Understandig BKP job who follows a REPLICA job

[max86]

Good morning all,
this is my simple infrastructure, where you can see 1 main HyperV host with few VMs. There are 2 BKP jobs on 2 different NAS and 1 replica job on the second HyperV host.
2 days ago the first host had a crash, so I powered on the replicas on the second host, with only a bunch of minutes of interruption. All my production is still working on these vms, untill the main HyperV host will be fixed and repaired, and it will take at least one week.

My question is:
how can I "say" to the 2 BKP jobs that they now have to "follow" the replicated vms and keep backupping as normal? is it possibile?
Or do I have to create a new bkp job and chain for replicas? This will happen again when I will replicate again all the vms on the main host.
In my mind I think that Veeam should follow the vms from position 1 to position 2, and keep following them when I move them again to the normal position, keeping all the BKP jobs as nothing happened. But probably I'm thinking wrong.

Thanks for support and suggestions.



[Moderator: Edited to fix image inlining]

[Andreas Neufert]

Hi Massimo,

Veeam uses the VMware moref ID field to detect which VM is what. There is no way for us to migrate this ID to another vcenter/ESXi, as there is no API for this. As well there is no unique VM ID that can move with the VM.

So the second (replica target) VM is a new VM for the backup job. As the primary system is down, there would be no reliable way for us to compare the stages and failover the backup processing to a specific (restored) point in time. As well the replica is NOT a 1:1 mirror in case of blocks on the datastore disk as VMware writes the data differently within any VM.

When we would failover the backup job, this means that we would start as well from a new full. Doing this automatically for the customers would potentially have negative consequences because of the additional data amount needed. For this reason a manual decision is needed from the admin. Based on the name you can go to the jobs and add the VMs again. As well this could be automated with PowerShell if you are worried about the amount of VMs.