Using tape as a backup target
Post Reply
tom4784
Novice
Posts: 5
Liked: 1 time
Joined: Jun 21, 2022 8:06 am
Full Name: Tom
Contact:

Tape from StoreOnce - nothing visible in repo

Post by tom4784 »

Hi,

I'm trying to implement the exact scenario mentioned in this thread:

tape-f29/backups-created-by-backup-copy ... 68603.html

Which is this:
1. Backup to StoreOnce repository
2. Optimized StoreOnce copy of this repository to another StoreOnce on another site
3. Copy to Tape (second site) of optimized copy
...and I'm having the same issue - I can create the job fine, selecting 'Backup repository' as the source and selecting the StoreOnce repo in question. It say '0 bytes' in the size column which was an instant red flag, but when running the job it simply ignores the StoreOnce repo. The repo definitely has things in it which I can browse through using the Veeam console connected to the same Veeam server.

In that thread, it was noted this was an upcoming feature and that was a couple of years ago, I'm running the latest version of Veeam and it appears this hasn't yet been implemented? Could someone confirm?

Kind regards,

Tom

Dima P.
Product Manager
Posts: 13589
Liked: 1332 times
Joined: Feb 04, 2013 2:07 pm
Full Name: Dmitry Popov
Location: Prague
Contact:

Re: Tape from StoreOnce - nothing visible in repo

Post by Dima P. »

Hello Tom,

Thank you for your post, can you please raise a support case and share the case ID?

tom4784
Novice
Posts: 5
Liked: 1 time
Joined: Jun 21, 2022 8:06 am
Full Name: Tom
Contact:

Re: Tape from StoreOnce - nothing visible in repo

Post by tom4784 » 1 person likes this post

Hi,

Case raised, it's: 02739409

tom4784
Novice
Posts: 5
Liked: 1 time
Joined: Jun 21, 2022 8:06 am
Full Name: Tom
Contact:

Re: Tape from StoreOnce - nothing visible in repo

Post by tom4784 »

Support got back to me:

Even though the backup files are visible in the repo (within backup copy jobs in the console) because the original backup wasn't done on the Veeam server it will skip them - despite being visible and readable by the local Veeam server front end.

It was suggested that the backups are copied again to local storage to solve the problem - but this seems a really inefficient way of going about things and wastes (a lot) of space I don't have - I can't really think of another way around it though - has anyone done anything similar?

Seems odd that this has been done 'by design' ?

Thanks

Tom

Dima P.
Product Manager
Posts: 13589
Liked: 1332 times
Joined: Feb 04, 2013 2:07 pm
Full Name: Dmitry Popov
Location: Prague
Contact:

Re: Tape from StoreOnce - nothing visible in repo

Post by Dima P. »

The logic sounds right - tape jobs process only the data created by the B&R server jobs and skip imported backups (which seems to be the root cause with your case). I wonder if it's possible to register tape device with your primary B&R server and add original backup copy jobs from the primary B&R site to the tape job on the same B&R server? Thanks!

tom4784
Novice
Posts: 5
Liked: 1 time
Joined: Jun 21, 2022 8:06 am
Full Name: Tom
Contact:

Re: Tape from StoreOnce - nothing visible in repo

Post by tom4784 »

Sorry for the delay in replying,

Adding the tape library on the other server doesn't work - presents message 'Tape device already registered with another backup server'

I'm really pulling my hair out with this one, to get round it I am doing a copy job from locally attached storage over the wire to locally attached storage on the server with the tape drive, then using a file backup to get the vbk files onto tape. This at least means there is a backup, but it is far, far, far from ideal and I cannot leave it like this long term:

It's massively inefficient - all the backups are already being replicated across the wire, doing it as I am above means it's now doing it twice on the nights tapes are required.
It's a pain to manage - instead of using normal simple jobs to manage it, I now have to have file copies and file level backups, which then need to be restored in a way different to how I could restore from a similar imported backup...this makes it a pain (an pretty much unacceptable) to handover to the team which will manage it in the future.

Ideally I would go back to the drawing board and redesign how it's all set out, but unfortunately that isn't an option and I must say I'm *slightly* frustrated that my design which explicitly stated this was how I was going to manage the backups - storeonce > tape was looked over and given the OK by several Veeam experts.... :roll:

It's really frustrating seeing the backups in the inventory, all imported and looking fine in the console...being able to add them to the tape jobs with no error...everything looking fine...and then they just get skipped...'by design'

If anyone has any suggestions...I would be more than grateful :D

soncscy
Expert
Posts: 636
Liked: 305 times
Joined: Aug 04, 2019 2:57 pm
Full Name: Harvey Carel
Contact:

Re: Tape from StoreOnce - nothing visible in repo

Post by soncscy » 1 person likes this post

> tape was looked over and given the OK by several Veeam experts.... :roll:

I'd definitely take those "experts" to task ;)

Do get it right you have:

Multiple sites (Let's say A, B, C)
Veeam servers at each sites
Storeonce Devices at each site
Tape Server only at one site (let's say A)
Catalyst Copies from Sites B and C to Site A

And you want to do tape-out on site A?

If I am correct, biggest question is why 3 VBR servers? All three sites could be managed by the same VBR server and this would absolutely solve your issue as best I know. (I recall there was an issue with Tape-out from Catalyst Copies, but I think I remember reading it was solved? Maybe someone from Veeam can comment)

I understand this would require a bit of restructuring, but my team and I have done consolidations like this before for clients and at worst it takes a weekend of some clicking; at the absolute worst, if it completely fails, just keep the old VBR servers around until it's up and you can just bring them back online again until you figure it out.

Alternatively, if your tape device supports library partitioning, you can spin up a partition + server for each site, add said servers to each site, and then initiate the tape jobs from the respective sites. I don't like this solution personally as it's a lot of complexity, but if there are absolute roadblocks to consolidation, this might work. (if your tape hardware supports partitions over iscsi, this can even be done with VMs; just make sure the iscsi NICs sit on a 10Gbit network else you'll be capped with 1Gbit speeds and choke your tape backups)

tom4784
Novice
Posts: 5
Liked: 1 time
Joined: Jun 21, 2022 8:06 am
Full Name: Tom
Contact:

Re: Tape from StoreOnce - nothing visible in repo

Post by tom4784 »

Thanks for your comments, it is much appreciated.

It's only site A and B currently - other than that you have described it perfectly.

Why we don't have 1 VBR server to manage both sites: One site has to be physical as the Tape drive won't connect directly to the VM's. That site may be powered down occasionally and we still need backups of the other site - so that's why we went with a server at each end.....if I had been advised by the consultants otherwise I wouldn't have gone this way and went with a solution like you describe and used proxies etc - thought I was going with the simplest solution :D

I think I'm going to have to make the case to do some rejigging.

Thanks for your input.

soncscy
Expert
Posts: 636
Liked: 305 times
Joined: Aug 04, 2019 2:57 pm
Full Name: Harvey Carel
Contact:

Re: Tape from StoreOnce - nothing visible in repo

Post by soncscy » 1 person likes this post

Hi Tom,

Sure, happy to advise.

Okay, your situation is more clear. Few questions on my side then:

1. Do both sites take a full power-off periodically or just one?
2. Is there any way to get the VBR server on a machine/host that just doesn't go down? (will discuss DR situation in a moment)
2a. Remember, only the tape _server_ needs to be physical. One of the VBR servers (the physical one) can be converted to "just a tape server" and then added to the main Veeam

I think this is still manageable; for DR purposes if the Veeam server dies, the configuration backup is simple enough. Normally I just either rsync the config backup directory to a mounted S3 bucket for clients; you can likely cheap-out and even use BackBlaze free tier. 10 GiB should be enough -- I don't have any particular love for BackBlaze, but I do have love for a truly "free" (as in beer) edition with no credit card and no functionality limitations. A recovery is as simple as get a new Veeam server up and running, then restore from config; on even a modest system, this adds about 10-30 minutes to your full DR scenario for good machines, and maybe an hour for slow ones.

Hopefully the rejigging is accepted, as I think you'll sleep a lot better once it's running like clockwork ;) The rejigging should just be recreating some jobs on the centralized server, adding the repositories, and then editing the recreated jobs and mapping to the newly imported backups. Bit of clicking, but thankfully it's nothing super spooky you need to do.

Post Reply

Who is online

Users browsing this forum: No registered users and 5 guests