Bizzare IP routing behavior

From: Eric Shao (ericshao@xxxxxxxx)
Date: Sat Jun 09 2001 - 12:31:34 GMT-3


   
Routing Protocol: OSPF, default static route(Last Resort of Gateway) points to
a Nokia firewall which runs OSPF.

Wks --- Rtr --- Firewall --- DMZ(172.16.10.0)

ip route 0.0.0.0 0.0.0.0 IP-of-firwall
ip route 172.16.10.0 255.255.255.0 IP-of-firewall (bandaid static route)

A workstation behind the Router can ping 172.16.10.10(DMZ). Without the
bandaid static route, multiple core OSPF routers' ping/extended ping packets
were simply dropped by the router himself. (debug ip icmp and tcpdump on
firewall did not capture any packet), but one OSPF core router was able to
ping the DMZ machine. As soon as the bandaid route was added, ping worked fine
on all the problem routers.(which confirms firewall is not the issue here)

The question is, has anyone ever encountered this problem that appears default
route/Last resort of gateway does not function for the router while,
workstations behind the problem router still can get to the destination? The
interim solution for now is to redistribute this bandaid static route into
OSPF core. which is a duplicate effort with the existing default static
route.

My hunch is, Any packet initiated from the router is route processed while the
external packet from the workstation is switch processed. Somehow the route
process screwed up on several routers, hard to understand?! IOS is 12.0(10)

I can't post the configuration here because of the company policy, nor can I
recycle one of the problem router to verify my thought till next scheduled
downtime.

Please send me email directly if you had similar experience or know what
causes the bizzare behavior.

TIA,
Eric
**Please read:http://www.groupstudy.com/list/posting.html



This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 10:31:21 GMT-3