And since this is Catalyst, I cannot simply re-create a .vbm file for the whole lot and then copy the file to the repo.

[foggy]

Hi Jim, basically, if you can backup to a Catalyst repository, you should be able to restore backups from tape there. We've just checked this internally and backup files were correctly written (including metadata specifics, since this is the main issue with why you cannot simply copy backups to a Catalyst Store) and VBM was there. So these backups can be used for seeding a backup copy job.

The question why there were no VBM in your first test is still open though.

[Vejete]

Hi Alexander,

That was our question, too. The original restore finished successfully, and we saw the resulting .vbk items in the Catalyst store. But because the .vbm item was not created at the end of the job, Veeam refuses to acknowledge the presence of those .vbk items in Disk (Imported).

Trying to seed with tape has turned into a real nightmare because when you map a copy job to the target store, you see only a list of job names which in our case is multiple instances of the very same "Copy from Tape..." name.

So we're still none the wiser. We have two sites separated by a pathetically undersized WAN link (50 Mb), and we need to do backup copy from prod site to DR site. Seeding the DR site is the only solution I can envision to get the first batch of .vbk items to the DR site via tape. Seeding as documented in https://www.veeam.com/kb1856 is not possible.

[foggy]

Haven't you opened a case to investigate the VBM issue? Seems using some other media for seeding would be a better choice.