Comment 5 for bug 1178273

Cool, thanks for your response.

I think that the required change will probably need to be more sophisticated than the one you showed (and I advised against hot-fixing your quantum server this way); reason being, the dhcp agent's behavior might be unexpected, especially if there was a port already associated with the gateway ip.

I also noticed that by using the quantum client, I am unable to reset the gateway-ip, as None or null do not parse correctly (they end up being marshalled into a string). I will follow that up with the client-side of the project, but for now, I think it would be sensible to gather more input/opinion from whom is more experienced with dhcp than I am, as I suspect that _removing_ a gateway from an existing subnet is not something that was initially planned.