Kill Switch - VPN shared containers need to be manually restarted

[rofocalus]

When the transmission kill switch activates, and the container restarts - the other containers in the stack that attach to the VPN via network_mode: "service:transmission-openvpn" lose internet access as well (by design, good)

However, when the transmission container reboots and gets internet access again the other containers that attached to the container network are still left without any internet access until I go and restart those containers manually.

Is there something I'm missing? What would be a good fix for this.

Thanks

[ilike2burnthing]

https://docs.docker.com/reference/compose-file/services/#depends_on

[ilike2burnthing]

I believe restart: always is also required.

[rofocalus]

version: '3.3'
services:
    transmission-openvpn:
        image: haugene/transmission-openvpn
        cap_add:
            - NET_ADMIN
        volumes:
            - '/redacted:/data'
            - '/redacted:/config'
            - '/redacted:/etc/openvpn/custom'
        environment:
            - OPENVPN_PROVIDER=redacted
            - OPENVPN_CONFIG=redacted
            - OPENVPN_USERNAME=redacted
            - OPENVPN_PASSWORD=redacted
            - OPENVPN_OPTS=--mute-replay-warnings --inactive 3600 --ping 10 --ping-exit 60
            - CREATE_TUN_DEVICE=true
            - TRANSMISSION_WEB_UI=flood-for-transmission
            - TRANSMISSION_RATIO_LIMIT=100
            - TRANSMISSION_RATIO_LIMIT_ENABLED=TRUE
            - LOCAL_NETWORK=redacted/redacted
            - TRANSMISSION_DOWNLOAD_DIR=redacted
            - TRANSMISSION_INCOMPLETE_DIR=redacted
            - PUID=redacted
            - PGID=redacted
        logging:
            driver: json-file
            options:
                max-size: 10m
        container_name: transmission
        hostname: transmission
        networks: 
          - transmission-network
        dns:
         - 8.8.8.8
         - 8.8.4.4
        restart: always
    prowlarr:
        image: lscr.io/linuxserver/prowlarr:latest
        container_name: prowlarr
        environment:
          - PUID=redacted
          - PGID=redacted
          - TZ=redacted/redacted
        volumes:
          - redacted:/config
        restart: unless-stopped
        network_mode: "service:transmission-openvpn"
        depends_on:
          transmission-openvpn:
            condition: service_healthy
            restart: true
        healthcheck:
          test: curl -f http://localhost:9091 || exit 1
          interval: 1m00s
          timeout: 10s
          retries: 2
    flaresolverr:
    # DockerHub mirror flaresolverr/flaresolverr:latest
          image: ghcr.io/flaresolverr/flaresolverr:latest
          container_name: flaresolverr
          environment:
            - LOG_LEVEL=info
            - LOG_HTML=false
            - CAPTCHA_SOLVER=none
            - TZ=redacted/redacted
          restart: unless-stopped
          volumes:
            - redacted:/config
          network_mode: "service:transmission-openvpn"
          depends_on:
            transmission-openvpn:
              condition: service_healthy
              restart: true
          healthcheck:
            test: curl -f http://localhost:9091 || exit 1
            interval: 1m00s
            timeout: 10s
            retries: 2
    slskd:
         image: slskd/slskd
         container_name: slskd
         environment:
           - SLSKD_REMOTE_CONFIGURATION=true
           - USER=redacted:redacted
           - SLSKD_UMASK=0000
         volumes:
           - /redacted:/app
           - /redacted/:/media
         restart: always
         network_mode: "service:transmission-openvpn"
         depends_on:
            transmission-openvpn:
              condition: service_healthy
              restart: true
         healthcheck:
            test: wget http://localhost:9091 || exit 1
            interval: 1m00s
            timeout: 10s
            retries: 2
    dispatcharr:
          image: ghcr.io/dispatcharr/dispatcharr:latest
          container_name: dispatcharr
          volumes:
            - /redacted:/data
          environment:
            - DISPATCHARR_ENV=aio
            - REDIS_HOST=localhost
            - CELERY_BROKER_URL=redis://localhost:6379/0
            - DISPATCHARR_LOG_LEVEL=info 
          runtime: nvidia
          deploy:
            resources:
              reservations:
                devices:
                  - driver: nvidia
                    count: all
                    capabilities: [compute, utility, video]
          restart: always
          network_mode: "service:transmission-openvpn"
          depends_on:
            transmission-openvpn:
              condition: service_healthy
              restart: true
          healthcheck:
             test: curl -f http://localhost:9091 || exit 1
             interval: 1m00s
             timeout: 10s
             retries: 2
    mousehole:
          container_name: mousehole
          restart: unless-stopped
          network_mode: "service:transmission-openvpn"
          depends_on:
            transmission-openvpn:
              condition: service_healthy
              restart: true
          image: tmmrtn/mousehole:latest
          volumes:
            - redacted:/srv/mousehole
          healthcheck:
               test: curl -f http://localhost:9091 || exit 1
               interval: 1m00s
               timeout: 10s
               retries: 2
networks:
  transmission-network:
    name: transmission-network
    driver: bridge
    ipam:
      driver: default
      config:
        - subnet: redacted/redacted

i tried like this, took down the bridge with setting "ip link dev br-transmission down" waiting for it to go unhealthy and restart, waited for all containers to be unhealthy, set the docker bridge back up, transmission is back to healthy however the other containers are still unhealthy and unreachable

[ilike2burnthing]

Did you try just:

depends_on:
  - transmission-openvpn

https://haugene.github.io/docker-transmission-openvpn/tips-tricks/#let_other_containers_use_the_vpn

[rofocalus]

yes thats what i started with beforehand

[ilike2burnthing]

If you just manually restart the Transmission container, do the others restart?