Storage replication for Veeam Primary Repo, RTO 24h

[Eson]

Hi

I’m looking for advice on how you guys do if your primary backup repo becomes unavailable and you need to be able to resume incremental backups the same day and its to much data to run new full backups, RTO is 24h.

We are currently running:

A primary backup repository on Windows Server with ReFS
A secondary Linux immutable backup repo in another site that we run backup copy jobs to.

Veeam does not support re‑mapping an existing primary backup job to continue running incremental backups against Backup Copy chains.
I read on the forum that it has been a feature request since a few years back though I havent read any info that it has been implemented in v13.

This means that if the primary repository is lost or inaccessible, we cannot simply point the backup job to the immutable copy repository and keep going with incrementals.

How are you guys out there solving this with the similar requirements?

Variants I thought about.

Buy two QNAP or Synology NASes with enough performance to run block level snapshot replication. I think I need to use iSCSI to get enough performance but I assume since its block level replication should work anyway.

Use some existing server hardware and setup TrueNAS or some similar software based NAS that support block level replication

Anyone running NAS/SAN or something similar with storage replication for Veeam repo?

Any insights, best practices, or real-world solutions from your environments would be greatly appreciated.

[Brad.Barker]

Hello Kalle! I'm trying to better understand how your reimplementing the Windows Repository after it's failure. I attempted to replicate the behavior you described through several methods and every time the backup copy only performed an incremental copy (except under one scenario).

Test 1: Same exact repository server after blowing it away and adding back into Veeam and adding it back to the existing primary job. Result: Full backup on Primary Backup Job; Incremental Backup on Copy Backup Job
Test 2: Completely introducing a different repository platform (a hardened repository) and updating the existing primary job to use the new repository. Result: Full backup on the Primary Backup Job; Incremental backup on Copy Backup Job.

The only way I could get the Copy job to run a new full Backup was to create a brand new source job and add the objects to that. So basically, when you lose the primary repository, are you creating new primary backup jobs, rather than updating the existing primary backup jobs to use a different repository?

[Eson]

Thank you for all the testing

I think you might have misunderstood my scenario or i did a poor job explaining it

So we are working under the assumption that i have lost all my data on the primary repo, the server/repo isnt available. I need to resume running backups the same day for my customers and running new full backups isnt an option since it will take several days.

Veeam doesnt support remapping the backup job to my backup copy data since they are their own chain.

So I need an easy way to have my primary repo fully replicated in my secondary data center, restore the Veeam server instance and resume running my backups with a new incremental backup.

[Brad.Barker]

So, for clarification, your ask is the ability to disassociate your backup copies from their job and then be able to map primary backup jobs to those backup copies to continue the backup chain?

The reasoning is that running new full backups at the primary location will take too long in some circumstances.

Is this all correct?

[Eson]

That would be the dream scenario but I know that is not supported by Veeam as I mentioned Also mentioned by Hannes in this thread:

veeam-backup-replication-f2/feature-req ... 96585.html

So with that feature not being available I have no other option than to create a 1:1 data replica of my primary backup repo in my second data center with some other underlying technology. So Im curious how others do this type of storage replication with good results.

[Brad.Barker]

There is an existing feature request to support mapping of backup jobs to backup copy chains. That's why I wanted to make sure I understood your ask, as I can get you added as a +1 for that feature.

[Eson]

Ah, now I understand why you asked. Sure you can go ahead and add me, I will need to solve this issue in the coming months though so Im not holding my breath that Veeam will add it before that