CPU Usage Issue

[rharvey]

Hello everyone, have not been here in a long time as my last Unraid build has been running very well but I do have an odd issue I wonder if anyone has seen or has ideas what may be causing this.  My server is based on a Supermicro X10SL-F motherboard with a Xeon E3-1231 running at 3.4GHz, 32 GB ECC memory,  8 total spinning drives all 4 or 8 GB and Iron Wolf NAS drives.  I have a single Windows 10 VM using the last 2 sets of cores (2/6 and 3/7) the other cores are fair game for everything else.  Also running a Plex Server, OpenVPN and DelugeVPN all in docker containers.  Under normal conditions the first 2 core pairs (0/4 and 1/5) see little CPU usage as shown in the load stat's on the Dashboard page.  Also under normal conditions the cores dedicated to the Windows VM are rather active running in the 30-40 % area but that's because the Windows VM is supporting a Blue Iris camera system with 7 - 1080p cameras, I know this is not a commonly suggested implementation but it's been rock solid for me and I run the VM itself from an SSD.

 

Now the issue - Every once in a while however the CPU usage shown in the load stat's for cores 0/4 and 1/5 are very erratic and one of the 4 will be pegged at 100%, but then quickly the 100% will jump to one of the other cores of these 4 and that bouncing around continues for ever, the cores that are not at 100% are running far higher than normal but not 100%, more like 50 to 60 %.  The 4 core set running the Windows VM remains normal.  I can stop all docker apps but still the first 4 cores are going crazy (or at least the load stat's think so).  If I stop the array and re-start it everything falls back to normal and will stay that way for several days.  I thought this was tied to my scheduled weekly parity check but I have turned that schedule off and it still happens.  Everything seems to be running fine during these crazy CPU usage times but it concerns me, any thoughts.  Running Unraid 6.4.1 by the way....

Edited March 12, 2020 by rharvey