Server Version#: Version 1.31.2.6810
Player Version#: Version 4.100.1
Plex Media Server.log (8.0 KB)
How to fix this? This cause a lot of CPU load for NOTHING and I can’t stop this crap. I cut the file cause this is spamming since days.
Can plex read write to the transcode directory you configured?
Yes, it can. Why I know? Cause I can play all content with no issues. And also there is plenty of space left. Other then that, I also set chmod 777 to the folder this was my first idea I had.
Just a thought, I’d run a SMART test on the hard drive(s) involved.
I did that, all drives tell me this:
=== START OF READ SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED
Note: If I restart Plex service, it run some hours without any issue after a while it starts spamming.
1 Like
I’m no Linux user, but this error with the EAE component usually means either
- the transcoder temp folder is mounted without
EXECution permission - the kernel is running out of
inotifymemory
OttoKerner
both should be given, I can see that docker could allocate 32GB Ram, and as I mentioned before I set 777 permission of transcodefolder, which include execute rights. Hm not sure what to change.
Look at the mounting point. AFAIR, data folders are often explicitly mounted with the NOEXEC parameter, which prevents execution, no matter the permission bits.
But I don’t really know how Docker differs from regular Linux, so I can’t tell where to look exactly.
Where did you map the temp and transcoder temporary directories PMS (in the container) uses ?
Did you map them to a network shared folder ?
Hey, no there are directly maped, no network share.
/mnt/disk1/Transcode/ → /transcode
same for content which is /mnt/diskx/ x = 1,2,3 etc.
The next question.
If you were to count the total number of directories which your media library includes (from top of each library section) , How many directories ?
‘find’ is the best way to count.
find /top/dir/section_1 /top/dir/section_2 (etc) -type d -print | wc -l
If the base OS kernel is < 5.0. your default max monitored directories (inclusive of the 2 that the EAE needs for transcoding) is 8192
What do you get for total ?
Hey, I check it and it is less than 1500 directories.
I use Linux 5.19.17-Unraid.
My Unraid box:
[chuck@glockner media.2004]$ find animated* tv* movies* uhd -type d -print | wc -l
2102
[chuck@glockner media.2005]$
Still learning Unraid here:
root@BabyNUC:~# uname -a
Linux BabyNUC 5.19.17-Unraid #2 SMP PREEMPT_DYNAMIC Wed Nov 2 11:54:15 PDT 2022 x86_64 Intel(R) Core(TM) i3-7100U CPU @ 2.40GHz GenuineIntel GNU/Linux
root@BabyNUC:~#
Would you please turn on DEBUG logging .
Get it to capture / recreate the problem you’re seeing.
Download the logs ZIP file from PMS
Attach here. I’ll look because I can’t recreate (yet) what you’re seeing.
How the transcoder sets up BEFORE the playback is important.
If there’s an error there – then you’ll see the errors like you’re getting.
My hope is to see where the problem is prior to start of playback so we can correct it.
I can do it tomorrow, cause I restartet the docker plex container for now and at beginig it has no issues. It should start being bad again I think towmorrow and then I can attach the logs.
Does this happen AFTER It’s had one night to do it’s maintenance ?
Not 100% sure if this is the case, but it could be. At least after a while, what ever a while is. Cause I don’t monitor that often, but I think around after a day or so.
Credit detection wasn’t deleting its temporary files and mucked up the whole system.
I’m on 1.31.3.6819-2ef591a4c. The fix is in this one.
Unfortunately there is no plexpass for docker from plex itself there is only Plex-Media-Server. I also thought that this is check by account and not by installation. There are third party plex pass, but I like to stay on the official one.
Here is my Log with Debug on as attachement. As I said it start doing errors after a while.
Question if this is also not offical from plex? I could swtich to other docker with plex pass. It may solve the issue.
€: I did a fresh intsall of plex pass, I hope it will solve the issue.
