NCM ConfigMgmtJob.exe is using too many resources. The NCM nightly inventory is scheduled for 11pm at night, and it was running this morning, burning 1/2 gig, and using most of a CPU.
We have NPM 9.5 SP5 and NCM 5.5.1 on the same box
There has been a performance issue with NCM 5.5.1 regarding the size of the policycacheresults table - and the resulting impact upon the inventory run times.
Here is the Thwack thread that discussed this issue
The release notes for 5.5.2 has a note regarding policycache changes - which I'm hoping addresses this issue. I have just installed 5.5.2 today, so will have an idea myself after the weekend just what has been changed. Just before the upgrade, this table was >15M rows in our implementation....
Dave.
I checked our policycacheresults and it has about 15k rows
ConfigMgmtJob.exe was running again this morning, nearly 12 hours since its start. I suspect it does this every day. I am not sure exactly how long this process takes to complete, or if it just burns CPU and memory 24x7
Our platform has under 500 devices
Hey Dave, I'm very curious to know how your inventory goes after the upgrade. I have been running on 5.5.2 since it came out. My inventory still takes about 12 hours. I have about 600 devices that are being inventoried.
I have also checked the policycacherults table. The table has 3,020,590 rows in it. Yes, you read that right, 3 million.
Not sure if this table still impacts the inventory scan or not with the new NCM version.
Well, things have changed a little. The inventory process no longer seems to drive a select from the policycacheresults table, and an index has now also been placed on this table.
My inventory however was still running some 13 hours after the scheduled time - but was the result of having all the inventory options selected. Previously we didn't collect the L2 Bridge Port information due to how long this took (the upgrade may have reset this?).
I have unselected this inventory option and will wait for another run tonight.
What hasn't changed is the number of rows in our policycacheresults table - still 15 million +.
There might be a weekly cleanup type job that sorts this out - but if you have been running 5.5.2 for a while, maybe this is something still be be addressed.
