OK, this might just be my misunderstanding of OSPF, so just want to run this by you and see if it is a mistake on my behalf. Let me try and explain:
In this case I have a number of routers (OpenBSD 5.0 boxes running ospfd and bgpd, except .106 which is a Cisco) which all share a common network to peer traffic (vlan50 213.133.64.96/27). The also are all on a management network (vlan1 192.168.111.0/24). They each have various other vlans connected to various other networks. The problem I have is that I don't seem to be able to control which router is used to send traffic to a host on the management vlan (e.g. 192.168.111.31). So, on 213.133.64.103: # ospfctl sh dat | grep 192.168.111. 192.168.111.0 213.133.64.98 924 0x800000ce 0x91ba 192.168.111.0 213.133.64.101 500 0x80000094 0xf38f 192.168.111.0 213.133.64.102 1361 0x80000005 0x0d05 192.168.111.0 213.133.64.104 1963 0x80000011 0x674c 192.168.111.0 213.133.64.106 19 0x80000012 0x5957 # ospfctl sh rib | grep 192.168.111. 192.168.111.0/24 213.133.64.98 Type 1 ext Network 110 02:22:39 192.168.111.0/24 213.133.64.101 Type 1 ext Network 110 02:22:39 192.168.111.0/24 213.133.64.102 Type 1 ext Network 110 02:22:34 Firstly, why don't all the routes in the ospf database appear in the rib? The problem I'm seeing is a packet going via 213.133.64.103 as a router destined to 192.168.111.31 gets sent via 213.133.64.98, but 192.168.111.31 has a default gateway of 192.168.111.1 which is the other interface on the router 213.133.64.106. So I get asymmetric packet flows. I want to somehow say that 213.133.64.106 is the preferred router when trying to reach 192.168.111.0/24. Yet seeing as all of the routers are on the same network segment I don't see how I can set the cost. Even if i could get the cost on vlan50 to take any effect (I've tried different values to no joy) it would surely affect *all* routes to that router, not just 192.168.111.0/24. -Matt