Exchange failing in SureBackup after Veeam 7 patch #3

[zoltank]

Case ID#00579066

Our basic SureBackup job's Application Group includes one domains controller (DC1) which houses the AD roles, and the Exchange hub/transport server (ExchangeHub). We were running Veeam v.7.0.0.690 and it worked fine.

However, when we patched to v.7.0.0.839 yesterday, ExchangeHub immediately started failing in SureBackup with the Exchange Transport service being unable to start. Upon investigation ExchangeHub's preferred domain controller is DC2 and not the domain controller in its Application Group (DC1). If I add the other domain controllers to the Application Group then ExchangeHub starts correctly. Using another Veeam server still running 7.0.0.690 I created a new backup of DC1 and ExchangeHub and it started up just fine in SureBackup, so that means the issue is with the new Veeam version and not the VM itself.

Any ideas?

Worst comes to worst, how can I uninstall v.839 and go back to v.690?

[zoltank]

It appears that ExchangeHub will only work if our sole Windows 2008 R2 domain controller is in the Application Group with it. If it's one of the Windows 2003 domain controllers, including DC2, then it won't work. Could the patch have somehow changed how SureBackup deals with Win2003 domain controllers?

[Gostev]

There definitely were changes around Domain Controllers handling during backup, which may have potentially impacted Windows 2003 DC recovery within SureBackup. I have not seen any other reports about this issue, but nevertheless, it is quite possible that you are facing it because of the patch.

[zoltank]

What are the chances of it impacting Win2003 DC recovery outside of SureBackup?

[Gostev]

This fix was actually all about making full DC recovery to work correctly in certain scenario, without breaking DC replication, so I doubt normal recoveries are affected. But SureBackup recovery has a special step that applies additional behavior modifier to DC, and Windows 2003 may have been overlooked specifically in SureBackup testing. Just guessing though.

[foggy]

According to QC, this one is a known issue with Windows 2003 DC being recovered in non-authoritative mode in SureBackup. Unfortunately, this was not fixed in patch 4 and should be addressed in the next version.

Normal recoveries are not affected.

[zoltank]

I found if I run SureBackup using a restore point prior to our update to v.839 then the Win2003 domain controllers work just fine. I only get this problem in SureBackup if I use a restore from after the update to v.839.

Based on that, it doesn't really make sense to me that this is just a SureBackup issue because then a Win2003 domain controller should fail in SureBackup regardless of when it was backed up.

By the way, updating to v.871 (patch 4) did not resolve it.

[foggy]

Based on that, it doesn't really make sense to me that this is just a SureBackup issue because then a Win2003 domain controller should fail in SureBackup regardless of when it was backed up.

I will confirm that with QA on Monday, but I'm pretty sure this is explained by the fact that the way DC is restored depends on the way how it was backed up. If application-aware processing logic was changed in patch 3, it affected restores from backups created after patch installation.

By the way, updating to v.871 (patch 4) did not resolve it.

Yes, this is expected, as I've pointed out earlier.

[zoltank]

Is this issue still on track to be fixed?

[foggy]

Yes, planned to be fixed in the next version.

[zoltank]

Sooooo...

...when is the next version coming out?

[Vitaliy S.]

This fall > http://go.veeam.com/availability-suite-v8

[chri5]

I have tested this in V8 hoping this bug would be fixed as promised but exchange services are still failing with one 2003 DC in the application group.

Can anyone confirm whether this was addressed in V8?

Thanks

[zoltank]

Since Win 2003 is going EOL next year, I'd recommend getting a couple 2008 DCs installed on your network. Once you get a 2008 DC installed you can add it to your application group and problem solved.

[foggy]

Chris, it was finally decided not to include the fix into v8 release due to more testing/research required.