Skip to content

Use Google DNS servers

Some have encountered problems with DNS resolving inside the docker container. This causes trouble because OpenVPN will not be able to resolve the host to connect to. If you have this problem use Docker's --dns flag and try using Google's DNS servers by adding --dns 8.8.8.8 --dns 8.8.4.4 as parameters to the usual run command.

Restart the container if the connection is lost

If the VPN connection fails or the container for any other reason loses connectivity, you want it to recover from it. One way of doing this is to set the environment variable OPENVPN_OPTS=--inactive 3600 --ping 10 --ping-exit 60 and use the --restart=always flag when starting the container. This way OpenVPN will exit if ping fails over a period of time which will stop the container and then the Docker daemon will restart it.

Let other containers use the VPN

To let other containers use VPN you have to add them to the same Service network as your VPN container runs, you can do this by adding network_mode: "service:transmission-openvpn". Additionally, you have to set depends_on to the transmission-openvpn service to let docker-compose know that your new container should start after transmission-openvpn is up and running. As the final step, you can add healthcheck to your service.

Example (Jackett):

As an example, let's add Jackett to the transmission-openvpn network based on the example from Running the container:

version: '3.3'
services:
    transmission-openvpn:
        cap_add:
            - NET_ADMIN
        volumes:
            - '/your/storage/path/:/data'
        environment:
            - OPENVPN_PROVIDER=PIA
            - OPENVPN_CONFIG=france
            - OPENVPN_USERNAME=user
            - OPENVPN_PASSWORD=pass
            - LOCAL_NETWORK=192.168.0.0/16
        logging:
            driver: json-file
            options:
                max-size: 10m
        ports:
            - '9091:9091'
            - '9117:9117'  # This is Jackett Port – managed by VPN Service Network
        image: haugene/transmission-openvpn
    jackett:
        image: lscr.io/linuxserver/jackett:latest
        container_name: jackett
        environment:
            - PUID=1000
            - PGID=1000
            - TZ=Europe/London
            - AUTO_UPDATE=true #optional
            - RUN_OPTS=<run options here> #optional
        volumes:
            - <path to data>:/config
            - <path to blackhole>:/downloads
        # You have to comment ports, they should be managed in transmission-openvpn section now.
#       ports:
#           - 9117:9117
        restart: unless-stopped
        network_mode: "service:transmission-openvpn" # Add to the transmission-openvpn Container Network
        depends_on:
            - transmission-openvpn # Set dependency on transmission-openvpn Container
        healthcheck: # Here you will check if transmission is reachable from the Jackett container via localhost
            test: curl -f http://localhost:9091 || exit 1
            # Use this test if you protect your transmission with a username and password 
            # comment the test above and un-comment the line below.
            #test: curl -f http://${TRANSMISSION_RPC_USERNAME}:${TRANSMISSION_RPC_PASSWORD}@localhost:9091 || exit 1
            interval: 5m00s
            timeout: 10s
            retries: 2

Check if the container is using VPN

After the container starts, simply call curl under it to check your IP address, for example with Jackett you should see your VPN IP address as output:

docker exec jackett curl -s https://api.ipify.org

You can also check that Jackett is attached to the VPN network by pinging it from the transmission-openvpn Container localhost:

docker exec transmission-vpn curl -Is http://localhost:9117
HTTP/1.1 301 Moved Permanently
Date: Tue, 17 May 2022 19:58:19 GMT
Server: Kestrel
Location: /UI/Dashboard

Example (Dante):

As an example, let's add Dante socks5 proxy to the transmission-openvpn network based on the example from Running the container:

version: '3.3'
services:
    transmission-openvpn:
        cap_add:
            - NET_ADMIN
        volumes:
            - '/your/storage/path/:/data'
        environment:
            - OPENVPN_PROVIDER=PIA
            - OPENVPN_CONFIG=france
            - OPENVPN_USERNAME=user
            - OPENVPN_PASSWORD=pass
            - LOCAL_NETWORK=192.168.0.0/16
        logging:
            driver: json-file
            options:
                max-size: 10m
        ports:
            - '9091:9091'
            - '1080:1080'  # This is Dante Socks5 Port – managed by VPN Service Network
        restart: unless-stopped
        image: haugene/transmission-openvpn

    socks5-proxy:
        image: wernight/dante
        restart: unless-stopped
        network_mode: service:transmission-openvpn
        depends_on:
            - transmission-openvpn
        command:
            - /bin/sh
            - -c
            - |
                echo "Waiting for VPN to connect . . ."
                while ! ip link show tun0 >/dev/null 2>&1 || ! ip link show tun0 | grep -q "UP"; do sleep 1; done
                echo "VPN connected. Starting proxy service . . ."
                sed -i 's/^\(external:\).*/\1 tun0/' /etc/sockd.conf
                sockd

Test Dante socks5 proxy

curl -x socks5h://{docker-host-ip}:1080 http://ip.ip-check.net

Bonus socks5 tip

With the Proxy SwitchyOmega Chrome/Edge extension, you can configure specific websites on your local machine to route through your socks5 proxy server.

Reach sleep or hibernation on your host if no torrents are active

By default, Transmission will always scrape trackers, even if all torrents have completed their activities, or they have been paused manually. This will cause Transmission to be always active, therefore never allow your host server to be inactive and go to sleep/hibernation/whatever. If this is something you want, you can add the following variable when creating the container. It will turn off a hidden setting in Transmission which will stop the application to scrape trackers for paused torrents. Transmission will become inactive, and your host will reach the desired state.

-e "TRANSMISSION_SCRAPE_PAUSED_TORRENTS_ENABLED=false"

Running it on a NAS

Several popular NAS platforms support Docker containers. You should be able to set up and configure this container using their web interfaces. As of version 3.0 of this image creates a TUN interface inside the container by default. This previously had to be mounted from the host which was an issue for some NAS servers. The assumption is that this should now be fixed. If you have issues and the logs seem to blame "/dev/net/tun" in some way then you might consider trying to mount a host device and see if that works better. Setting up a TUN device is probably easiest to accomplish by installing an OpenVPN package for the NAS. This should set up the device and you can mount it. There are some issues involved in running it on Synology NAS, Please see this issue that discusses solutions

Systemd Integration

On many modern Linux systems, including Ubuntu, systemd can be used to start the transmission-openvpn at boot time, and restart it after any failure.

Save the following as /etc/systemd/system/transmission-openvpn.service, and replace the OpenVPN PROVIDER/USERNAME/PASSWORD directives with your settings, and add any other directives that you're using.

This service is assuming that there is a bittorrent user set up with a home directory at /home/bittorrent/. The data directory will be mounted at /home/bittorrent/data/. This can be changed to whichever user and location you're using.

OpenVPN is set to exit if there is a connection failure. OpenVPN exiting triggers the container to also exit, and then the Restart=always definition in the transmission-openvpn.service file tells systems to restart things again.

[Unit]
Description=haugene/transmission-openvpn docker container
After=docker.service
Requires=docker.service

[Service]
User=bittorrent
TimeoutStartSec=0
ExecStartPre=-/usr/bin/docker kill transmission-openvpn
ExecStartPre=-/usr/bin/docker rm transmission-openvpn
ExecStartPre=/usr/bin/docker pull haugene/transmission-openvpn
ExecStart=/usr/bin/docker run \
        --name transmission-openvpn \
        --cap-add=NET_ADMIN \
        -v /home/bittorrent/data/:/data \
        -e "OPENVPN_PROVIDER=TORGUARD" \
        -e "OPENVPN_USERNAME=bittorrent@example.com" \
        -e "OPENVPN_PASSWORD=hunter2" \
        -e "OPENVPN_CONFIG=CA Toronto" \
        -e "OPENVPN_OPTS=--inactive 3600 --ping 10 --ping-exit 60" \
        -p 9091:9091 \
        --dns 8.8.8.8 \
        --dns 8.8.4.4 \
        haugene/transmission-openvpn
Restart=always
RestartSec=5

[Install]
WantedBy=multi-user.target

Then enable and start the new service with:

$ sudo systemctl enable /etc/systemd/system/transmission-openvpn.service
$ sudo systemctl restart transmission-openvpn.service

If it is stopped or killed in any fashion, systemd will restart the container. If you do want to shut it down, then run the following command and it will stay down until you restart it.

$ sudo systemctl stop transmission-openvpn.service
# Later ...
$ sudo systemctl start transmission-openvpn.service

Running with Traefik reverse proxy

A working example of running this container behind a traefik reverse proxy can be found here: Config

Running this container with Podman

The podman run command is almost identical to the one mentioned in README.md but with the following exception:

Instead --cap-add=NET_ADMIN you have to specify --cap-add=NET_ADMIN,NET_RAW,MKNOD. MKNOD and NET_ADMIN are necessary so Podman can create the tunnel, NET_RAW is needed for ping.