I noticed this bizarre behaviour and wondered if anyone has encountered this before and know why this is happeneing.
The scenario is as follows:
Each router has a loopback address corresponding to its router number. For example
R1's loopback 0 ip address is 188.8.131.52/32 in area 0
R2 is 184.108.40.206/32 in area 0
R3 is 220.127.116.11/32 in area 51 NSSA
[router id's are configured in the OSPF process of all routers]
R4 is 18.104.22.168/32 in RIPv2
The LAN 22.214.171.124/24 is in area 0. The WAN 126.96.36.199/30 is in area 51 NSSA. The WAN 188.8.131.52/30 is running RIPv2. R4's loopback is in RIP.
All routers are 7206 VXR NPE400 running 12.4(12) service provider feature set. I have also test this scenario with 2600 [ip plus] and 2800 routers.
On the NSSA ASBR [R3] which is rediatributing RIP into OSPF, it will redistribute 184.108.40.206/32 with the forwarding address of it's router id 220.127.116.11 because the WAN 18.104.22.168/30 is not enabled in OSPF. This bit I understand and expect.
The wierd bit is that once you configure a new loopback on R3 and add it to the OSPF process, the forwarding address of the NSSA external LSA 22.214.171.124 is changed to that address immediately!!! IT TOTALLY IGNORES THE ROUTER ID THAT WAS CONFIGURED!!! If you add another new loopback and add it to OSPF it changes the forwarding address again!!![this problem only affect NSSA ASBRs]
Anyone know why this is??