Server access is removed when running PAL docker

[Metuant]

When running PAL docker I lose the ability to "see" my Plex server via app.plex.tv and local connection. Stopping the container allows access to the server. Config.yaml (redacted token info) and docker-compose config are below.

plexautolanguages:
  # Update language for the entire show or only for the current season
  # Accepted values:
  #   - show (default)
  #   - season
  update_level: "show"

  # Update all episodes of the show/season or only the next ones
  # Accepted values:
  #   - all (default)
  #   - next
  update_strategy: "all"

  # Whether or not playing a file should trigger a language update, defaults to 'true'
  trigger_on_play: true

  # Whether or not navigating the Plex library should trigger a language update, defaults to 'false'
  # Set this to 'true' only if you want to perform changes whenever the default track of an episode is updated, even when the episode is not played.
  # Setting this parameter to 'true' can result in higher resource usage.
  trigger_on_activity: false

  # Plex configuration
  plex:
    # A valid Plex URL (required)
    url: "http://plex:32400"
    # A valid Plex Token (required)
    token: "XXXXXXX"

  scheduler:
    # Whether of not to enable the scheduler, defaults to 'true'
    # The scheduler will perform a deeper analysis of all recently played TV Shows
    enable: false
    # The time at which the scheduler start its task with the format 'HH:MM', defaults to '02:00'
    schedule_time: "07:00"

   notifications:
    # Whether of not to enable the notifications through Apprise, defaults to 'false'
    # A notification is sent whenever a language change is performed
    enable: false
    # An array of Apprise configurations, see Apprise docs for more information: https://github.com/caronc/apprise
    apprise_configs:
      - "XXXXXX"

Compose config

### PAL###
  PAL:
    image: remirigal/plex-auto-languages:latest
    container_name: PAL
    environment:
      - TZ=${TZ}
    volumes:
      - ${APPDATA_ROOT}/PAL/config.yaml:/config/config.yaml
    restart: unless-stopped

[RemiRigal]

Do you have any relevant output from the container ?

[Metuant]

Nothing significant from the PAL container. The Plex container, however, throws a ton of the following:

Scan :: files: []
Scan :: mediaList: []
Scan :: subdirs: ['_insert random media directory here_']

Before finally crashing and producing a .dmp for the event. Does this help? Is there additional information I can provide to help troubleshoot?

[RemiRigal]

I don't see how the access to Plex can be compromised by PlexAutoLanguages, it only creates a WebSocket connection with the instance and uses the exposed API. Maybe you are using an outdated version of Plex ?

[Metuant]

I'm also a little confused by the behavior. Running 1.26.0.5715 - things run fine for about 3 - 5 minutes. I'll notice via an app like PlexDash that the server is wiped out (similar on plex web, local access). Already running streams are not interrupted.

Stopping the PAL container and restarting Plex restore connectivity.

I'll continue to explore other variables and update on what I find. If you have any leads as to where to continue my own troubleshooting i'm open to them.

[RemiRigal]

Some leads that come to my mind in no particular order:

• Do you have Plex running behind a reverse proxy ?
• Is Plex running with Docker ? If so, are both containers in the same Docker network ?
• Your Plex URL uses a hostname, does the issue still happen if you put an IP instead ?

[Metuant]

Do you have Plex running behind a reverse proxy ?

• Yes (Traefik).

Is Plex running with Docker ? If so, are both containers in the same Docker network ?

• Yes. Yes. I also tried running on different network (same behavior was exhibited).

Your Plex URL uses a hostname, does the issue still happen if you put an IP instead ?

• Yes.

This issue persists whether utilizing docker or running outside of docker.

[Metuant]

@RemiRigal - have you had a moment to look further at this?

[RemiRigal]

Yes I have tried to reproduce the issue in several ways. I've created multiple VMs with various docker/docker-compose and Traefik versions but I'm unable to reproduce the issue.

The HTTP requests from PAL are probably not causing the issue so my guess is that your configuration is messed up by the WebSocket connection. Since WebSockets are supported by default by Traefik and bypassing Traefik (by using the Plex machine IP) produces the same behavior, I'm out of idea...

Can you try with the latest version of PAL with the debug mode enabled ? You just have to add debug: true in your configuration file, see the project's README for example.

Can you give me the following information as well:

• Your host OS and Kernel version
• Your docker/docker-compose version
• Your Traefik version
• Whether or not you are using secure connection in PAL (HTTP/HTTPS)
• The network configuration of the Plex container (host mode, bridge mode...)

[Metuant]

Running latest version of PAL with debug enabled shows no errors. but every minute or so there is a stopping of the cache proces and then it kicks off again: [INFO] Scanning all episodes from the Plex library, this action should only take a few seconds but can take several minutes for larger libraries [DEBUG] [Cache] Refreshing library cache [INFO] Received SIGINT or SIGTERM, stopping gracefully

Your host OS and Kernel version

• Ubuntu 18.04
• 5.11.11-051111-generic Your docker/docker-compose version
• Docker version 20.10.14, build a224086
• docker-compose version 1.27.4, build 40524192 Your Traefik version
• 2.6.6 Whether or not you are using secure connection in PAL (HTTP/HTTPS)
• I've tried both secure and non-secure connection The network configuration of the Plex container (host mode, bridge mode...)
• bridge mode

[Metuant]

@RemiRigal using a client-side token seemed to have resolved this particular issue.

EDIT: For others who have problems - DO NOT use the token from preferences.xml - use token generated https://github.com/Cloudbox/Cloudbox/blob/master/roles/scripts/files/plex_token.sh