Hello!
I got my Lemmy server up and running locally yesterday yay (I set up some subs and test-posted), but today I must have tinkered with something because now the server only shows “Server error” (check code block 1 below for the server log).
Checking the logs, it seems it can’t “load” the site thumbnail. Curiously if I ask for it in firefox:
http://0.0.0.0:1236/pictrs/image/730840b6-d6ec-4a40-8668-36b89c6c1d33.png
it loads up (code block 2)
Any idea why Lemmy gets this “ECONNREFUSED” error ?
Cheers !
lemmy-ui_1 | FetchError: request to http://0.0.0.0:1236/pictrs/image/730840b6-d6ec-4a40-8668-36b89c6c1d33.png failed, reason: connect ECONNREFUSED 0.0.0.0:1236
lemmy-ui_1 | at ClientRequest.<anonymous> (/app/node_modules/node-fetch/lib/index.js:1505:11)
lemmy-ui_1 | at ClientRequest.emit (node:events:512:28)
lemmy-ui_1 | at Socket.socketErrorListener (node:_http_client:495:9)
lemmy-ui_1 | at Socket.emit (node:events:512:28)
lemmy-ui_1 | at emitErrorNT (node:internal/streams/destroy:151:8)
lemmy-ui_1 | at emitErrorCloseNT (node:internal/streams/destroy:116:3)
lemmy-ui_1 | at process.processTicksAndRejections (node:internal/process/task_queues:82:21) {
lemmy-ui_1 | type: 'system',
lemmy-ui_1 | errno: 'ECONNREFUSED',
lemmy-ui_1 | code: 'ECONNREFUSED'
lemmy-ui_1 | }
proxy_1 | 172.18.0.1 - - [18/Jul/2023:14:31:31 +0000] "GET / HTTP/1.1" 500 12 "-" "Mozilla/5.0 (X11; Linux x86_64; rv:109.0) Gecko/20100101 Firefox/115.0"
pictrs_1 | 2023-07-18T14:32:21.858505Z INFO HTTP request{http.method=GET http.route=/image/original/{filename} http.flavor=1.1 http.scheme=http http.host=pictrs:8080 http.client_ip=172.18.0.7:59168 http.user_agent=Mozilla/5.0 (X11; Linux x86_64; rv:109.0) Gecko/20100101 Firefox/115.0 http.target=/image/original/730840b6-d6ec-4a40-8668-36b89c6c1d33.png otel.name=HTTP GET /image/original/{filename} otel.kind="server" request_id=65877368-96ae-4ad9-a928-67a9bc3e3700}: tracing_actix_web::root_span_builder: new
pictrs_1 | 2023-07-18T14:32:21.861635Z INFO HTTP request{http.method=GET http.route=/image/original/{filename} http.flavor=1.1 http.scheme=http http.host=pictrs:8080 http.client_ip=172.18.0.7:59168 http.user_agent=Mozilla/5.0 (X11; Linux x86_64; rv:109.0) Gecko/20100101 Firefox/115.0 http.target=/image/original/730840b6-d6ec-4a40-8668-36b89c6c1d33.png otel.name=HTTP GET /image/original/{filename} otel.kind="server" request_id=65877368-96ae-4ad9-a928-67a9bc3e3700 trace_id=00000000000000000000000000000000 http.status_code=200 otel.status_code="OK"}: tracing_actix_web::root_span_builder: close time.busy=1.43ms time.idle=1.73ms
proxy_1 | 172.18.0.1 - - [18/Jul/2023:14:32:21 +0000] "GET /pictrs/image/730840b6-d6ec-4a40-8668-36b89c6c1d33.png HTTP/1.1" 200 12511 "-" "Mozilla/5.0 (X11; Linux x86_64; rv:109.0) Gecko/20100101 Firefox/115.0"
is it only images? Can API clients get comments, posts, users, communities?
I don’t really know, from FF it just blurts out the “Server error” (and keeps running). Is there some way I can test some API command from the commandline, or from the web maybe?
Cheers & tanks!
Edit: the working request in firefox seems to ask directly pictrs for the image. It feels like pictrs doesn’t like the Lemmy docker…
Here is a simple client API call you can do with wget or curl on the terminal:
curl "https://lemmy.ml/api/v3/community/list?sort=Hot&Limit=1"Change lemmy.ml to your domain name. This test assumes you created at least 1 community so far. This particular call works fine even in a browser, like Firefox: https://lemmy.ml/api/v3/community/list?sort=Hot&Limit=1
Thanks, they both work with http instead of https.
Https on firefox:
SSL_ERROR_RX_RECORD_TOO_LONG
curl:
curl: (35) error:0A00010B:SSL routines::wrong version number
Can it be so that the lemmy docker server tries to hit up pictrs on ssl?
That’s why I’m trying to get you to just fetch routine stuff that doesn’t involve picts.
Is the communities fetch giving you the failure with https? Then it’s likely your proxy config, nginx. Server ins’t setup right.
Yep everything fails with https.
I’ll try to validate my nginx.conf file and the nginx_internal.conf and see if I can figure what’s wrong, nginx is all new for me …
At the moment, I’m just trying to make it run locally, but your idea is great and I sure will use it when things start to roll! I mean if I ever get it to work locally :-)
Edit: https://0.0.0.0/pictrs/image/730840b6-d6ec-4a40-8668-36b89c6c1d33.png now works, but the https://0.0.0.0:1236 does still not
SSL works by hostname, not by IP address… so your teat isn’t valid to be using SSL. Nginx does the conversion (proxy) of SSL to non-SSL in Lemmy setups.
Well, it may be failing ONLY locally on the server, that’s why I emphasize to test remotely. As you may be bypassing nginx which does the SSL to non-SSL translation.
Seems very reasonable, I checked it out, but it’s the exact same problem (after I forwarded the 443 port too. BTW I haven’t tinkered with the ssl port);
https://lemmy.mindoki.com shows the server error
but
https://lemmy.mindoki.com/pictrs/image/730840b6-d6ec-4a40-8668-36b89c6c1d33.png
shows the image …
Also,
curl “https://lemmy.mindoki.com/api/v3/community/list?sort=Hot&Limit=1”
Works!
It drives me crazy :-) but I feel even more some ssl problem between the lemmy docker server and pictrs, no?
ok, based on those test results (which they work for me too)… then your lemmy_server is running, the API engine is responding, giving me the list of communities.
You could, at this point, try another client - a smartphone client - and see if you can use your server?
The front-page https://lemmy.mindoki.com not working is the lemmy-ui service not talking to the backend (which is what your logs in the post show). I don’t use Docker, so I’m not very good at figuring out how to get the two to talk together. But if a smartphone client works for you, then you know it’s lemmy-ui that’s your only real issue.