Hello, on my OpenBSD 3.9 borderrouter I configured a BGP session to my core-router and to several external bgp-neighbors. The core-router announces my prefixes via iBGP to my borderrouters. These announced this prefixes via eBGP to my neighbors and thus to the world.
For my local transfer network between my borderrouter and the corerouter I use a /30 network of one of my prefixes which my corerouter announces (via the borderrouters) to the world. Let's say I announce 192.168.0.0/23 to the world and uses 192.168.1.0/30 as my transfer network to the one borderrouter (Cisco) and 192.168.1.4/30 to the other (OpenBGPd). This constellation works fine. To get a fully meshed borderrouter-setup I set up a direct link between the openbsd and the cisco router and configured 192.168.1.8/30 on the interfaces. Now I can notice a really strange behavior: the prefix 192.168.0.0/23 is correctly announced to the world, I can reach targets within 192.168.0.0/24 without any problem. But I can't reach any target within 192.168.1.0/24, neither from the world nor from the openbsd-router. On the router I get the error message "network unreachable". A "bgpctl sh rib 192.168.1.1" gives me a correct routing-information for 192.168.0.0/23. Any idea? Regards, Falk [demime 1.01d removed an attachment of type APPLICATION/DEFANGED which had a name of fb.14675DEFANGED-vcf]
