2024-02-25, 10:47 PM
(This post was last modified: 2024-02-25, 10:58 PM by powergrim1234. Edited 4 times in total.)
I am having a problem that I just cannot wrap my head around.
I used to have a set-up of a Jellyfin server with some companion apps bare-metal on a fairly heavyweight and power-hungry server. I have switched that over to a more energy-efficient solution.
The former solution was exposed to the internet via NGinx Proxy Manager(NPM). This was working just fine. No additional configuration was set via de advanced tab in NPM.
But now that I have switched over to my newer machine things start working worse. I expose it through the same configuration as my former jellyfin server and it works for about 30 seconds. After that I get 502 Bad Gateway errors for about 5 tot 10 minutes and after that the next attempt works fine for about 30 seconds.
What I have tried sofar:
- Run with advanced configuration, several even. One from the officials docs, one from Reddit. All in all, no change in behaviour.
- Spun up a container instead of bare-metal. The result is the same
- Tried different setups regaring websocket support, no change at all
- Compared absolutely everything with my old server. I can't see anything different, and yet that machine is working fine (still is, haven't wiped it as of yet)
I am really lost as to how to fix this, I have read up how to spin up Traefik. But honestly, it just doesn't fit my usecase and I like the ease of NPM.
Couple of facts:
- DNS externally is Cloudflare. It is a full DNS passthrough
- Firewall has not been touched, it points to the reverse proxy
- Internal and external DNS names are fully different. So no weird internal lookups.
- Allow remote connections is enabled
- The proxy is listed in the networking tab of Jellyfin. Allthough I don't know if it will work. I put my dockerhost IP in there, would that be the issue?
- NPM is fully IP-based, and HTTPS enabled and forced for external connections
- NPM listens on 443, and is configured to redirect to port 8096.
Does anyone have the slightest idea what logfile would help me out of where to look to resolve this?
I used to have a set-up of a Jellyfin server with some companion apps bare-metal on a fairly heavyweight and power-hungry server. I have switched that over to a more energy-efficient solution.
The former solution was exposed to the internet via NGinx Proxy Manager(NPM). This was working just fine. No additional configuration was set via de advanced tab in NPM.
But now that I have switched over to my newer machine things start working worse. I expose it through the same configuration as my former jellyfin server and it works for about 30 seconds. After that I get 502 Bad Gateway errors for about 5 tot 10 minutes and after that the next attempt works fine for about 30 seconds.
What I have tried sofar:
- Run with advanced configuration, several even. One from the officials docs, one from Reddit. All in all, no change in behaviour.
- Spun up a container instead of bare-metal. The result is the same
- Tried different setups regaring websocket support, no change at all
- Compared absolutely everything with my old server. I can't see anything different, and yet that machine is working fine (still is, haven't wiped it as of yet)
I am really lost as to how to fix this, I have read up how to spin up Traefik. But honestly, it just doesn't fit my usecase and I like the ease of NPM.
Couple of facts:
- DNS externally is Cloudflare. It is a full DNS passthrough
- Firewall has not been touched, it points to the reverse proxy
- Internal and external DNS names are fully different. So no weird internal lookups.
- Allow remote connections is enabled
- The proxy is listed in the networking tab of Jellyfin. Allthough I don't know if it will work. I put my dockerhost IP in there, would that be the issue?
- NPM is fully IP-based, and HTTPS enabled and forced for external connections
- NPM listens on 443, and is configured to redirect to port 8096.
Does anyone have the slightest idea what logfile would help me out of where to look to resolve this?