VPN Bind or Killswitch

[thob97]

I have a question regarding the setup this repo offers, and I would greatly appreciate any insights you can provide.

I was wondering if it directly binds transmission to the VPN, or if it incorporates a killswitch mechanism to disconnect the internet in case the VPN connection is lost. I've come across some information suggesting that killswitches may not be entirely secure, which has raised my curiosity about the specific approach employed in this setup.

Thank you kindly in advance for your assistance. I am grateful for any clarifications you can offer.

Best regards,
Thor

[edgd1er]

2 protections ( 1 optional) + 1 kill switch).
based on route-up/route-pre-down:

docker-transmission-openvpn/openvpn/start.sh

Line 232 in 42eb2ee

TRANSMISSION_CONTROL_OPTS="--script-security 2 --route-up /etc/openvpn/tunnelUp.sh --route-pre-down /etc/openvpn/tunnelDown.sh"

• tunnelUp script bind transmission to private ip:
  

  docker-transmission-openvpn/transmission/start.sh

  Lines 28 to 32 in 42eb2ee

  	echo "Updating TRANSMISSION_BIND_ADDRESS_IPV4 to the ip of $dev : $ifconfig_local"
  	export TRANSMISSION_BIND_ADDRESS_IPV4=$ifconfig_local
  	# Also update the persisted settings in case it is already set. First remove any old value, then add new.
  	sed -i '/TRANSMISSION_BIND_ADDRESS_IPV4/d' /etc/transmission/environment-variables.sh
  	echo "export TRANSMISSION_BIND_ADDRESS_IPV4=$ifconfig_local" >>/etc/transmission/environment-variables.sh

• when openvpn disconnects, tunnelDown script is launched and transmission is killed (stop.sh):
  

  docker-transmission-openvpn/transmission/stop.sh

  Lines 27 to 34 in 42eb2ee

  	# Check whether transmission-daemon is still running
  	if [[ -z "$(pidof transmission-daemon)" ]]
  	then
  	echo "Successfuly closed transmission-daemon"
  	else
  	echo "Sending kill signal (SIGKILL) to transmission-daemon"
  	kill -9 "$PID"
  	fi

• OPTIONAL: default route is removed after vpn is up. (off by default)
  

  docker-transmission-openvpn/transmission/start.sh

  Lines 90 to 95 in 42eb2ee

  	if [[ "true" = "$DROP_DEFAULT_ROUTE" ]]; then
  	echo "DROPPING DEFAULT ROUTE"
  	# Remove the original default route to avoid leaks.
  	/sbin/ip route del default via "${route_net_gateway}" || exit 1
  	fi

[thob97]

I want to express my sincere appreciation for your prompt response. Thank you very much.

If this script utilizes route-pre-down, then it should provide a secure solution, as it would halt transmission before removing the disconnected VPN connection. I'm curious why I've come across various sources stating that killswitches are considered unsafe. Perhaps other killswitches use different implementations which are less secure.

Furthermore, I would like to inquire whether you recommend utilizing both protections or if the default option alone is sufficient in ensuring a secure setup.

Once again, thank you for your assistance. I am grateful for any further insights you can provide.

Best regards

[edgd1er]

There are also questions when using wireguard (another vpn solution): https://unix.stackexchange.com/questions/733430/does-wireguard-postup-predown-really-work-as-a-kill-switch.
Killing the app on route change and binding it to the vpn interface is pretty safe for me. I don't see how on vpn termination, the application wound connect directly to the internet. bound interface is down, vpn ip is removed, and quickly transmission is killed.

I also add a blocklist to transmission.

Deleting default route is an extra protection, depending on the risk of leaking you are willing to take.

You may run some tests with https://bash.ws/torrent-leak-test to check if leakage is detected when you disable the vpn interface ( ifconfig )