June 11, 2012 at 12:21 pm
We have a Two node "Active/Active" cluster. We are running Windows Server 2008 as well as SQL Server 2008 R2 SP1. Last Friday morning (1 am), our engineer applied the Windows updates to the second node of our cluster. I felt that this would perform an orderly migration of the three instances on that node to the other node... then do the restart. When I came into work and checked the system, the very large databases on the three instances that were on the node that rebooted were in "Recovery" mode. They all recovered after a period of time, but this is not what I expected to have happen. I could not find anything in the SQL Server Error Log that proved that a migration of resources did or did not occur, but it appears to me that a dirty shutdown occurred. Can anyone shed light onto this for me? Are there any other logs I should be checking for more information? Thank you in advance!
Charlie
June 12, 2012 at 9:50 am
I would take a look at the system error log and the SQL error logs to see if there was anything untowards. The SQL error log will also show you what node they were running on when started so you can go back and look to see if failover occurred.
You mentioned that it took a long time for the large databases to recover, have you checked your VLF count recently? Extremely high VLF numbers can lead to very long recovery periods for databases.
June 15, 2012 at 9:53 am
The updates will not perform a manual failover for you. You need to do that part yourself. Or if you've configured it to reboot automatically as part of the updates, the nodes will failover when it reboots. Either way, it's not a hard crash of SQL Server. SQL will shut down gracefully unless it hits some bug that causes it to crash or you force it to shutdown with no wait.
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply