Backup of VM replica during failover

[t_ahrens]

Hi all,

since I haven't found any information on this topic in the documention or previous threads, I'm asking this question to the community:

We are plannig to use replication of VMs for the first time. In case of failover from productive to replica VM, is it possible to use the same backup target (NAS) for the backup of the replica, or is it necessary to create a new full backup for the replica VM? Full backup would mean a lot more storage space consumption to us, which isn't available by now.

What are the recommendations for this case, in which the replica VM is in "permanent failover" state = productive for some time, let's say a week or so? What is / are the best practice(s) for continuous backup during this time?

Greetings and thanks in advance!
Thorger

[veremin]

You can select permanently failed over VMs as new backup job source, but full backup will be still performed. Basically, there is no way to avoid full backup cycle, as both VMs (source and replica ones) have different MorefIDs and, thus, are treated as independent entities. Thanks.

[t_ahrens]

Hi Vladimir,

that is exactly, what I was afraid of... I'm aware, that both VMs have different IDs, but isn't there a possibility to implement the feature in Veeam B&R to choose the option to map the backup of a new - restored or replicated - VM to an old backup of this VM, which contains exactly the same data, where only the ID of the new VM is different from the old one?

It's quite irritating, if it is necessary to run a new full backup cycle after some special Veeam B&R actions, even if the data of the VM hasn't changed at all. Especially, if you don't hove enough storage space at your backup target, and you are forced to delete the old backup of the VM. If something goes wrong, you won't have any proper backup of the VM left. And apart from that, you might get a problem with your retention policies.

Best regards
Thorger

[foggy]

Hi Thorger, what backup method do you use? In case of reverse incremental, even though the replica VM has different ID, it's data will be deduplicated in the full backup file if you map the new job to the existing backup chain, so will not consume double space.

[t_ahrens]

Hi Alexander,

meanwhile I did some testing with a small Linux VM, and what you describe is exactly what I have found out as well

For the test I have used the reverse incremental method, and different than expected, no new full backup file was created, the new backup was merged into the existing full backup file and a new incremental file was created.
Having a look at the backup in the VBR Console, it seems like it would even be possible to restore data of the VM or the complete VM at a restore point of the "old" / original VM before switching to the replica, is that right?

Regards
Thorger

[foggy]

Yes, until the moment it is deleted according to deleted VMs retention settings.

[t_ahrens]

Ok, thanks a lot!

[t_ahrens]

Just another question...

Vladimir wrote:

You can select permanently failed over VMs as new backup job source...

As I noticed, it is as well possible, to select a "normal" failed over VM as new backup job source. Are there any disadvantages if I do so, besides the fact that CBT can't be enabled, due to present snapshot(s)?

[foggy]

You can back up the failed over VM as well.

[erycucafreitas]

Morning,

I want to do a failover test and let vm's on the secondary site work for a week, is it possible to configure the backup for failover replicas while they are working on the secondary site?

Eridson

[foggy]

Yes, this is possible. Exactly what the OP was asking about.