IP forwarding with Nginx Proxy Manager (NPM)

[TheDreadPirate]

Put 172.17.0.1 as a known proxy.

[Ploutos]

Well I added 172.17.0.1 to the known proxies list (https://litter.catbox.moe/stki3l.png) and restarted Jellyfin. Nothing has changed.

---

The IP showed is always the same (https://litter.catbox.moe/g1kdet.png)

[TheDreadPirate]

I looked at your Jellyfin docker compose again. It's using host networking? But Nginx is using bridge networking with port publishing? Usually it is the other way around. Nginx is running on the host network while Jellyfin is in a bridge network.

From a client PC, can you run a tracert/tracepath to your domain name to see the hops it is going through?

Also, remove these from the Advanced tab of Nginx.

Code:

proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_set_header Host $host;

Looking at Nginx Proxy Manager's git, those properties are already part of Nginx's config and aren't necessary and could be interfering with Nginx's default config.

[Ploutos]

I cant put my Nginx Proxy Manager in Host network mode because there is already a Nginx installed in DSM on my Synology NAS. So I have to put NPM in a docker network then do a strange port forwarding because my 80 and 443 ports are already busy.

I could use the native Nginx but it dosnt have an UI and I have no idea on how to use it (well I find it kinda difficult)

I removed the code of the Advanced tab of NPM restarted both container (NPM first then Jellyfin) nothing has changed

I did a tracert to jellyfin.mydomain.com but there is only one hop on my public IP (from a client in the local network)

From a client using 5G internet it is not working well I have a lot of timeout results

[Ploutos]

Here the result with the tracert to jellyfin.mydomain.com

C:\Users\user>tracert myserver.mydomain.com

Tracing route to jellyfin.mydomain.com [XXX.XXX.XXX.XXX]
over a maximum of 30 hops:

1 4 ms 9 ms 6 ms isp-gateway.example.com [XXX.XXX.XXX.XXX]

Trace complete.

[TheDreadPirate]

What you could try doing is switching Jellyfin to the bridge network. Then in Nginx Proxy Manager you would use the Jellyfin container's bridge IP. That reduces how many times traffic switches between host and bridge networking.

[Ploutos]

Ok I"ll try to put Jellyfin and NPM in the same bridge network.

Good idea.

[Ploutos]

Now both container are in the same network (https://privatebin.net/?eeabc2f956e5940d...cw3f1hK82K)

(https://litter.catbox.moe/37zl0a.png)

I added the IP adress in Jellyfin and restarted Jellyfin (https://litter.catbox.moe/ew6reh.png)

It is still the IP of the gateaway that is showed in Jellyfin

My nginx proxy manager configuration is still the same exept the IP that I changed for the IP of jellyfin in the bridge network (https://litter.catbox.moe/e7g5qk.png)

---

It is odd because the IP showed in Jellyfin is the IP of the gateaway (https://litter.catbox.moe/gv1pt4.png) even though they are in the same network.

Why would the HTTP request go through the 172.17.0.1 (IPV4 IPAM Gateway)

[TheDreadPirate]

172.30.0.1? But you removed it from the known proxies. It would need to still need to be there. Also, why is NPM in a separate bridge network?

---

Oh. And disable "Cache Assets". There is no benefit for Jellyfin and consumes a lot of resources.

---

Backing up. What is the IP for Nginx? 172.17.0.4?

[Ploutos]

No they are both in the same bridge network I changed their network and put both of them in the default bridge network created natively by docker
(Sorry I messed up my link in the previous post, I fixed it)

(https://litter.catbox.moe/37zl0a.png)

---

Because I changed their network, Docker changed their IP addresses

---

NPM has now 172.17.0.4
and Jellyfin has now 172.17.0.3

---

I will disable cache assets thanks