"Nix, Robert P." wrote (with some snippage): > HOME > 192.168.36.1 VLINK1 > 192.168.30.6 CHIP2 > 192.168.31.6 CHIP3 > 192.168.37.3 LINUX1 > > GATEWAY > 192.168.37 = LINUX1 4096 0.0.0.255 0 > ;
Technically, this is not correct, since CTC is a point-to-point link. I use a HOST route, like the following, to define the address at the other end of the link: GATEWAY 192.168.10.4 = CTC0 1500 HOST > BSDROUTINGPARMS TRUE > VLINK1 4096 1 255.255.255.0 0 > CHIP2 4096 0 255.255.255.0 192.168.30.1 > CHIP3 4096 0 255.255.255.0 192.168.31.1 > ; LINUX1 4096 0 255.255.255.0 192.168.37.2 > ENDBSDROUTINGPARMS > > We've tried this with and without the GATEWAY statement and BSDROUTE for > LINUX1. I assume you're using a dynamic routing protocol (RIP, OSPF)? BSDROUTINGPARMS is intended as configuration for the OROUTED process, and if you're not actually running one then results may be undefined. > W/o it, LINUX1 doesn't show up in NETSTAT GATE at all. This might be the case if you are running OROUTED, and there is not a corresponding process (gated or similar) running on your Penguin. It might also be because the CTC is a non-broadcast medium, and defining the link 'incorrectly' (refer above) might be confusing OROUTED... > With it, it > shows up, and we can ping the Linux system from VM and VM from Linux. But > Linux can't get to anything past VM and nothing outside of VM can get to > Linux. If you're not running gated or similar on your Penguin, this is the correct config for VM TCPIP, but you need a little more to get everything going. Looks like you need one or both of the following: * Linux needs to have a default route (static) defined, or to learn one via dynamic routing from VM TCPIP. Unless you intend to have multiple VM TCPIP stacks providing CTCs to your Penguin (as we started to explore in the ISP/ASP Redbook) a static default route via your VM TCPIP will be fine. * the rest of the network needs to know/learn what the route to your Penguin is. This can be either static routes, or again using a dynamic routing protocol. If you are running OROUTED and RIP, for example, you can set up OROUTED to import static routes into the RIP domain for re-advertisement to the Ciscos (so the routers learn about the route coded in the GATEWAY statement). Similar configs can be done with OSPF. >From what you've said, I don't believe your problem to be associated with the VIPA... Have a look at the routing table on Linux (the output of the 'netstat -r' or 'route' commands are useful here) and in the Ciscos (show ip route... maybe? My IOS is a little rusty) Hope this is helpful. > ---- > "In theory, theory and practice are the same, > but in practice, theory and practice are different." How true! :) Cheers, Vic Cross
