Maintenance Plans All Died…See the Magic
Last night on one of my production clusters, all of the maintenance plans died. On my drive in to work, I thought it was probably a drive space issue, but it was also strange that I did not receive an alert for that. Finally at the office looking at the Job Activity Monitor then view history, I get a lot of useless information that says simply “the step failed” during my reorganize index routine. This is not much help.
When looking at the Maintenance Plan view history, then I actually see a useful error: “Alter failed for Server <servername>.” When looking up this error ,I narrowed it down to two plausible solutions. The first being to check to make sure that “Automatically set I/O affinity mask for all processors” is checked. We don’t generally change that from the default. Next.
The second option was to make sure sp_configure “allow updates” was set to zero. I thought it was, especially since this is a 2005 box and the feature is supposed to be deprecated (books online confirmed that for me, interesting though it says the functionality is unavailable). Low and behold “allow updates” was set to one. A quick reconfigure and the maintenance plans are running again.
This is a temporary fix because the new server I am migrating these to at the end of the year will be running Ola Hallengren’s backup solution, like my other new servers. Bye bye maintenance plans!