Replicating and Backing Up at Same Time

[Stone_Deaf]

Can you replicate a virtual machine while veeam is doing a backup at the same time? or vice/versa. Cheers.

[Vitaliy S.]

Hello Nick,

You cannot process the same VM with multiple jobs (backup/replication) at the same time.

Thanks!

[Jude.Clermont]

Hello Nick,

You cannot process the same VM with multiple jobs (backup/replication) at the same time.

Thanks!

Vitaliy S,

Can you backup and replicate a VM on separate jobs at separate times?

[Vitaliy S.]

Yes, for sure

[jbsengineer]

[merged into existing discussion]

How do you guys handle backups and replications on the same VM's? Can you backup the replicated machine? Can you run a backup on a machine that is being replicated?

[jbsengineer]

I know you can't do it. I guess I'm asking how do you handle this situation?

[Gostev]

Schedule backup and replication job at separate times?

[jbsengineer]

Schedule backup and replication job at separate times?

How does that scale though? I'm curious how if there is a trick? Can you back up the replicated side, etc?

[Gostev]

You can backup a replica, but you are going to have to disable the use of CBT. Please search around this forum, all of these questions have already been discussed before.

[tsightler]

It's annoying to have both a replica job and a backup job as they are completely unaware of each other. We have systems that replicate every hour, and are backed up once a day. About once a week one of the replicas kick off while the backup is being performed (or sometimes vice-versa) and I get a lot of "failed" emails. It self corrects, but is truly an annoying. You can script up stuff to stop the replication during a backup window but because you can't control the exact order of backups and times it's very difficult to get it all working with random failures here and there due to conflicts between the jobs.

I wish replication and backup jobs had at least some knowledge of each other and would just wait and/or silently retry without all of the errors/warnings. Even better would be a policy based backup/replication schedule that was completely aware of machine's backup queue, but that's probably well beyond where they want to go with the product.

[Gostev]

In theory, this limitation can potentially be removed so that multiple jobs can process the same VM at the same time without causing any conflicts. I have actually discussed this with issue with devs a few months ago. Thing is, a few product pieces still do not have the best implementation for historical design decisions going all the way from v1. Most of these design quirks have been addressed already by now, and it is only a question of when we get to re-writing this particular piece of code.

[jbsengineer]

You can backup a replica, but you are going to have to disable the use of CBT. Please search around this forum, all of these questions have already been discussed before.

Understood. However, the primary reason we went to VEEAM was to utilize CBT.

We want to scale the implementation up to 1000 VM's. Just trying to weed through the glossies....

[SteveNeruda]

I'm struggling with this also. Rather than running what almost amounts to two backups a night I would love to be able to have the backup file replicate automatically to our DR site once the backup completes along with it's metadata so I would have restore point. I was really surprised that Veeam didn't work this way. I really don't like having to snapshot and unsnapshot the VM twice since there is a small pause while it STUN/UNSTUNs.

Non-CBT backups are a non-starter as not only are they slower but also are more work for the SAN/NAS infrastructure. One of the items I'm backing up is a 4TB repository of very small images that are stored as a virtual mode RDM. Our current backups system is taking 8-10 hours daily to stat() it's way through the millions of files to see what's changed. The VStorage API and Veeam are unbelievable. The initial Veeam backup took 31 hours across a 1GB network link (very sensitive data and didn't have the nerve to present the LUN to the Veeam server for direct SAN mode) The daily Veeam backups with CBT take 20 minutes. Simply incredible.

[Gostev]

Hi Steve, have you thought about using reversed-incremental backup mode, syncing the produced VBK each day to the remote site with RSYNC, and automatically restore it to DR host. I know this is not ideal, but this is best you can do to achieve single snapshot and yet additional replica in your remote site. Thanks.

[okrehan]

Just to make sure I have a correct understanding, it is currently not possible to have the same VM being backed up by a daily backup job with CBT enabled and the same time (but in antoher job) replicated several times a day to another datastore, also using CBT?

[Gostev]

Just to make sure I have a correct understanding, it is currently not possible to have the same VM being backed up by a daily backup job with CBT enabled and the same time (but in antoher job) replicated several times a day to another datastore, also using CBT?

This is possible (and even covered in sticky FAQ). Just make sure job do not overlap. Even if they do though, no big deal - second one will simply fail out.