Kacheong Poon writes: > Could Seb please comment on the IP tunnel service? I guess tunnel > links can only be setup after there is network connectivity. So > it makes sense to have the tunnel service dependent on milestone/ > network.
Requiring that network interfaces be set up before setting up tunnels sounds to me like a bug, not a feature. A tunnel is a detached object. It doesn't explicitly use any particular underlying network interface. The network interface actually used at any moment in time depends on both the kernel forwarding table (routing) and the set of interfaces that might be up at any given instant. This means that as routes and interfaces come and go with time, a given tunnel may or may not actually be usable. It might be nice to have the dynamic local routing and interface state reflected in the tunnel as IFF_RUNNING, but it seems like a mistake to me to require this state to be present in order to configure the tunnel in the first place. It's a mistake at least for the reason that it makes configuration indeterminate: there's no way to know whether any given configuration attempt will succeed or fail, as the required underlying facilities might slip out from under you as you configure the tunnel. It makes configuration itself unreliable. Yes, I know that's how it works today. I think it's a bug, though. I don't like the idea of codifying bugs and design errors into SMF dependencies. -- James Carlson, KISS Network <james.d.carlson at sun.com> Sun Microsystems / 1 Network Drive 71.232W Vox +1 781 442 2084 MS UBUR02-212 / Burlington MA 01803-2757 42.496N Fax +1 781 442 1677
