Route redistribution can cause frequent DDR connections
© 2OO2, Cisco Systems, Inc. All rights reserved. Cisco CCIE Prep v1.O—Module 7-83
Reason 3: Redistribution from a classful routing protocol is being performed on one of the routers that connect to the OSPF demand circuit.
This is probably the most common reason that ISDN links configured as OSPF demand circuits still flap. It is also the hardest to troubleshoot and fix. In the example above, the ISDN link between R4 and R1 is 172.16.14.0/24 and is configured as an OSPF demand circuit. R4 is redistributing Enhanced Interior Gateway Routing Protocol (EIGRP) routes into OSPF. This causes many problems for an OSPF demand circuit on an ISDN link.
Since the encapsulation type on the ISDN link is Point-to-Point Protocol (PPP), both routers install a host route for the other side of the link as shown below.
R4# show ip route 172.16.14.1 Routing entry for 172.16.14.1/32
Known via "connected", distance 0, metric 0 (connected, via interface) Routing Descriptor Blocks: * directly connected, via BRI0/0
Route metric is 0, traffic share count is 1
EIGRP, IGRP, and Routing Information Protocol (RIP) are classful routing protocols. Therefore, the network statement in R4's EIGRP configuration is for the classful network of 172.16.0.0. This classful network statement causes the router to believe that the host route of 172.16.14.1/32 is being originated by EIGRP and gets redistributed into OSPF as an external route as shown below.
R4# show ip ospf database external 172.16.14.1
OSPF Router with ID (220.127.116.11) (Process ID 1) Type-5 AS External Link States
Options: (No TOS-capability, DC) LS Type: AS External Link
Link State ID: 172.16.14.1 (External Network Number )
Advertising Router: 18.104.22.168
LS Seq Number: 80000001
Length: 3 6
Network Mask: /32
Metric Type: 2 (Larger than any link state path) TOS: 0 Metric: 20
Forward Address: 0.0.0.0 External Route Tag: 0
Here is the problem: when the ISDN link goes down the /32 host route will disappear from the routing table. OSPF understands this as a change in topology and the demand circuit brings the link up again to propagate the MAXAGE version of the /32 host route to its neighbor. When the link comes up, the /32 mask gets inserted into the routing table again and the LSA age gets reset. After the first dead time interval on the link, the link goes down again. This process repeats itself, and the demand circuit link keeps flapping.
Solution: Use the no peer neighbor-route command
Under the ISDN BRI interface, configure the no peer neighbor-route command. This command prevents the /32 host route from being installed in the routing table. This command is only needed on the router performing redistribution, but is recommended for both sides of the ISDN link for consistency.
Was this article helpful?