EAE timeout! EAE not running, or wrong folder? Could not read

Plex Media Server
1

Server Version#: 1.41.5.9522
Player Version#: 1.108.1.307-dd5b87aa

I migrated my instance from Debian to a container on TrueNAS Scale recently and am no longer able to analyze media. Have tried with and without hardware decoding enabled but errors are the same.

EAE is looking for the following ..EasyAudioEncoder/Convert to WAV (to 8ch or less)/8d6edd1a-b955-420d-8203-b0b5cbce6310_555-0-61.wav which does not exist even after purging the /tmp folder to let pms recreate it.

While the .wav file does not exist, the folder does contain 8d6edd1a-b955-420d-8203-b0b5cbce6310_555-0-61.ec3.

Something seems to be misaligned here. Any help appreciated
(attachment removed)

2

Biggest problem I see is that you’ve turned off DEBUG logging.

It’s not possible to track where it’s failing without the DEBUG server logs
Can’t tell if you’ve run out of inotify_watch slots or not

Please enable debug logging,
Restart PMS
Recreate the failure

Then attach the full server debug logs zip file

3

Debug enabled and reproduced:
(attachment removed)

4

@jdhicks

There MUST BE a problem outside the container on the host side.
( to set a kernel fs.inotify.user_watch trigger )

I don’t know TrueNAS. Last I did know, they were using Kubernetes. Is this still the case or have they gone back to real Docker ?

5

It has fully moved over to Docker as of version 24.10. Plex (PMS) is in their App library and is in fact one of their ‘stable’ Apps added by TrueNAS dev team vs ‘community’ Apps. This host was a fresh install of 24.10. I read another thread where you helped a user with a similar error message but I don’t us automatic scans.

A couple of things come to mind that could be contributing:

  1. At one point, I mistakenly switched the container to pull the Plex Pass image
  2. I migrated (rsync) my bare metal /var/lib/plexmediaserver/Library to a zfs dataset and then attached it to the container. The library migration was successful. My library is intact, I can add media, my EPG updates successfully, etc
  3. I changed the UID and GID to match the values on my previous bare metal install.

If nothing else pops out in the logs, the current container was only intended to be temporary anyway. Planning to rebuild another host this weekend on a fresh Scale 24.10 install and will create a ‘clean’ plex deployment to test on a small folder or two to see if I can replicate the issue with Analyze.

Part of me wonders if I should have done a zvol instead of a dataset and/or if I should strip the NFSv4 ACL off that I used during the migration and switch it to POSIX. These are things I’ll test but if it’s an issue with the filesystem, I’m not sure why everything else seems to be functional.

6

Just circling back here to close the loop. The clean install did not replicate the issue. I went back to the migrated container and switched the UID/GID back to the original value (568 for both) and the issue was resolved.

For anyone deploying PMS as an ‘application’ on TrueNAS Scale (my version was 24.10), changing these values seems to break the ability to Analyze media:

image
image1262×486 13.1 KB

7

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