Plex Transcoder stops whilst transcoding a stream

I've been running my plex container on the same hardware for years now, I have 8gb of ram installed and this has never been an issue, transcodes have always worked.

• I have recently added new drives and went from a 16TB array to a 36TB array.
• My memory usage never seems to go above 50%.
• When I look in htop after an OOM kill the process is still in the list.
• My transcode directory is mapped to my cache SSD
• I have added --memory=8G to extra parameters for the plex container

This is all I have in the syslog

Jan 10 06:21:09 Highrise kernel: [ 15581] 0 15581 1127 144 45056 0 0 prerunget.sh
Jan 10 06:21:09 Highrise kernel: [ 16323] 0 16323 51 1 28672 0 0 s6-supervise
Jan 10 06:21:09 Highrise kernel: [ 16324] 0 16324 51 5 32768 0 0 s6-supervise
Jan 10 06:21:09 Highrise kernel: [ 16325] 0 16325 51 1 28672 0 0 s6-supervise
Jan 10 06:21:09 Highrise kernel: [ 16327] 0 16327 51 1 28672 0 0 s6-supervise
Jan 10 06:21:09 Highrise kernel: [ 16329] 0 16329 396 13 40960 0 0 crond
Jan 10 06:21:09 Highrise kernel: [ 16331] 0 16331 6709 2580 94208 0 0 fail2ban-server
Jan 10 06:21:09 Highrise kernel: [ 16332] 0 16332 58669 2566 143360 0 0 php-fpm7
Jan 10 06:21:09 Highrise kernel: [ 16360] 99 16360 58674 1903 122880 0 0 php-fpm7
Jan 10 06:21:09 Highrise kernel: [ 16361] 99 16361 58674 1903 122880 0 0 php-fpm7
Jan 10 06:21:09 Highrise kernel: [ 17682] 0 17682 1256 277 49152 0 0 prerunget.sh
Jan 10 06:21:09 Highrise kernel: [ 17686] 0 17686 596 21 40960 0 0 sleep
Jan 10 06:21:09 Highrise kernel: [ 18065] 99 18065 137693 16758 307200 0 0 deluged
Jan 10 06:21:09 Highrise kernel: [ 18523] 99 18523 305818 503 192512 0 0 privoxy
Jan 10 06:21:09 Highrise kernel: [ 18533] 99 18533 20396 10891 192512 0 0 deluge-web
Jan 10 06:21:09 Highrise kernel: [ 16869] 99 16869 1209887 1197534 9723904 0 0 Plex Transcoder
Jan 10 06:21:09 Highrise kernel: [ 18826] 0 18826 596 22 40960 0 0 sleep
Jan 10 06:21:09 Highrise kernel: [ 24282] 0 24282 543 11 36864 0 0 run
Jan 10 06:21:09 Highrise kernel: [ 25139] 0 25139 1053 677 40960 0 0 cache_dirs
Jan 10 06:21:09 Highrise kernel: [ 25141] 0 25141 1053 677 40960 0 0 cache_dirs
Jan 10 06:21:09 Highrise kernel: [ 25143] 0 25143 1053 677 40960 0 0 cache_dirs
Jan 10 06:21:09 Highrise kernel: [ 25161] 0 25161 662 187 40960 0 0 timeout
Jan 10 06:21:09 Highrise kernel: [ 25162] 0 25162 1044 578 45056 0 0 find
Jan 10 06:21:09 Highrise kernel: [ 25176] 0 25176 662 187 40960 0 0 timeout
Jan 10 06:21:09 Highrise kernel: [ 25177] 0 25177 1042 596 40960 0 0 find
Jan 10 06:21:09 Highrise kernel: [ 25181] 0 25181 662 188 40960 0 0 timeout
Jan 10 06:21:09 Highrise kernel: [ 25182] 0 25182 1011 273 40960 0 0 find
Jan 10 06:21:09 Highrise kernel: [ 25267] 0 25267 940 741 45056 0 0 sh
Jan 10 06:21:09 Highrise kernel: [ 25268] 0 25268 940 63 45056 0 0 sh
Jan 10 06:21:09 Highrise kernel: oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),cpuset=/,mems_allowed=0,global_oom,task_memcg=/docker/4749f7790f7ed85cedcaa8f6f91a1f36a8267a138dbf248bf41425810255f8a3,task=Plex Transcoder,pid=16869,uid=99
Jan 10 06:21:09 Highrise kernel: Out of memory: Killed process 16869 (Plex Transcoder) total-vm:4839548kB, anon-rss:4790136kB, file-rss:0kB, shmem-rss:0kB, UID:99 pgtables:9496kB oom_score_adj:0
Jan 10 06:21:09 Highrise kernel: oom_reaper: reaped process 16869 (Plex Transcoder), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB
Jan 10 08:00:16 Highrise crond[1734]: exit status 1 from user root /usr/local/sbin/mover &> /dev/null
Jan 10 10:00:16 Highrise crond[1734]: exit status 1 from user root /usr/local/sbin/mover &> /dev/null
Jan 10 12:00:16 Highrise crond[1734]: exit status 1 from user root /usr/local/sbin/mover &> /dev/null
Jan 10 13:15:37 Highrise kernel: kmem.limit_in_bytes is deprecated and will be removed. Please report your usecase to [email protected] if you depend on this functionality.

It looks like the OOM reports total-vm is around 4GB instead of 8GB but I don't really know why that would be.

EDIT: So now I'm looking into this more, it might not be OOM. I am seeing in the syslog but the time doesn't match. It seems like the plex transcoder just stops transcoding after a few minutes and I have to stop and restart the stream for it to start a new transcode session. It sounded similar to this https://forums.unraid.net/topic/110169-plex-suddenly-having-problems-with-audio-transcoding-i-think/ but I am able to play the file for a while until it stops.

I've deleted the codecs folder in my plex data as suggested by that thread but I still have my transcodes just stop after a while. The process isn't killed, it just seems to hang.

Edited January 10, 20224 yr by evoactivity
Might not be OOM