there’s no secret formula to this. you’ll just have to roll up your sleeves and do it. apparently, this used to be in some old article Q234508 which has been removed for whatever reason since this works with sms 2.0, sms 2003, and configmgr (sccm) 2007.
identifying the instance
on your troubled server, use these steps to get perfmon to show you where the problem is occurring:
- fire up perfmon (obviously).
- add thread object with the following counters:
- % processor time
- id thread
could be quite a bit. i had close to 90.
if you think it’ll help, you can try the report view (ctrl+r) to isolate the thread causing the problems. otherwise, you can enable highlighting (ctrl+h) and with zen like patience, move through all of the smsexec threads until you see the thread that’s eating up % processor time. sometimes it’s easier watching it in histogram view, then double-clicking the line that’s bouncing around like an ice cream charged 4 year old.
once you find it, write down the instance number. (i don’t trust you to remember it). don’t close out perfmon yet.
isolating the thread
once you’ve identified the instance, these steps will locate the thread value.
- in perfmon, scroll through your counter list until you’re in the id thread list.
- find the corresponding instance number that you wrote down.
- highlight the instance and note the values in last, average, minimum, and maximum. they should all be identical. write this down, too.
- convert the value to hex using calc.
in my case, the value is 100980 which translates to 18A74. go to your sms\logs directory and use any assortment of find that you prefer using the hex value of the thread as your search criteria. why, here’s a good example, right HERE of how to do it. :)
your results should paint a nice picture for you.
you can get the full article HERE if you need more.