I have some unreliable gateways (wireless links), and when they go in a “fault” state as detected by the ping test in net balancer, I would think this should disable attempting to use the gateway, until it tests good again.
However, while the link is down (the AP gets turned off for hours, so it’s not just flaky and coming up and down), ZS/openVPN constantly tries to bring the VPN up, and on the remote end/VPN server, I see attempts, and via the wrong gateway.
It would be nice if the net balancer and vpn scripts could work together, so that when net balancer detects the gw is in fault state, it will pause/stop openVPN. The reason for this is that sometimes, rarely, the VPN will actually manage to come up over the wrong gateway, and since I have my vpns in a bond interface, if it’s a priority vpn/bond link, then it will take over, which results in the vpn being unusable.
So there’s 2 issues, that the VPN keeps trying to connect over a faulty/down gateway, and that the traffic is being sent via the wrong gateway than it is assigned to in the netbalancer.