For unknown reasons, my UnRAID server will suddenly have all 4 CPUs pegged to 100%.  The result is the whole house loses internet access as I've got Pi-hole DNS, and Unifi Controller dockers running on this box.  It happens about once a week.  I first thought it was the Pi-hole docker so I  pinned it to a single CPU.  Then I thought it was the Plex docker, so I pinned that to 2 CPUs.  Neither action has solved the problem.  Running top doesn't seem to indicate what is maxing out the CPUs, no process is using over ~20%.  I can fix it without rebooting by restarting the plex docker via command line.  I've been having this problem for almost a year.   I've looked at the diagnostics but can't figure this out.  Could anyone please help?

 

 

scour-diagnostics-20210430-0910.zip

Occurred again today.  I noticed the cache drive was being read at a very high rate while CPUs were pegged to 100% again.

Attached new diags.

 

Any help appreciated.

scour-diagnostics-20210503-0959.zip

Looks more high load due to i/o wait, if you stop all cache activity (turn off docker service/VMs if needed to test) does it get better?

16 hours ago, JorgeB said:

Looks more high load due to i/o wait, if you stop all cache activity (turn off docker service/VMs if needed to test) does it get better?

@JorgeB thanks for the reply.  
Is there anyway to check besides disabling docker/VMs? As this doesn’t occur predictably, and sometimes not for many days, I’d be without many services unless I built another system to take over network duties.

thanks again. 
 

I meant does it get better if you stop everything after the problem starts?

Sorry, I misunderstood. 

If I restart only the plex docker, it goes back to normal operation.  I haven't tried disabling/re-enabling the docker service and VM service.

1 hour ago, DingHo said:

If I restart only the plex docker, it goes back to normal operation.

Sounds like a plex problem then, you can try the plex forum or the docker support thread.

Could it be the infamous Plex health check? AFAIK that issue is resolved in later versions of Unraid. May I ask why you are not upgrading? It seems like an easy thing to apply if you're having issues.