SQL Always on Configuration questions and behavior

[backupquestions]

First off, I have an open service request with ticket number: 04146154

I have submitted all the requested logs and they are reviewing I believe. The engineer I worked with was great, but did not know the specifics on some of my questions. I am hoping to hear it on the ticket once they get more info but I wanted to post here to try to get some more clarification and visiblity on this, as it is a very important project we have right now.

We have a 2 node SQL Always on Cluster. ServerA and ServerB. ServerA is primary server, for all the time it can be, only if it had issues or was being patched would we fail over to B as primary.

Ok, so I've read the Veeam documentation regarding Always ON located here https://helpcenter.veeam.com/archive/ba ... pport.html It seems to say that Veeam can handle availability groups fine, and that agents are NOT REQUIRED and that Veeam will query each node and find which is PRIMARY and which is SECONDARY. It will then TRUNCATE the sql logs on PRIMARY and COPY ONLY on SECONDARY.

In my job, it seems to have a successful run (but it truncates both systems!) a few days, and then it will give me a warning every 2 or 3 runs instead of success (warning says it failed to truncate on secondary).

So, why is it truncating the logs of the secondary node every single time in the first place? It should be doing a "copy only".

My other confusion is exactly how this should be set up in the Veeam job. Do I configure the primary to do "backup logs every 15 min" if I want that functionality, and set the secondary to "copy only"? The way the support article reads, it sounds like it shouldn't matter what I select as veeam should intelligently determine which is primary and which is secondary itself.

Anyway, I need to both understand why it acts like this, but also how do I configure things so that the primary gets logs backed up every 15 min and the secondary is only a copy, or what best practice is supposed to be for always on cluster.

Last notes: Veeam 9.5 update 4b... SQL 2016 enterprise.... the availability groups are "basic" and not "distributed".

[backupquestions]

A quick add on to this. The support team has pointed out a query result in the logs that indicates the group is not "basic", but from what I can tell in SQL management studio we do not use distributed either. So maybe just standard group? Distributed groups seem to indicate requiring multiple windows server failover clusters, and that is not what we have. Just one cluster with 2 nodes. So it seems they are just standard groups but not "basic".

[foggy]

Hi Alabaster, just to confirm, are we talking about Windows Failover Cluster setup? In this case, you need to use Veeam Agent for Windows to back it up.

[mamosorre84]

Hi Foggy,

it's a suggestion or a best practice?!

I'm interested on it!

Thanks

Marco S.

[backupquestions]

Just an FYI, in my case since we did have a failover cluster set up, and in the logs veeam was complaining about the smb share that is used as the quorum in that cluster, it did require agents. I used the cluster backup job type and everything has been solid for quite a while now with 0 issues or errors. Very happy!

[foggy]

it's a suggestion or a best practice?!

Hi Marco, it's the proper way of action in case of the setup in question. Thanks!