From: Chad Marsh (chad@xxxxxx)
Date: Sat Oct 02 1999 - 02:40:50 GMT-3
Consider the following:
AS100 | AS100 | AS8
| |
Confed 65456 | Confed 65107 |
| |
| R10 |
| | |
| | |
R5------|------R7-------|-----R8
| |
| |
There are quite a few other routers in this scenario (10 total), but the
problem lies in this area.
AS 100 is addressed primarily out of the 172.16.0.0/24 block, is running
OSPF internally, and is advertising 172.16.0.0/16 via eBGP to AS8, no
problem.
AS 8 is advertising 192.168.1.0/24 to AS100 via eBGP no problem.
The link between R7 and R8 is on the 150.10.x.x network, and is not
participating in any IGP.
So here comes the problem:
R10 has a loopback address of 200.200.200.1/24, and is inserting this
network into bgp like so:
router bgp 65107
bgp confederation identifier 100
neighbor 172.16.70.7 remote-as 65107 <this is R7
network 200.200.200.0
The loopback interface is also participating in OSPF, and is set to 'ip
ospf network point-to-point' (12.05XK) so it is advertising
200.200.200.0/24 into OSPF, and all of the AS100 routers have a route to
200.200.200.0/24
R7 is running BGP syncronized
R7 has the following entry in it's BGP table:
Network Next Hop Metric LocPrf Weight Path
*>i200.200.200.0 172.16.70.10 0 100 0 i
R7 is advertising 200.200.200.0 to AS8 (as it should, since the route is
synchronized) but is not advertising it to it's confederation peer R5.
Damn, nevermind, I just figured it out. My route-map applied out to
neighbor R5 was only permitting the network from AS8.
Oh well, I spent so much time typing this email, I'm gonna send it
anyway.
Looks like fun stuff eh?
Chad Marsh
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 08:21:52 GMT-3