Plex linusserver docker - 100% CPU, 1 core loop after starting a 'The Boys' episode. On Unraid

Plex Media Server
1

Server Version#: 1.20.2.3402
Player Version#: 4.34.4

As of a few days ago I’ve been having trouble with my Linuxserver plex container. Whenever I start a episode from a show called ‘The Boys’ 1 of my CPU cores randomly goes to 100%, stays there for a 1-2seconds then goes back down to 1%. Afterwards another single core goes up to 100% and then goes down. This keeps repeating itself infinitely until I reboot the container.

It only happens when I watch any episode from The Boys, any other show seems to work perfectly and when Plex has to transcode it does. The Boys is in h264 codec I believe so correct me if I’m wrong its just a direct stream?

I have no idea why this is happening, one week ago I watched season 1 & 2 without any problems. But now I get this bug and whenever someone else starts one of the episodes I have to restart the plex server or else non of the shows work when the 100% CPU core rotating bug is happening.

I saw older post that have this problem, but most of they point out Upnp is the source of the problem. I never have Upnp enabled in my router, so I’m not sure how to fix this problem other then delete the show from my library.

For some reason when I watch ALL logs in the console I see the error in the image below, which are not in the debug log at the bottom. Does anyone have any idea what is going on and how I can fix it?

image
image1874×238 239 KB

Here are my debug logs:
plex_debug_logs.txt (68.7 KB)

This happens to my cpu cores, it goes to 100% and switches. Sometimes starting a different show makes to seem to make it go away. But starting The Boys again makes it bug.

20201011_172220
20201011_172220841×594 27.3 KB

PS: I also like to point out for some reason playing ‘The Boys’ serie seems to work om some android devices running the Plex app and a browser now and then. There isn’t really consistency there, but my friend can watch it at his home from any device, yet I can only watch on the Plex app from my tablet, but not my phone or PC.

2

Unraid with NVIDIA plugin?

3

No Nvidia plugin, currently a CPU only system.

These are my settings:

image
image1064×462 30.5 KB

6

These usually mean that you’ve run out of inotify watches. You need to increase the limit in your system configuration or reduce the number of watches in use (such as disabling the Scan my library automatically pref).

7

How is that even possible to have to many watches? It’s a fresh linuxserver container with maybe 10 shows and 200gigs of movies. I lost like 5 TB of shows when I swapped to unraid, and never had to many iwatches before :S. I manually refresh libraries too.

When I run “find / /movies /tv -type d -print | wc -l” in both docker and the host i have about 145k directories. The iwatch limit is 545k.

8

Did you get it fixed?

Where is your transcode folder /mnt/user/media/transcoding located? Unraid Cache or Array? Array, possibly Parity protected, would be a very bad descision.

9

I fixed the problem by reinstalling plex all together. Right now it is qorking as intended. And my transcoding folder is located in my media share, wich is on a disk in my array. How come is this a bad idea? Does the parity make it slow? If i change the location to my cache disk will that prevent the problem in the future?

10

Array with single or dual parity is approx. 3 times slower than a native disk. Many user, me included, map /tmp to the transcoding folder - enough RAM assumed. Transcoding should take place on a fast disk. There’s no need to use parity secured operation for that task.

11

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.