I've got a server that's hosting about 100 log shipping secondary databases. I frequently see messages like this in the sqlagent.out file.
"2012-03-23 08:00:11 - + [398] The job (LSRestore_dbname) has been queued because the maximum number of working threads (200) are already running. This job will be executed as soon as one of the working thread finishes execution."
I've got a couple of questions.
Where is the 200 coming from? My max worker threads setting is 512.
Which subsystem do the LSRestore jobs use? Isn't it CmdExec? My CmdExec subsystem's worker threads is set it 20 not 200. What I typically see is about 25 LS jobs executing and > 100 waiting for worker threads.
If I want to increase the max workers on the correct subsystem to avoid this error, which subsystem do I do it on?
Chuck