I get this error more often than I want to see

Today I tried to install SQL Server 2005 SP2 Cumulative Update 13 on one of servers and I received an error "Error code: 11009 No passive nodes successfully patched". This is a Windows 2003 32-bit cluster and after six successful installations of CU 13, it had to fail on a production server. Luckily this production server is not yet in production (he he) and thankfully Microsoft included rollback inside the update. It did update workstation tools but not the engine.
Usually it fails when task scheduler or WMI is not running on the passive node. In today's case both services were running and I am yet to hear from Microsoft web support about the issue. What I did last time to fix was take the passive node out of the cluster and patch individually one by one each node. This was a special case because for some weird reason, the application vendor requires task scheduler to be a cluster-aware service and it must fail-over so I had to find another solution and that's what I had to do if I didn't want to touch the task scheduler. Well, I searched for the hell of it today and found this very helpful result. It is really quite genius, I must say. I can't believe I didn't think of it before. This article is as of April 20th, 2009, which is way after I had my last issue. The article describes that this issue first appeared after Cumulative Update 6 and plagued DBAs since then. It is fixed in SP3, more reason to upgrade! Anyway, my previous workaround is described there as method 2 but method 1 is pure DBA happiness! What I did today to quickly patch was pause the passive node, patch the active node, unpause, failover, reboot, patch the now active node with rebooted node now paused, reboot it and test! Voil-la you now have a working database engine version 9.0.3325!


Popular posts from this blog

Vista Vulnerability Report mentions Ubuntu 6.06 LTS

Running CockroachDB with Docker Compose and Minio, Part 2

Doing "print screen" on a Mac is a pain in the ass