Gateway timeout on one of the services/containers

Traefik Traefik v2 (latest)
1

I have the following docker-compose setup. I have two nginx containers that are connected with
traefik. The one labeled with foobar.test works, the one labeled with foobar.test && pathPrefix /admin doesn't work most of the time. Sometimes it works, but most of the time after bringing the containers down and back up it doesn't work anymore.

version: '3.7'

services:
  traefik:
    image: traefik:v2.0
    command: --api.insecure=true --providers.docker --providers.docker.exposedByDefault=false
    ports:
      - '80:80'
      - '8080:8080'
    volumes:
      - /var/run/docker.sock:/var/run/docker.sock
    networks:
      - proxy

  mysql:
    container_name: mysql
    image: mysql:8.0.19
    command: --default-authentication-plugin=mysql_native_password
    environment:
      - MYSQL_ROOT_PASSWORD=$DB_PASSWORD
    ports:
      - 3306:3306
    volumes:
      - mysql_data:/var/lib/mysql
    networks:
      - mysql

  client:
    image: nginx:1.17.10-alpine
    volumes:
      - ./client/public:/usr/share/nginx/html
    labels:
      - traefik.http.routers.foobar_client.rule=Host(`foobar.test`)
      - traefik.enable=true
    networks:
      - proxy

  admin_nginx:
    build:
      context: ./admin
      dockerfile: ./docker/nginx/Dockerfile.dev
    volumes:
      - ./admin/public:/app/public
    networks:
      - proxy
      - admin
    labels:
      - traefik.http.routers.foobar_admin.rule=Host(`foobar.test`) && PathPrefix(`/admin`)
      - traefik.enable=true

  admin_php-fpm:
    container_name: php-fpm
    build:
      context: ./admin
      dockerfile: ./docker/php-fpm/Dockerfile.dev
    volumes:
      - ./admin:/app
    networks:
      - admin
      - mysql

networks:
  proxy:
    name: proxy
    driver: bridge
  mysql:
    name: mysql
    driver: bridge
  admin:
    name: admin
    driver: bridge

volumes:
  mysql_data:
    name: mysql_data

Any ideas what I'm doing wrong? Thnx in advance!

2

One other thing, if I add the admin network to the traefik container everything works. But that is rather strange because they (traefik and admin_nginx) already share the network called proxy.

3

As the container has two networks, you need to tell traefik which network to use. It is currently flipping between the two and only one works.

- traefik.docker.network=proxy

4

Thank you very much! That was it. In hindsight pretty logical, but I didn't knew that could happen.

5

Found out the hard way myself.

I kind of think that Traefik should be able to match networks. Actually TraefikEE supports this and will connect the Proxy to the correct network, so might just be a pay to play only feature.