November 4, 2009 at 11:58 pm
I'm wondering if anyone is having the same issues that I am with SQL 2008 clustering and using the rolling patching feature.
Currently, I have installed 3 clusters with the same result.
1. I install a slipstream of SQL 2008 SP1 with CU1 on the active node.
2. I add a node to the sql cluster with the same slipstream.
3. I then apply SP1 CU3 to the passive node (partially following this KB: http://support.microsoft.com/kb/958734).
4. I then manually failover the SQL cluster group to the patched node.
5. The SQL Agent repeatedly fails to start on the passive, patched node, then eventually returns back to its previous active node.
It's not until I patch the previous active node that the Agent will start successfully.
Each time, the messages in the event logs are:
SQLServerAgent could not be started (reason: Unable to connect to server 'sql-server-virtual-name-here'; SQLServerAgent cannot start).
SQLServerAgent service successfully stopped.
SQLServerAgent service successfully started.
SQLServerAgent could not be started (reason: Unable to connect to server 'sql-server-virtual-name-here'; SQLServerAgent cannot start).
SQLServerAgent service successfully stopped.
[sqagtres] CheckServiceAlive: Service is dead
I honestly cannot tell what I'm doing wrong as far as this is concerned. The workaround I have implemented is to change the sql agent to "do not restart", and that allows the SQL Engine to be up and running in a patched mode, so the agent becomes unavailable for the duration of the patch on the other node, but it's just bugging me as to why that won't load up the agent successfully.
Any thoughts/comments would be appreciated.
Thanks in advance!
Steve
December 14, 2009 at 10:47 pm
No one has experienced this issue?
December 21, 2009 at 7:09 pm
Guess I should open a case with MS?
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply