On 11/30/12 7:15 PM, John Neiberger wrote:
I thought I'd post an update since I found my answer. Marko Milivojevic
answered on another mailing list. As it turns out, the router still
compares metrics for the next hop even if they're not both learned from
IGPs. So, the path with an OSPF metric of 101 is losing out to a path with
a BGP-learned next hop with a MED of 0. I wouldn't have expected that
behavior at all!
I'll certainly defer to Marko, but I think your next-to-last sentence
should finish with "with an IGP metric of 0". IGP metric and MED are
two separate concepts, which unfortunately appear in the 'sh ip b'
output both as "metric":
R2#show ip bgp 100.100.100.0/24
BGP routing table entry for 100.100.100.0/24, version 12
Paths: (3 available, best #3, table Default-IP-Routing-Table)
Flag: 0x900
Advertised to update-groups:
1 2 3
Local, (Received from a RR-client)
5.5.5.5 (metric 102) from 5.5.5.5 (100.100.100.100)
Origin incomplete, metric 0, localpref 100, valid, internal
Here, "(metric 102)" is the IGP cost to egress point, and "metric 0" is
the BGP MED value.
When the IGP path disappears and the router reverts to 0/0 as
reachability for the egress point, (metric 102) becomes (metric 0),
which as 0 disappears from the command output.
R2#show ip bgp 100.100.100.0/24
BGP routing table entry for 100.100.100.0/24, version 13
Paths: (3 available, best #1, table Default-IP-Routing-Table)
Flag: 0x900
Advertised to update-groups:
1 2 3
Local, (Received from a RR-client)
5.5.5.5 from 5.5.5.5 (100.100.100.100)
Origin incomplete, metric 0, localpref 100, valid, internal, best
It sounds to me like you're experiencing stale BGP routes. You might
want to explore shortening your BGP timers and/or whether BFD would
allow BGP to dump paths learned from an (now-unreachable) OSPF neighbor.
Or, random thought, no idea if this would work, what if you put "ip ospf
cost 103' on the interface towards the source of your EBGP 0/0?
Grabbing at straws here...coffee hasn't kicked in.
Also, what about carrying 0/0 in OSPF instead?
pt
_______________________________________________
cisco-nsp mailing list [email protected]
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/
