Support / questions about Lemmy.
I've opened some pull requests on the main app repo and the docs repo with changes that should fix the docker install issues people have been having.
Specifically: https://github.com/LemmyNet/lemmy/pull/2908 / https://github.com/LemmyNet/lemmy-docs/pull/186
The hero we need.
Ok, looking at it after it was committed:
The instructions are not very clear on the proxy setup. It kind of hand-waves a suggested config after the "To make Lemmy available outside the server, you need to set up a reverse proxy, like Nginx. You can use the following simple proxy:" preface, but is not really specific which file this goes in. Is this appended to the nginx.conf?
Thank you
I mostly didn't touch that part, it seemed to be outside the scope of the article and suggesting a config for each person's exact setup isn't going to be possible.
This would be for putting in front of the nginx proxy built into the compose file. In theory you could do SSL offloading here or whatever you want, as long as you don't change the headers needed by the nginx.conf file further up the page.
I'm not sure if I should make a new post or keep this in a single thread, anyway, I've managed to get the front page to load, but can't get any farther.
I used the instructions from @jax@lemmy.cloudhub.social's branch, particularly their nginx config.
I'm running the whole docker-compose setup behind my own nginx reverse proxy, which is set to proxy requests to the docker nginx (and also handles SSL). I modified the docker-compose.yml to bind the host port on 9001, then my main nginx config looks like this:
http {
server {
listen 443 ssl http2;
server_name lemmy.villa-straylight.social;
ssl_certificate /etc/letsencrypt/live/<redacted>
ssl_certificate_key /etc/letsencrypt/live/<redacted>
location / {
proxy_set_header Host $host;
proxy_pass http://localhost:9001;
}
}
}
The front page loads, but pretty much anything else I do results in an endless spinner. The only possible problem I see is this in the docker terminal output:
WARN Error encountered while processing the incoming HTTP request: lemmy_server::root_span_builder: WebSocket upgrade is expected. NoWebsocketUpgrade
Your reverse proxy is doing websockets incorrectly, there's some bogosity in the spec for websockets that makes them single hop and the proxy needs to propagate the upgrade request. Search for "nginx reverse proxy websocket", e.g. https://www.nginx.com/blog/websocket-nginx/
Thanks, that's the missing piece. Ironically I am a software engineer in my day job, but not the right kind of software engineer. It's fun to broaden my horizons though. I copied the proxy settings from the @jax@lemmy.cloudhub.social's nginx.conf with the right edits for the hostname into my reverse proxy config and now everything appears to be working!
Glad it's working for you!
Yeah, that sounds like a web sockets issue. I think one part of the nginx.conf file from my docs has proper web sockets handling that you can copy into your config.
I just finished my setup and I had to do A LOT of tweaking to get it to all stand up and talk it everything correctly. The documentation (and even the Ansible script) is not up to date sadly.
Feel free to message me and I can share my configs with you. They worked on a dedicated VM.
There were some issues for me when I set mine up, what issues are you having? I am using 0.17.3 personally.
Should only need docker compose, lemmy.hjson, and an nginx proxy conf file if you use it, but sounds like you're using something else
I'm running into the problem mentioned 2 months ago on Reddit with trying to install via docker. I'm on Ubuntu Server.
I solved it just copying the nginx.conf file
The commits of @jax@lemmy.cloudhub.social have the solution.
https://github.com/LemmyNet/lemmy/pull/2908
https://github.com/LemmyNet/lemmy-docs/pull/186
You can just update the config of nginx to another port and then use your own reverse proxy, in my case I'm using caddy.
The instance is up, but I can't search communities of other instances yet '^-^
Some people are reporting outbound networking issues/DNS resolution in the lemmy matrix chat, I wonder if you're hitting the same thing. Do you see anything about DNS resolution errors in the logs when trying to search for other communities?
The issue in this post for me was already fixed.
Now I do have another issue, not sure if it's related to what you're mentioning.
I'm getting this error when trying to search for a community in another instance.
HTTP request{http.method=GET http.scheme="https" http.host=lemmy.pe1uca.dev http.target=/api/v3/ws otel.kind="server" request_id=36d10fe3-81a9-45f0-b39a-5299b2760ed6 http.status_code=101 otel.status_code="OK"}: lemmy_server::api_routes_websocket: couldnt_find_object: Request error: error sending request for url (https://lemmy.one/c/privacyguides): operation timed out
Here's the post https://lemmy.one/post/16124
You should be able to use any reverse proxy to replace the one it includes, but you'll want to understand the various things that need to happen. Specific HTTP methods and content-types must be routed to the lemmy and not the lemmy-ui container, and not all of these (e.g. paths) are well documented...I just found this out the hard way: watch the caddy logs carefully as you are dealing with federation issues.