Replication/DR repositories

[FrenchBlue]

Hello,

We have 2 DC/2 Veeam instances, each one replicating backups on a remote repository (Linux) in the other DC. For the DRP (but first, DRP test), we plan to import the replicated backups repository in the local Veeam instance and use backups there to restore VMs. Will this break the backup copies chain of the production site? I remember that, unless it changed recently, Linux repos can't be shared between 2 VBR instances, we tried to do this but had certificates issues, it failed. So once the DR test is finished, we'll probably have to re-import back the remote repo in the other VBR instance, I'm afraid it could break the backup copies chain at each DR test, which would be annoying.

Thanks.

[Mildur]

Hi FrenchBlue

Linux Repositories (deployed as BYO linux machines) can be shared between two backup server, but not supported. Both backup server have to run the exact same product version/build.
And we recommend to use different folder paths for each backup server in case you write backups with both backup server.

Nevertheless, I heard of customers doing it. It should work without „re-import on primary backup server“, but in case there are any issues our support may not be able to assist.

If you are looking for a clean and supported solution, then please have a look at our new Veeam Software Appliance with its HA feature. You can failover to the backup server in the second datacenter and start recovering immediately.

Or check out our Veeam Recovery Orchestrator.

Best,
Fabian

[FrenchBlue]

Thanks, we will make some tests. A couple of years ago it was already supposed to be working but not supported, but it didn't work at all, importing the repo in the 2nd VBR broke the connection of the 1st one due to certificates issues, maybe this has been solved in between.

[Mildur]

Hi FrenchBlue

importing the repo in the 2nd VBR broke the connection of the 1st one due to certificates issues, maybe this has been solved in between.

It should have worked (but not supported) since v11a and some patches.

Best,
Fabian

[FrenchBlue]

If you are looking for a clean and supported solution, then please have a look at our new Veeam Software Appliance with its HA feature. You can failover to the backup server in the second datacenter and start recovering immediately.

Hello, this supports re-ip of the appliance? Also, what's the status of migration from V12 or V13 Windows to the appliance?

Thanks.

[Mildur]

Hi Frenchblue,

What do you mean — re‑IP a node in an HA relationship?
At the moment, I expect that re‑IP’ing a node (especially the secondary node) would require disassembling and reassembling the cluster. However, I’ll verify this internally and ensure we add a note about it to the Help Center.

status of migration from V12 or V13 Windows to the appliance?

There is no ETA yet for when we may start mass conversions.
We have processed a few migrations in internal labs and selected customer environments, and based on the results, we are continuing to optimize the conversion workflow. The next patch for v13.0.1 will include additional stabilization improvements and fixes for the conversion process.

Best,
Fabian

[FrenchBlue]

Hello,

Sorry I wasn't clear enough, the DR datacenter has a different ip addressing range, with L3 routing between the 2, so the secondary node would need to be in a different subnet.

[Mildur]

Different subnets for both HA nodes is not yet supported. But it‘s on our roadmap for a future update.

Best,
Fabian

[FrenchBlue]

Ah OK, thanks, so this can't yet be a solution for us. I'm going to try the Linux repo sharing for now.

[FrenchBlue]

And I confirm the hardened Linux repo "sharing" now works fine (V13), no more issues like when tested back in V11. Maybe it was fixed in V12 already, we didn't re-test in between. Will make the DR easier, just rescan the backup copies repo and backup copies will be ready to restore.

[Mildur]

Thank you for testing.
Sharing Hardened Repositories should work (although it’s unsupported) starting with v11a and one of the subsequent cumulative patches — but I don’t recall the exact patch.

Best regards,
Fabian