I upgraded to Orion V9 and APM V2 tonight and I have found that the service, SWJobEngineWorker.exe and SWJobSchedulerSvc.exe are constantly rampaging and spiking out the CPU on this machine. The memory for these processes seems to be rediculous as well. Please look at the graph below, there is clearly something wrong and the graph is slowly getting worse. Could this be a memory leak?
Please help!
Your graph isn't visible. Not sure why.
We want to look into this. We need more detail. How many monitors are you running on how many servers?
Can you open a Support ticket so that we can gather more detail?
Case # 55989 Created: APM V2 Processes Rampage
I too am seeing this issue - since going to APM2 there are now 3 SWJobEngineWorker.exe processes running at all times now - they are all using between 160-200Mbytes of real memory and a similar amount of virtual. CPU utilisation runs anything up to 20-30% for any one process at times (and other times as low as 0%). Killing one just brings it back up again. I also see the SWJobSchedulerSvc.exe running around 15% CPU a lot of the time... So far my server is holding up - but it's WAY busier than it used to be - maxing out the CPU from time to time...
SMcdonald, I wish I had an answer for what was causing this but I don't. I ended up doing a complete uninstall and reinstall on the server and everything is working great now....
smcdonald:I too am seeing this issue - since going to APM2 there are now 3 SWJobEngineWorker.exe processes running at all times now - they are all using between 160-200Mbytes of real memory and a similar amount of virtual. CPU utilisation runs anything up to 20-30% for any one process at times (and other times as low as 0%). Killing one just brings it back up again. I also see the SWJobSchedulerSvc.exe running around 15% CPU a lot of the time... So far my server is holding up - but it's WAY busier than it used to be - maxing out the CPU from time to time...
How many component monitors are you using? Did you change the number after you upgraded?
We are also seeing a similar issue after going to APM2 and Orion9sp2. It is now starting to cause timouts on our HTTPS monitors if they poll while the server is at 100% CPU and one or all of the SWJobEngineWorker.exe's are running. We haven't change the number of component monitors since the upgrade but are certianly hoping to add a lot more monitors in the near future if we can work through the performance issues.
- David
Please open a support ticket. If it's the same issue that a couple of other users have seen, then we need you on a list to do a trial of the fix.
After APM 2.0 SP1 upgrade, CPU went to 100% and stayed there. 50% of that Netflow. We addedd two processors to the VM it was on for a total of 4, and CPU dropped to 7%. Go figure.