Troubleshooting guide

Additional BGP Configuration

The peering router would also use a loopback interface in this scenario. The neighbor address

configured on each router would be the IPv4 address of the loopback interface on the peering router.

Since the loopback address enables the BGP neighbors to use an IPv4 address that is not reachable by

a directly connected interface, separate static routes would need to be specified for each individual link

that can reach the neighbor’s loopback IPv4 address.This ensures that the loopback address on the

BGP peer remains reachable if one of the links goes down.

Establishing Routing Preference

When a route is redistributed into BGP from a static route, another routing protocol, or when a prefix is

specifically advertised using a network statement in BGP Configuration mode, BGP must consider it in its

route calculations.

Assuming the LOCAL_PREF value of the two routes is equal, the BGP decision logic prefers routes

learned by the local router from statically configured routes or routing protocols over routes learned from a

BGP peer.

After the best valid route has been selected by the BGP decision logic, the administrative distance for BGP

is applied to the route. The routing table then chooses the best route to install based on the lowest

administrative distance. If the BGP algorithm chooses a locally learned route with a higher administrative

distance that is already installed in the routing table, it will not replace that existing route since it points to

the same next-hop IPv4 address. As this specific scenario is manifested, it can be observed in the output of

the show ip route command. The routing source will show the locally learned route from its original

source instead of a BGP sourced route (denoted with a B).

The best way to make a static route or a prefix learned from another routing protocol more or less

preferable than any route learned using BGP is to redistribute the protocol into BGP (using the

redistribute [connected | ospf | rip | static] [metric <value> | route-map <map>] command) and use a

route map to modify the local preference of the intended routes. If this approach is used, an additional

prefix list might need to be configured and applied outbound to all neighbors to prevent unwanted

redistributed subnets from being advertised.

The default LOCAL_PREF value for BGP routes is 100. To prefer the redistributed route over a route

learned through BGP, the LOCAL_PREF of the redistributed route must be set to a value higher than 100.

To prefer a route learned through BGP over a redistributed route, an inbound route map should be applied

to the appropriate BGP neighbor to set the LOCAL_PREF of the intended route higher than 100.

The following example configuration demonstrates the basic commands used to prefer a static route or a

route learned through another routing protocol (local prefix) over a BGP learned route:

route-map REDISTRIBUTE permit 10

match ip address prefix-list <name>

Assuming both routes have an equal LOCAL_PREF value, it doesn’t matter if the

administrative distance of a locally learned route is configured higher than that of BGP.

The local route will still be used instead of the route learned from the BGP peer. This

scenario occurs only when the same route is learned locally and by BGP.

Despite administrative distance and assuming equal LOCAL_PREF value, BGP learned

routes are always less preferable than routes learned by the local router from statically

configured routes or routing protocols.