A cross sites cluster protection group rescan issue

[Ahmed Shakran]

When we have a cluster that has three nodes, two of them in main site and third node in DR, in addition there’s different repository for each site.
So we may have to create protection group for each site, and in main site protection group we exclude DR node and in DR protection group we exclude main site nodes.
Then we created two backup jobs one for main site nodes with main site repository as a target and main protection group in Protected Computers and other for DR node with DR repository as a target and DR protection group in Protected Computers.

Although above setup when we ran two backup jobs simultaneously both jobs took backup form the same node.

After investigation I found that both backup jobs took backup from the node which in the last rescanned protection group before backup jobs.
Example if DR site protection group rescan ran at 9:00pm then main site protection group rescan ran at 10:00pm and main and DR protection group backup jobs ran at 11:00pm, both backup jobs took backup from main site nodes. Moreover if we changed the schedule of DR protection group rescan job to be at 10:30pm both backup jobs took backup from DR site node.

So I think that at somewhere in Veeam database it marks the nodes of cluster depends on the last rescan operation regardless the name of protection group and its exclusion list. as we have the same cluster name in both protection groups with different exclusion list.

As a workaround to take backup from main and DR sites nodes, I rescheduled the backup job for main site after main site protection group rescan operation then DR backup job in different time after DR site protection group rescan operation, so I've different schedule for main site protection group rescan operation and DR site protection group rescan operation.

[rovshan.pashayev]

Good day,

Please clarify more details about the cluster environment. High level information regarding Operation Systems.

Thanks,
Rovshan.

[Ahmed Shakran]

Hi Rovshan,
Microsoft Windows Server Failover Cluster

[HannesK]

Hello,
that setup sounds like a SQL AlwaysON cluster, right? Let me check the recommended configuration for that scenario, because it's not covered in https://www.veeam.com/kb2463

Best regards,
Hannes

[HannesK]

Hello,
I got the answer, that the setup is not supported, but there should be a workaround possible

1. Create a protection group and add cluster to it. Add third node (that is in DR site) to exclusions of this protection group.
2. Create another protection group and add third node (that was excluded earlier) to it.
3. Create a failover cluster backup job, add the first protection job to backup scope and use main site repository.
4. Create a server backup job, add the second protection job to backup scope and use DR site repository.

If that does not work, please open a support case and post the case number here. Although it's an unsupported scenario, we would look at it.

Best regards,
Hannes

[Ahmed Shakran]

Hello Hannes,
Yes, I’ve already performed above 4 steps and it worked fine however on one condition the rescan time for each protection group, if the rescan of DR protection group was run so the DR backup job must be run next, it’s related to the last rescan job before start time of backup job regardless which protection group in this backup job.
I’ll open support case and post its number here.

[Ahmed Shakran]

Case # 06198126

[HannesK]

Hello,
thanks for the case number: one question... is that a SQL always ON cluster, or what type of cluster are you running?

Thanks,
Hannes

[Ahmed Shakran]

Yes, this is a SQL always ON cluster

[HannesK]

thanks. Just to let you know... we are still discussing options. I will come back when I have something.

[HannesK]

for agent-based backup, we need to change the product on our site for your scenario.

If it's about backing up the database itself, then the V12 SQL plugin might help https://helpcenter.veeam.com/docs/backu ... ml?ver=120