Monday, February 18, 2013 3:09 PM
I am looking at applying SP1 to my production WSFC cluster/Always On AG.
I am struggling to find any information on best practice for doing this, does anyone have any links? I presume the thing to do is patch the secondary first, fail over to it, then patch the old primary server. I am using an AG Listener so this should work well for me. I assume that after patching the secondary server and failing over to it, the AG will go into suspended mode due to the version mismatch? Therefore I will have to apply SP1 to the old primary server a.s.a.p. so that data synchronisation can continue?
I'd love to hear from anyone who's done this already.
Monday, February 18, 2013 3:49 PM
Please check the below KB
SQL Server failover cluster rolling patch and service pack process
- Proposed As Answer by Noam BinyaminiMicrosoft Employee Monday, February 25, 2013 10:40 PM
Monday, February 18, 2013 3:54 PMThanks, but that article refers to a clustered instance, which mine is not, it is an Always On AG which is not the same thing.
Monday, February 18, 2013 4:27 PM
I don't know if there is a Whitepaper on that... The simple breakdown is: (Like you suggested...) Patch the secondary, Failover, patch the other node.
Key Thing is that the secondary should in General be higher in patch Level than the Primary. So you can patch the secondary any time, but once you do the Failover you should immediately patch the old Primary node.
- Marked As Answer by Maggie LuoModerator Tuesday, February 26, 2013 9:38 AM
Monday, February 18, 2013 4:32 PMThanks Lucifer!
Monday, February 18, 2013 11:07 PM
Below procedure can be followed for applying patches to SQL 2012 instances having always on
- Patch the secondary servers
- Make one secondary synchronous if there isn’t one
- Failover to that secondary
- Patch the previous primary’s server
- If you made a secondary synchronous, make it asynchronous again
Ensure that cluster quorum is maintained during this process (e.g. by not restarting the servers simultaneously).
You may be able to run one server on RTM and one on SP1, but it would be best if you could get them both on to SP1 one after the other.
Friday, February 22, 2013 7:11 PM
Just to rephrase your scenario you have Multiple Local Instances on Windows Clustered Server and AG is setup across the Instances
In this case, you will need to
1. Ensure that AG is residing on the Node which is NOT getting patched
2. Patch the Local Instance on Non-AG Node
3. Failover the AG to Patched Local Instance
4. Apply Patch to the Local Instance on Non-AG Node.
I completed this in my Lab setup without any issues