Docker Compose + Nginx reverse proxy: error 502 Bad Gateway, Connection reset by peer

[Andrewskiz]

Hey, I just installed Trilium via Docker Compose and I am having trouble with nginx and a reverse proxy.

Here is my docker-compose.yml (I changed the port since I have another service running on port 8080)

version: '2.1'
services:
  trilium:
    image: zadam/trilium
    restart: unless-stopped
    environment:
      - TRILIUM_DATA_DIR=/home/node/trilium-data
    ports:
      - "127.0.0.1:8050:8050"
    volumes:
      - trilium:/home/node/trilium-data

volumes:
  trilium:

and the reverse proxy section in my nginx config:

location /trilium/ {
    proxy_pass http://127.0.0.1:8050/;
    proxy_http_version 1.1;
    proxy_set_header Upgrade $http_upgrade;
    proxy_set_header Connection 'upgrade';
    proxy_set_header Host $host;
    proxy_cache_bypass $http_upgrade;
}

There are no errors in the docker container logs, and this is the status on docker:

CONTAINER ID   IMAGE                         COMMAND                  CREATED        STATUS                  PORTS                                NAMES
************   zadam/trilium                 "docker-entrypoint.s…"   23 hours ago   Up 23 hours (healthy)   127.0.0.1:8050->8050/tcp, 8080/tcp   triliumdocker_trilium_1

When I try to visit https://mydomain.com/trilium/ I get error 502 Bad Gateway. This is the error I am getting in my nginx log:

2023/01/24 17:12:01 [error] 1110637#1110637: *1565 recv() failed (104: Connection reset by peer) while reading response header from upstream, client: my ip, server: mydomain.com, request: "GET /trilium/ HTTP/1.1", upstream: "http://127.0.0.1:8050/", host: "mydomain.com"

I'm still new to docker and reverse proxies, so I've probably made a error somewhere. Any help would be greatly appreciated, thank you!

[LukeSkywalker993]

same here since last Update of nginx

i using teslamate and ip-symcon.

[Nriver]

Check the output of curl http://127.0.0.1:8050/.

[Andrewskiz]

curl: (56) Recv failure: Connection reset by peer

[Nriver]

I suppose you did not change the running port of Trilium in the ini config file, and make the port mapping to an empty port. Try to modify the ports part of the compose file to - "8050:8080" and restart your docker conatiner.

[Andrewskiz]

I suppose you did not change the running port of Trilium in the ini config file, and make the port mapping to an empty port. Try to modify the ports part of the compose file to - "8050:8080" and restart your docker conatiner.

I changed the ports section set to - "8050:8080" I also tried it without the -d argument and got this output:

trilium_1  | No USER_UID specified, leaving 1000
trilium_1  | No USER_GID specified, leaving 1000
trilium_1  | DB not initialized, please visit setup page - http://[your-server-host]:8080 to see instructions on how to initialize Trilium.
trilium_1  | DB size: 4 KB
trilium_1  | Trusted reverse proxy: false
trilium_1  | App HTTP server starting up at port 8080

I am not sure if it is still trying to run on port 8080 despite specifying 8050 in the docker-compose.yml file

Docker ps shows this:

CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
************ zadam/trilium "docker-entrypoint.s…" 4 minutes ago Up 46 seconds (healthy) 0.0.0.0:8050->8050/tcp, :::8050->8050/tcp, 8080/tcp triliumdocker_trilium_1

And I'm still get the 502 Bad Gateway from nginx / Recv failure: Connection reset by peer from curl

[return215]

My setup is similar, but I use nginx-proxy and acme-companion containers to also setup HTTPS for my site and my notes.

Here is my docker-compose config output:

services:
  server:
    container_name: nginx-proxy
    image: nginxproxy/nginx-proxy:1.1
    ports:
    - published: 80
      target: 80
    - published: 443
      target: 443
    restart: always
    volumes:
    - certs:/etc/nginx/certs:rw
    - vhost:/etc/nginx/vhost.d:rw
    - www:/usr/share/nginx/html:rw
    - /var/run/docker.sock:/tmp/docker.sock:ro
  site:
    container_name: site
    depends_on:
    - ssl
    environment:
      LETSENCRYPT_EMAIL: [email protected]
      LETSENCRYPT_HOST: mydomain.xyz
      VIRTUAL_HOST: mydomain.xyz
    image: nginx:stable
    restart: always
  ssl:
    container_name: acme-companion
    depends_on:
    - server
    environment:
      DEFAULT_EMAIL: [email protected]
      NGINX_PROXY_CONTAINER: nginx-proxy
    image: nginxproxy/acme-companion:2.2
    restart: always
    volumes:
    - certs:/etc/nginx/certs:rw
    - vhost:/etc/nginx/vhost.d:rw
    - www:/usr/share/nginx/html:rw
    - /var/run/docker.sock:/tmp/docker.sock:ro
    - /var/run/docker.sock:/var/run/docker.sock:ro
    - acme:/etc/acme.sh:rw
  trilium:
    container_name: trilium
    depends_on:
    - ssl
    environment:
      LETSENCRYPT_EMAIL: [email protected]
      LETSENCRYPT_HOST: notes.mydomain.xyz
      VIRTUAL_HOST: notes.mydomain.xyz
      VIRTUAL_PORT: 8080
    image: zadam/trilium:0.58-latest
    ports:
    - published: 8080
      target: 8080
    restart: always
    volumes:
    - certs:/home/node/certs:rw
    - acme:/home/node/acme:rw
    - /var/local/trilium-data:/home/node/trilium-data:rw
version: '3.3'
volumes:
  acme: {}
  certs: {}
  vhost: {}
  www: {}

The logs from nginx-proxy returns this when I access my site and my notes subdomain.

nginx-proxy | nginx.1     | mydomain.xyz ***.***.***.*** - - [10/Feb/2023:02:19:23 +0000] "GET / HTTP/2.0" 304 0 "-" "Mozilla/5.0 (X11; Linux x86_64; rv:109.0) Gecko/20100101 Firefox/109.0" "***.***.***.***:80"
nginx-proxy | nginx.1     | 2023/02/10 02:19:24 [error] 77#77: *33 upstream prematurely closed connection while reading response header from upstream, client: ***.***.***.***, server: notes.mydomain.xyz, request: "GET / HTTP/2.0", upstream: "http://***.***.***.***:8080/", host: "notes.mydomain.xyz"
nginx-proxy | nginx.1     | notes.mydomain.xyz ***.***.***.*** - - [10/Feb/2023:02:19:24 +0000] "GET / HTTP/2.0" 502 157 "-" "Mozilla/5.0 (X11; Linux x86_64; rv:109.0) Gecko/20100101 Firefox/109.0" "***.***.***.***:8080"

All other containers run without any errors. However, I noticed something in docker-compose ps (also in docker ps but basically the same as this):

acme-companion   /bin/bash /app/entrypoint. ...   Up                                                                                     
nginx-proxy      /app/docker-entrypoint.sh  ...   Up               0.0.0.0:443->443/tcp,:::443->443/tcp, 0.0.0.0:80->80/tcp,:::80->80/tcp
site             /docker-entrypoint.sh ngin ...   Up               80/tcp                                                                
trilium          docker-entrypoint.sh ./sta ...   Up (unhealthy)   0.0.0.0:8080->8080/tcp,:::8080->8080/tcp  

trilium container is in "unhealthy" status but running. Is there any errors in my setup and how should I go about fixing it?

[return215]

Forgot to add that I can successfully access my site with HTTPS but it returns "502 Bad Gateway" on the notes subdomain. This is what causes the output of the reverse proxy server. Stopping and restarting the compose yields the exact same result.