I'll defer to David on this, since, in part, I don't have the exact
lab in front of me. But remember that BGP is there both to advertise
your routes to other AS, and to accept routes from other AS. Could
your configuration have been trying to bring up a session to listen
to the other AS, even if it didn't have anything to announce? Also,
remember that BGP does have a keepalive, although the keepalive often
is disabled in practice because other mechanisms will detect failures
faster.
>Hi All,
> I was just making my way through a couple ISDN/DDR Snapshot =
>routing scenarios and made a unlikely observation.
>
>For reference purposes I was making use of my CZone =
>privileges(disclaimer) in mocking up David wolsefer ISDN lab exercise =
>and got the following results. Before attempting this lab I worked =
>through another BGP related lab. In saying so I moved right into this =
>scenario without performing a write erase on the routers. To my benefit =
>one of the routers I used in this ISDN mock up was clean and not used in =
>the previous lab. =20
>
>OK, so I get everything all configured baseline snapshot (just the ISDN =
>circuit and 1 loopback) and it works great. I progressed to follow the =
>requirements of David's scenario which makes use of the Ethernet circuit =
>and commands to support ISDN backup of the Ethernet line. Here is where =
>things get interesting.... The ISDN line keep flapping up..down..up.. =
>down. =20
>
>The "debug ip packet", "debug dialer packet" , and debug dialer events" =
>revealed that the client-side of the snapshot circuit was trying to make =
>a connection to 11.1.1.2 (eth0). As mentioned before I was doing a BGP =
>lab and the Snapshot server router was one of the routers used in that =
>scenario. Although there was no ip configurations for the address =
>11.1.1.2, the bgp process on the snapshot server keep trying to make a =
>tcp connection on the segment(11.1.1.2). This caused the ISDN line to =
>try and route packets to that address as defined in the running bgp =
>process (neighbor 11.1.1.2 remote-as 1).=20
>
>Once I removed the bgp process everything worked as it should. My =
>questions now go to the fact that I have an ospf process running that =
>hasn't caused any problems at all. I'm trying to understand what I =
>experienced. =20
>
>In knowing BGP uses protocol TCP port 179, OSPF IP port 89, and RIP UDP =
>port 520. Now I recognize that the "dialer-list" used in the exercise =
>is baseline(dialer-list 1 protocol ip permit), but does this make any =
>sense at all. Why would this non-active bgp connection cause the ISDN =
>line to flap... There was no redistribution being performed so isn't =
>this a good example of "ships in the night" routing? =20
>
>Thoughts anyone.
>
>TIA
>Nigel
>
>
>
>
>
>_________________________________
>FAQ, list archives, and subscription info:
>http://www.groupstudy.com/list/cisco.html
>Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
_________________________________
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
