Unraid Version: 6.10.3
Plex Server version: 1.28.2.6106 in LSIO Docker

Entire server grinds to a halt in the middle of the night usually, and I've worked out it's Plex Media server, because if I stop the docker container everything recovers, and it seems to happen during Plex Maintenance, what I can't figure out is what exactly is causing it, the plex container is a fresh install too because the same problem was happening and I wiped the plex configs and completely reinstalled it, sadly doing that only fixed it for a short period and now it's happening again. I've attached Unraid Diagnostics.

Edit: Forgot to mention that I can usually see when the server is grinded to a halt by running 

watch -n 1 "(ps aux | awk '\$8 ~ /D/  { print \$0 }')"

And it'll be full of things trying to run, and as soon as I run `docker stop plex` and wait a few seconds for a stop that list completely clears.
 

home-server-diagnostics-20220819-0511.zip

Edited August 21, 20223 yr by deanosim
Add additional information and Version numbers.

Happened again and this time bugged out something using nvidia-smi and caused that to leave a ton of nvidia-smi proccess in ps aux in the uninterruptible sleep D state


home-server-diagnostics-20220822-0602.zip
 

Edit:
Just noticed theirs a page fault in the syslog relating to nvidia-smi, sooo that's probably not good. and I need to restart the server to get it to work properly again.

Edited August 21, 20223 yr by deanosim
add more information

I had a lot of Plex related weirdness, basically I built an Unraid server from scratch to replace my old faithful qNAP 853a.

 

The simplest fix was to give the Unraid Plex install a unique name, the error I had was due to a name conflict, try renaming "friendly name" of your Plex server to something unique, mine is "Socks-Plex". Once I did this all issues disappeared.

Interesting, I'll have to give that a try, currently it's just called PineappleFlix but I've been using that name for years and across reinstalls of plex including the recent one, I'll try adding a 1 or something to it and see if that helps.

Well it stopped for a little while, but happened again tonight and I stopped the plex docker container and everything went back to normal for awhile, but then it happened again like an hour later so it can't have been plex the second time.
I'm out of ideas so I've submitted a support request with Limetech with a bucketload of diagnostics files, Hopefully they can spot something I'm missing.