Is it possible to have Lemmy and Lemmy UI on separate subdomains? #876
Comments
|
docker-compose: https://github.com/LemmyNet/lemmy-ansible/blob/main/templates/docker-compose.yml nginx: https://github.com/LemmyNet/lemmy-ansible/blob/main/templates/nginx.conf It could be a port issue, I'm not sure. I don't use traefik so I can't help you there. We only support nginx. |
|
I don't think the issue was directly related to the nginx/traefik configuration. Both services were running and serving content. Lemmy UI was just trying to make websocket requests to itself rather than to Lemmy. So it was trying to connect to wss://lemmyui.thetamail.com/api/v3/ws instead of wss://activitypub.thetamail.com/api/v3/ws. With the sample nginx config this isn't a problem because it just splits requests between services based on path/request type. I ended up just setting things up such that they both ran on the same domain similar to the sample nginx config. My updated docker-compose.yml with routing rules similar to the nginx config, in case anyone else comes across this: version: '3.5'
services:
lemmy:
image: dessalines/lemmy:0.16.7
hostname: lemmy
networks:
- postgres
- proxynet
restart: always
environment:
- RUST_LOG="warn,lemmy_server=info,lemmy_api=info,lemmy_api_common=info,lemmy_api_crud=info,lemmy_apub=info,lemmy_db_schema=info,lemmy_db_views=info,lemmy_db_views_actor=info,lemmy_db_views_moderator=info,lemmy_routes=info,lemmy_utils=info,lemmy_websocket=info"
volumes:
- ./lemmy.hjson:/config/config.hjson
depends_on:
- pictrs
labels:
- "traefik.enable=true"
- "traefik.docker.network=proxynet"
- "traefik.http.routers.activitypub.rule=Host(`activitypub.thetamail.com`) && (
(
Headers(`Accept`, `^application/.*$$`)
) ||
(
Method(`POST`)
) ||
(
PathPrefix(`/api`) ||
PathPrefix(`/pictrs`) ||
PathPrefix(`/feeds`) ||
PathPrefix(`/nodeinfo`) ||
PathPrefix(`/.well-known`)
)
)"
- "traefik.http.routers.activitypub.entrypoints=websecure"
- "traefik.http.routers.activitypub.service=activitypub"
- "traefik.http.services.activitypub.loadbalancer.server.port=8536"
lemmy-ui:
image: dessalines/lemmy-ui:0.16.7
networks:
- proxynet
environment:
# this needs to match the hostname defined in the lemmy service
- LEMMY_INTERNAL_HOST=lemmy:8536
# set the outside hostname here
- LEMMY_EXTERNAL_HOST=activitypub.thetamail.com
- LEMMY_HTTPS=true
depends_on:
- lemmy
restart: always
labels:
- "traefik.enable=true"
- "traefik.docker.network=proxynet"
- "traefik.http.routers.lemmyui.rule=Host(`activitypub.thetamail.com`) && !(
(
Headers(`Accept`, `^application/.*$$`)
) ||
(
Method(`POST`)
) ||
(
PathPrefix(`/api`) ||
PathPrefix(`/pictrs`) ||
PathPrefix(`/feeds`) ||
PathPrefix(`/nodeinfo`) ||
PathPrefix(`/.well-known`)
)
)"
- "traefik.http.routers.lemmyui.entrypoints=websecure"
- "traefik.http.routers.lemmyui.service=lemmyui"
- "traefik.http.services.lemmyui.loadbalancer.server.port=1234"
pictrs:
image: asonix/pictrs:0.3.1
# this needs to match the pictrs url in lemmy.hjson
hostname: pictrs
# we can set options to pictrs like this, here we set max. image size and forced format for conversion
# entrypoint: /sbin/tini -- /usr/local/bin/pict-rs -p /mnt -m 4 --image-format webp
networks:
- proxynet
environment:
- PICTRS__API_KEY=API_KEY
user: 991:991
volumes:
- ./volumes/pictrs:/mnt
restart: always
labels:
- "traefik.enable=true"
- "traefik.docker.network=proxynet"
- "traefik.http.routers.pictrs.rule=Host(`pictrs.thetamail.com`)"
- "traefik.http.routers.pictrs.entrypoints=websecure"
- "traefik.http.routers.pictrs.service=pictrs"
- "traefik.http.services.pictrs.loadbalancer.server.port=8080"
networks:
proxynet:
external:
name: common_network
postgres:
external:
name: postgresNote that I'm still in the early stages of setup so this may not be fully functional. I know that I can browse around my instance, create posts, and federate with lemmy.ml well enough to at least subscribe to communities though. |
|
I just wanted to throw you a quick super big thanks @Cardinal8501. I can confirm this still works today with Traefik and Lemmy 0.17.2. 💙 |
|
You must now use |
|
@tgxn is correct, for my and for my |
|
Hi! I came here looking for a working solution. It would be nice if we could back it in a seperate docker-compose.override.yml to leave original config intact. While writing this I wish I could do this stuff myself - but I have very little time at hands (children). This is what GPT returned to me; I had no chance to verify until now |
|
This should be included in the default docker |
I was originally writing this issue out because I thought I had misconfigured something or encountered some bug, but re-reading it, thinking about it and looking at the Nginx config I'm not so sure. If this actually just isn't allowed everything written below doesn't matter. If it is intended to work then please read on.
I've been trying to set up a Lemmy instance on my domain. Lemmy itself seems to be working (I can use it through the Jerboa app). Lemmy UI hangs indefinitely when I try to log in. I can see errors in the console like this:
However, my Lemmy instance is actually hosted at activitypub.thetamail.com. From what I can tell lemmy-ui has no /api/v3/ws route but lemmy itself does so I think the requests should be going to, e.g., wss://activitypub.thetamail.com/api/v3/ws.
I can see that my lemmy-ui docker logs reference the websocket URI I'm expecting:
but the logs in the console do not:
Here's what the lemmy-ui section for my docker-compose.yml looks like. It's pretty similar to the sample one:
I tried explicitly setting the LEMMY_UI_LEMMY_WS_HOST environment variable as well, made no difference.
I can see if I get a shell in the container that the correct environment variable is set. What's more the lemmy-ui pages even include a reference to activitypub.thetamail.com in their metatags:
but it's not using the right domain in the websocket URI.
The text was updated successfully, but these errors were encountered: