Backup an Instant VM

[gsegree]

Can you perform a backup on a running Instant VM?

Thanks.

[foggy]

Yes, you can backup an Instantly Recovered VM just fine.

[gsegree]

I added it to an existing job and there is no record of it being backed up.

Scenario:
server1 failed and I created an instant vm called server1 then re-added it to the server 1 backup job.

When I try to do a restore there is no reference of the server 1 instant VM being backed-up it that job.
Should I have created a new job?

[Vitaliy S.]

Did you add this server to the job that previously was backing up its production copy? Do you see this VM being processed in the job session stats?

[foggy]

Gareth, we haven't specifically tested such scenario, but I do not see any obstacles preventing it from working. Check whether you're selecting the correct restore point (the one that contains newly backed up IR VM).

That said, IR is not designed to use VMs in this state in the long run, it is considered to be more of a spare tire and requires further migration to production datastore. Any reason why you do not want to do that prior to starting to back this VM up?

[nik_kit]

We have now the following Problem:
Big VM (couple TBs) was damaged (human factor) and cannot be used any more.
FLR takes too much time to complete (for so much Data), so it is not an Option.
We started IVR to be able to use this VM immediately and started Migration to production as well.
Now, with the actual restore Speed this will take approx. couple days. To avoid such human factor issues in these days I would backup this IVR VM as well.
I see the following Problems with it:
1. Regardless I would add IVR VM into existing Job or create a new one - initial backup have to be done, so couple TBs in Backup Rep should be free.
2. As Migration to production is done, we will need to backup restored VM as well. So an initial backup have to be done again with time and space needs.
Is there any possibility to optimize this task? Is the idea correct? Is there any Thing to make better?
Thanks
Nik

[foggy]

Nik, what backup mode does the original job use, forward or reversed incremental? With the latter one, you may not need additional space for the entire full if you add the VM to existing job, since similar blocks will be deduplicated within the full backup file. However time to read the entire VM data will be required, indeed.

[nik_kit]

Hi,
Forward incremental is used by the old backup Job. I will add the restored VM into old backup Job. Hope it will work as you describe, will save us space.
Thanks!

[foggy]

Additional space would not be required if reverse incremental was used.

[nik_kit]

Unfortunately initial backup was exactly as large as the origin Vm. I am not sure where Veeam looks for the same blocks, but it look like just by backup of a single VM. If you make a copy of a VM and backup it in the same or different task you will need the same space amount as by the first VM backup.
Probubly you will save some space if you dedup backup rep with windows dedup, but I think veeam does not combine blocks from different Vms regardless if those VM are in the same or different tasks.

[foggy]

You've said that you'r using forward incremental backup, it is expected in this case that the new increment has the size similar to the entire VM backup (and I've mentioned it earlier). Blocks can be deduplicated only in case of reversed incremental, since they will be placed into the same full backup file (the most recent restore point).

but I think veeam does not combine blocks from different Vms regardless if those VM are in the same or different tasks.

Not task, but backup file (restore point) matters in this case. Veeam B&R does deduplication between different VMs within a given backup file and this is actually what could allow to save space in your case in case of reversed incremental mode.