Backup Based Replication no longer works

[B.F.]

Site A ----------------------> Site B
VM Backup ----------------> VM Replication (using the Backup as a Source)

Both sites have Veeam B&R

After Site A Backup is complete, it runs a script that disables itself (prevent overlapping if the replication is still running when the next backup is kicked off) and then launches the replication job that is on Site B. Once the replication job that is on Site B completes, it then runs a script which re-enables the Backup job on Site A.

This setup has been working for years. I first talked about this type of setup back in 2016.

Some time, this year I'm guessing, the replication got switched to use the production VM on Site A instead of using the Backup on Site A. My hunch is that maybe it got flipped accidentally when the Veeam VM got rebuilt on Site B? Or maybe it got flipped after the 9.5.0.1922 upgrade? At any rate, I recently discovered that it wasn't using the Backup so was attempting to switch it the way we intended it. However, it keeps failing.

The error we keep getting is "Restore point is located in backup file that is not mapped to any backup job". I created a fresh test replication job based off a small VM backup on Site A and also get the same exact error. I thought I may have gotten somewhere when I set the initial seeding from the Site A's backup repository. It actually started and finished the initial replication. However, any new replications after a new backup will yield the same error.

The setup is quite simple. In the Virtual Machines settings of the replication, I added the VM in question and then select the backup repository on Site A from the Source... option.

Please advise.

Thanks

[foggy]

Just to clarify, do you have two Veeam B&R instances where the first (site A) is doing backup and the second one (site B) responsible for replication?

[B.F.]

In this scenario yes. Site A's Veeam server is backing up the VM. Site B's Veeam server is pulling data from Site A's backup for replication....or at least it used to.

[foggy]

Then you do need the backup job on the site B's instance that is mapped to the corresponding backup, along with periodic rescans.

[B.F.]

Well that is very interesting. Makes me now believe ours never really worked as intended since we didn't have such a dummy job before. Did as prescribed and now appears to be working! It needs to recalculate the digest which will take a long time unfortunately.

Thanks for the help!