Empty Configugration although Config exists

[isntfunny]

Describe the bug
Everytime i navigate to the config page /settings/config every field is empty. Checking the config file on the file system manually shows that there is an config.

Expected behavior
Load the configuration from the config file

Screenshots

Server:

• Docker Compose:

version: "3.8"
services:
 web:
   restart: always
   image: mrbrax/twitchautomator:master
   ports:
     - 8082:8080
   volumes:
     - data:/usr/local/share/twitchautomator/data
     - /mnt/Streams:/usr/local/share/twitchautomator/data/storage/
   environment:
     - NODE_ENV=production
     - TCD_ENABLE_FILES_API=1
     - TCD_EXPOSE_LOGS_TO_PUBLIC=1
     - TCD_MIGRATE_OLD_VOD_JSON=0

[BigFlubba]

Make sure that you copied the Docker Compose file as instructed.

version: "3.8"
services:
  web:
    build: .
    image: mrbrax/twitchautomator:master
    ports:
      - "8082:8080"
    volumes:
      - ./data:/usr/local/share/twitchautomator/data
    environment:
      - NODE_ENV=production
      - TCD_ENABLE_FILES_API=0
      - TCD_EXPOSE_LOGS_TO_PUBLIC=0
      - TCD_MIGRATE_OLD_VOD_JSON=0

After it builds and starts up your ENV's should look something like this.

TCD_MIGRATE_OLD_VOD_JSON=0
NODE_ENV=production
TCD_ENABLE_FILES_API=1
TCD_EXPOSE_LOGS_TO_PUBLIC=1
PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/home/node/.local/bin
NODE_VERSION=20.18.1
YARN_VERSION=1.22.22
IS_DEV=
VITE_IS_DEV=
BUILD_DATE=
VITE_BUILD_DATE=
TCD_TWITCHDOWNLOADER_PATH=/usr/local/bin/TwitchDownloaderCLI
HOME=/home/node
TCD_BIN_DIR=/usr/local/bin
TCD_FFMPEG_PATH=/usr/bin/ffmpeg
TCD_BIN_PATH_PYTHON=/usr/bin/python
TCD_BIN_PATH_PYTHON3=/usr/bin/python3
TCD_MEDIAINFO_PATH=/usr/bin/mediainfo
TCD_NODE_PATH=/usr/local/bin/node
TCD_DOCKER=1
TCD_WEBSOCKET_ENABLED=1
TCD_SERVER_PORT=8080
TCD_PYTHON_ENABLE_PIPENV=1

Also, here are my volume mappings.
Container: /usr/local/share/twitchautomator/data
Host: /home/administrator/livestreamdvr

If you are trying to save recordings to a NAS, I haven't had any luck with it. If you are able to get it to work, please let me know how, so I can add it to my instance.

[isntfunny]

I've been running my LivestreamDVR for years using my docker compose. I am not sure which difference in your compose would fix my issue.

[DerBunteBall]

I can't reproduce this with an Docker Instance behind a Reverse Proxy and Authentication enabled.

I think these things could be the reason (based on the docker-compose I guess you don't use a reverse proxy - pulling mode):

1. Authentication. It might be that these info isn't loaded if no password authentication is set.
2. Something is wrong with fetching info on non 80/443 port, non TLS or something like that. For this check the browser console to check whether there is strange stuff.

[isntfunny]

Hey there!

I am actually using traefik reverse proxy, I just stripped the labels out because of privacy concern.

The response of the API route /settings looks fine but inside the response, the config json field seems to be mostly empty. Just some bare minimum configurations.

Checking the logs I also see the error that some binarry paths would not be set... But I have them right in my configuration, even when catting them inside the container

2025-06-12T17:52:22.742Z | config.loadConfig <info> Migrated setting 'locale.date-format' to 'date_format'.
2025-06-12T17:52:22.746Z | config.saveConfig <success> Saved config from missing default values
2025-06-12T17:52:22.747Z | config.cfg <warn> Setting 'bin_path_python' does not exist.
2025-06-12T17:52:22.747Z | config.cfg <warn> Setting 'enable_files_api' does not exist.
2025-06-12T17:52:22.748Z | config.cfg <warn> Setting 'bin_path_python3' does not exist.
2025-06-12T17:52:22.748Z | config.cfg <warn> Setting 'expose_logs_to_public' does not exist.
2025-06-12T17:52:22.748Z | config.cfg <warn> Setting 'python_enable_pipenv' does not exist.
2025-06-12T17:52:22.749Z | config.cfg <warn> Setting 'docker' does not exist.
2025-06-12T17:52:22.749Z | config.cfg <warn> Setting 'migrate_old_vod_json' does not exist.

[DerBunteBall]

Hm.....

I use the Image from Docker Hub. Also I don't pulled for a longer period of time. The last thing I could imagine is that if you manually build or the Docker Hub Image Rebuild did it's job that some dpendency pulled in this situation did a breaking change.

I simply can't reproduce this.

Image ID: sha256:b1be775bbcf75a4a42ca533fe5a0aeaa8e526ef5d98f4e9c54d06b7bd9742bfb

As long as your config.json in ./data/config/config.json (typically relative to your docker-compose project dir) has the necessary JSON fields they should be in the API response.

To give a orientation for the length of a config file. Line count with wc comes out at something around 110 lines.

[isntfunny]

How does your config folder look like? Do you also have a lot of .bak files?

Here's my wc (it's also 110), pwd, ls... Everything looks good to me.

[DerBunteBall]

Yes I also have these .bak files. I guess that's because it simply creates them with a non existing random number at every change of the config or regularily by time.

When I open my Settings I have values in the input fields and I have a long API response. So the API response from the path /api/v0/settings gives me some settings. I checked this in Firefox not with RapidAPI/Postman/or similiar actually.

[isntfunny]

Is there any chance that my config file got outdated and some values are mismatching or non-existent now and it's therefore ignoring my config file?

I guess I'll have to debug some code tomorrow to figure more out.

If you could share your full config so that I can compare it, that would be amazing! Of course removed with any domains or authentications.

These are all my set keys in my config:

config.json | jq 'keys'
[
  "api_client_id",
  "api_secret",
  "app_url",
  "basic.language",
  "bin_dir",
  "bin_path.python",
  "capture.autosplit-enabled",
  "capture.fallbackcapture",
  "capture.hls_segment_queue_threshold",
  "capture.killendedstream",
  "capture.loglevel",
  "capture.retry_on_error",
  "capture.twitch-access-token-param",
  "capture.twitch-api-header",
  "capture.twitch-ttv-lol-plugin",
  "capture.use_cache",
  "capture.viewercount",
  "channel_folders",
  "chatburn.default.auto_chat_height",
  "chatburn.default.chat_font",
  "chatburn.default.chat_font_size",
  "chatburn.default.chat_height",
  "chatburn.default.chat_width",
  "chatburn.default.crf",
  "chatburn.default.horizontal",
  "chatburn.default.preset",
  "chatburn.default.vertical",
  "chatdump_notext",
  "checkmute_method",
  "contact_sheet.enable",
  "contact_sheet.grid",
  "contact_sheet.width",
  "create_kodi_nfo",
  "create_video_chapters",
  "debug",
  "debug.catch_global_exceptions",
  "debug.catch_global_rejections",
  "delete_only_one_vod",
  "disable_ads",
  "discord_webhook",
  "download_retries",
  "dump_payloads",
  "eventsub_secret",
  "exporter.auto.enabled",
  "exporter.default.directory",
  "exporter.default.exporter",
  "exporter.default.privacy",
  "exporter.default.title_template",
  "ffmpeg_path",
  "file_chmod",
  "file_chown_gid",
  "file_chown_uid",
  "file_permissions",
  "filename_clip",
  "filename_vod",
  "filename_vod_folder",
  "guest_mode",
  "hls_timeout",
  "keep_commented_vods",
  "keep_deleted_vods",
  "keep_favourite_vods",
  "keep_muted_vods",
  "locale.date-format",
  "locale.time-format",
  "localvideos.enabled",
  "mediainfo_path",
  "min_chapter_duration",
  "no_vod_convert",
  "node_path",
  "playlist_dump",
  "python.enable_pipenv",
  "python.virtualenv_path",
  "reencoder.crf",
  "reencoder.enabled",
  "reencoder.hwaccel",
  "reencoder.preset",
  "reencoder.tune",
  "reencoder.video_codec",
  "schedule.export_vods",
  "schedule_deleted_vods",
  "schedule_match_vods",
  "schedule_muted_vods",
  "scheduler.clipdownload.age",
  "scheduler.clipdownload.amount",
  "scheduler.clipdownload.enabled",
  "server_port",
  "storage.clean_empty_vod_folders",
  "storage.deleted_cloud",
  "storage.no_watch_files",
  "storage_per_streamer",
  "template.vodsplit.filename",
  "template.vodsplit.folder",
  "thumbnail_format",
  "trust_proxy",
  "twitch.voddownload.auth_enabled",
  "twitchapi.auth_type",
  "twitchapi.enable_gql",
  "twitchapi.eventsub_type",
  "twitchapi.eventsub_unsub_on_start",
  "twitchdownloader_path",
  "video.chapters.title",
  "vod_container",
  "vod_folders",
  "vods_to_keep",
  "websocket_client_address",
  "websocket_enabled",
  "youtube.client_id",
  "youtube.client_secret",
  "youtube.quota_override"
]

[DerBunteBall]

That structure is fine. It's nothing else in my case.

I don't think you configuration got invalid. As I said a thing I could imagine is that there is a slight difference in the images used.

I'm not a JavaScript/TypeScript Developer. But the normal approach would be to define the config as a object in the application data model. So that can be validated and so on. That's handed to the API component by the Controller and send out by the View.

That's a relatively complex process. It might be possible that the Container Image is different. I installed my setup a few months ago. I never pulled again in that time because the repo has no dependabot or so. If you build manually or the Docker Hub Repo builds new from time to time there can be differences in the Images depending on how dependencies are defined.

E.g. if the packages.json says I always need version 1.0.1 of X it would take this. But if it says I need "newest" of Y it takes this. So if some component e.g. is defined like Y is and something changed it's possible that your Image - because of the moment when it got pulled - has everything like my image except Y. If Y is part of the process described above a slight change might broke the validation or something like that.

Since I installed my setup the code in the repo never changed. So the only thing that might be changed is the dependency tree in the Docker Images. Because I never pulled again I have the Image from Docker Hub at I think something like mid of April. I'm not sure but I thought that Docker Hub builds the Image new every day or week from master. Just check that on Docker Hub.

If you pulled within mid of April and today you might have a slight different Container Image like my Image is. And I think there will be the reason. I posted my Imge Hash above.

[DerBunteBall]

Whoops forgot to say that my Image is the arm64 variant.

And the Images aren't rebuild on Docker Hub. So the thing just can happen if the Image is build manually.

I would try to increase logging. See the config keys with "debug" in it's name. Play around a bit with them and check whether you get more output with that in the container log.

[BigFlubba]

Sorry, I can't be much of help, but from my experience, all I did was run the compose file, let it build, then edit the ENV and settings to my preferences in Portainer. Since then, I have destroyed, edited, & repulled the image many times. Other than my mistakes and the bugs I created issues for, I have not had other major issues. If you want, I can share my config that I use. I have found these settings to be the most stable.

My instance is running on Docker inside of a VM with 16 CPU threads, 8GB RAM, & 300GB NVMe storage. Hardware-wise, I can use fewer threads, but I have plenty of CPU headroom, so I don't bother with it. I do need more RAM because when it is remuxing a long VOD, it uses all of it (granted, I am running 12 other containers on that VM.