Static DHCP leases held by VPN interfaces

[lggomez]

Hi everyone, after solving this issue that took me several days to identify I wanted to check if this is an expected behavior or otherwise a bug

I have multiple VPN client interfaces used by hosts with static DHCP leases on the following setup:

1. Multiple VPN client interfaces configured (both OpenVPN and wireguard). Only some of them turned on
2. Multiple static DHCP leases, including 2 hosts bound (or being bound before) to these VPN interfaces via VPN director rules
3. Custom DNS pointing to said hosts themselves (as they are my DNS resolvers)

After messing with the DNS and VPN client configurations I suddenly lost connection on these 2 devices. After removing all DNS overrides I found that the hosts still failed to resolve external hosts, so the next thing I did was removing the static DHCP leases. This is where it got weird; after removing the leases, rebooting the router several times and flushing the caches on the hosts, the associations were still standing

It was after a while and looking at a netstat of static routes that I found that the hosts were still bound to these static IPs via the VPN client interfaces (in this case, the 2 were wireguard clients). After I disabled the VPN clients and removed VPN director rules the hosts were free and leasing dynamic IPs as expected

So, to sum it up, the question is, are static DHCP leases expected to be preserved by VPN director rules?

Thanks in advance