Closed skyajal closed 8 years ago
Thanks, but this way you solve one issue by creating another one. The availability of eth0 is not always given but can vary. Making it possible to select which interface to bind to is a better approach in my opinion.
Why don't we solve this issue using openvpn route-up and down hook scripts in the conf itself. There is one env variable that can be used to fetch the interface which is
route_net_gateway The pre-existing default IP gateway in the system routing table. Set prior to --up script execution.
With "route-up" you can add the masquerade rules fetching the local lan with the previous env variable. With down you can delete the openvpn added rules.
Discard my last comment, that var is for the client. Will have to look for another option
Disregard this request as it is no longer valid.
Fixes a bug with the docker-gui plugin installed not to use docker0 as the binding port. eth0 has been hardcoded as the binding port.