With a new release 0.9.0, arm build is finally available. It is not a big deal to properly install it following instructions, but problem is with exposing your instance to the internet, bc of 2 open ports requirement (1235 for ui and 8536 for a server). I have tried to use tunneling services like ngrok and pagekite to expose localhost on 1235 port, but in this case I got access to ui only without ability to execute any action.
Have you managed to successfully expose your instance? How?
with rocket.chat for ex when you run an instance, it is available through localhost:3000. To expose it you use port 3000 with tunneling service like ngrok (./ngrok http 3000). So it looked logical to reproduce such approach :)
Lemmy consists of a couple different services, thats why its a bit more complicated 😃
I think the setup instructions should be exactly the same for a raspberry pi, since it can run both nginx and docker.
ok, I set nginx up based on your instructions and connected tunneling for 443 port. with default nginx lemmy.conf it gets
400 Bad Request The plain HTTP request was sent to HTTPS port. But if replaceserver { listen 433 ssl http2; ... }withserver { listen 433 ; ... }it works smoothly without any error. Any suggestions? Maybe there is a need to add force redirect to https on 433 in lemmy.conf?If you handle SSL elsewhere, then you can remove anything related to SSL from the nginx config (and also the first server block which redirects from http to https).
I’m not sure why not having ssl or http2 in there would fix it… you’d have to send me your whole nginx for me to see what’s wrong.
I have manged to expose it with
ssl http2in nginx, problem was in my tunneling setup, which I have fixed.sent :)