Frame-Relay & EIGRP

[mcechman]

I have a large Frame-relay network with multiple point to point PVC's written everywhere. When a PVC goes down usually I ping it and it responds that my router locally times out and or destination host cannot be reached, but in either case the response comes back from my defualt gateway. Here as of recent, I have noticed that when a PVC goes down and I attempt to ping it, the packet travels around the network and the destination unreachable response comes from another router somewhere else. What is causing these packets to be routed to a dead destination all of a sudden? What can I do to ensure that when I attempt to ping a down site that my timeouts respond from the default gateway as opposed to somewhere else? Like I said this has always been the case and this new qwirk just came up. We are using EIGRP and are using the network 123.456.000.000 statment on each router so that each router has the full route table.

[s-doyle]

EIGRP may take some time to converge, after PVC goes down. In the meantime, routing tables in th enetwork will be unstable and the route taken by the packet will be unpredictable.

[raymond.lucas]

See if you have a default or summary route floating around.

The command "show ip route 123.456.0.0 255.255.0.0 longer" might be of use here.

Ray

[mcechman]

I did that command, and I can see that there is no gateway of last resort set and that all the routes appear to be there and in the right places....

Matt

[paupot]

Did the FR network have some redundancy built in? If from a distant location a new gw of last resort is redistributed into eigrp, your gw might be overruled.

sh ip eigrp events might help, or add the log-eigrp-neighbor changes into your router eigrp config.

or - more work - summarize your remote locations towards the core of the fr cloud and set the default gw for each location separately....

hth

regards

paul