Constant 50% CPU Overhead

[drogg]

Hi all! 

 

I am having an issue that I just can't seem to put my finger on. I constantly have 50% or more of my CPU being used and I have no idea what's causing it. I've attached a pic of terminal with htop running beside it as well as graphs from a direct play on Plex. I've tried turning off Docker, VMs, and it's still hovering at 50%. I only have 1 cache and 1 data drive (system running on an HP290). I know this Celeron isn't the highest performing CPU, but this seems a bit odd. FWIW, I was having the same issue on my older Xeon 1226v3 on my old machine.

 

Meanwhile, my memory is also being hammered by Log and I don't know what that means. 

 

Any ideas?

 

scout-diagnostics-20201212-0730.zip

Edited December 12, 2020 by drogg
Added diagnostics

[JorgeB]

Do you have syslog server running? Also try booting in safe mode.

[drogg]

7 minutes ago, JorgeB said:

Do you have syslog server running? Also try booting in safe mode.

Not sure what the syslog server is. How would I check?

What should I be looking out for in safe mode? 

[JorgeB]

20 hours ago, drogg said:

Not sure what the syslog server is. How would I check?

Settings -> Syslog Server

 

20 hours ago, drogg said:

What should I be looking out for in safe mode? 

If it helps or not.

[drogg]

Ok, just want to follow up. I did a complete reinstall - only carried over my license number. The ONLY other thing I'm running is a mergerfs mount with rclone and Sonarr/Radarr/DelugeVPN, but I'm not seeing how those would be hammering the CPU all day. 

 

Any ideas? Diagnostics attached. 

scout-diagnostics-20201221-2009.zip

[tjb_altf4]

If you stop the array is CPU still pegged?

I'm battling a similar issue, although it didn't appear in 6.8.3, might be some suggestions that help you there.

 

 

[drogg]

9 minutes ago, tjb_altf4 said:

If you stop the array is CPU still pegged?

I'm battling a similar issue, although it didn't appear in 6.8.3, might be some suggestions that help you there.

 

Stopping the array drops it to 1% (if that). 

[tjb_altf4]

4 minutes ago, drogg said:

Stopping the array drops it to 1% (if that). 

I noticed in your screenshot plex is doing a scan, try starting your array but not running Plex and see what usage is like.

The culprit might be Plex being over zealous.

[drogg]

I've actually been watching it for the last 10 minutes without a scan, and after the rclone mount has been added. Plex is at 2%, max right now (no scans), while streaming a video. I'm going to start every docker one at a time to see if maybe a docker has gone wild.

 

EDIT: Things were going like they should be until I opened NZBGet. As soon as it opened, the CPU and process shot up and won't subside. Interestingly, this was happening on my previous install...when I was using Sabnzbd. 

 

EDIT EDIT: According to advanced view on the Docker tab, NZBGet is only using 1%. So I'm thinking it's not actually NZBGet after all. 

 

EDIT EDIT EDIT: It just appears out of nowhere 11 minutes after the server is rebooted. I installed the Open Files plugin and it showed me a random video file was open, so killing that has fixed the error. I will keep monitoring. 

Edited December 22, 2020 by drogg

[drogg]

Just going to keep bumping this in hopes of help.

 

It seems like I have extremely high iowait. What causes this? I run one NVMe, one HDD. I keep everything besides seeding downloads on the cache, and I wouldn't expect to slow my server to a crawl whenever Sonarr or Lidarr copy files. 

 

Any help would be much appreciated! I'm at a complete loss here,

[daze]

I have a similar issue with cpu usage being pegged with all VMs stopped and docker stopped. No shares are being accessed, and nothing is being written. CPU load goes up to 51% on a 4 core i7. I've disabled the one file caching plug in. And also noticed, since the -rc1/rc2 releases, disks are never spun own.