Plex server with error on console

Plex Media Server
8

very sorry. been swamped.

Apr 07, 2025 20:36:52.000 [22846242962232] ERROR - [Req#44f1d/Transcode/f3a89269-f362-450f-89c9-4ae55c765d8d/20f205c9-b351-4167-9f2e-cfb0b6d4dc0f] [eac3_eae @ 0x146b4db50b80] EAE timeout! EAE not running, or wrong folder? Could not read '/tmp/pms-f333e80e-d54a-484b-9744-a9aa0beeff5d/EasyAudioEncoder/Convert to WAV (to 8ch or less)/f3a89269-f362-450f-89c9-4ae55c765d8d_72835-0-2235.wav'
Apr 07, 2025 20:36:52.001 [22846214212408] ERROR - [Req#44ff2/Transcode/f3a89269-f362-450f-89c9-4ae55c765d8d/20f205c9-b351-4167-9f2e-cfb0b6d4dc0f] [eac3_eae @ 0x146b4db50b80] error reading output: -5 (I/O error)
Apr 07, 2025 20:36:52.001 [22846474734392] ERROR - [Req#44ff3/Transcode/f3a89269-f362-450f-89c9-4ae55c765d8d/20f205c9-b351-4167-9f2e-cfb0b6d4dc0f] Error while decoding stream #0:1: I/O error

This error means that the file/directory doesn’t exist.

Could not read ‘/tmp/pms-f333e80e-d54a-484b-9744-a9aa0beeff5d/EasyAudioEncoder/Convert to WAV (to 8ch or less)/f3a89269-f362-450f-89c9-4ae55c765d8d_72835-0-2235.wav’

How it’s supposed to work.

  1. PMS creates a temp folder
  2. PMS sets a Notify event for that folder
  3. PMS launches FFMPEG for the video
  4. PMS launches EAE for the dolby audio.
  5. Error occurs if there is no audio file to load and merge with the video

Most common causes:

  1. Notify table is full (too many directories being watched – fills up the max_users_watches table in the Linux kernel
  2. tmp directory is on a network drive where notify will not work.

Given the most common cause is max_user_watches fills.
( 8192 directories on some, 65535 directories on other Linux kernels )
the solution is to:

  1. Count to total number of directories being used for ALL media
  2. See what the kernel currently allows ( fs.inotify.max_user_watches )
root@lizum:/glock/media/testfiles# sysctl -a | grep max_user_watches
fs.epoll.max_user_watches = 14541022
fs.inotify.max_user_watches = 1048576
root@lizum:/glock/media/testfiles#

  1. If needed, increase it.
    – newer Linux kernels dynamically set the table size based on memory available. (be careful when limiting a container’s memory – it can hurt you here )

  2. Round up number of directories in use to next integral multiple of 32768
    (it’s ok to increase another 32768 if you’re frequently adding and have plenty of memory)

  • On regular Linux kernel,

– ON HOST, edit file /etc/sysctl.conf
– at bottom of file, add:

fs.inotify.max_user_watches=<YOUR_NEW_VALUE>

For general Linux, here’s how to do it.