/30 on interface but shows C /30 & L /32 on route table and r> on BGP

[hanadul1234]

2 cisco rourters, directly connected and assigned /30 ip addresses on each interfaces.
I can see /30 as Connected and /32 as Local on show ip route on both routers.
I enabled BGP and redistribute connected.  No filter, no static and no other protocols or configs.  It is really simple config.
I can see "r>" which means RIB failure on show ip bgp (eBGP each other).
I suspect I got "r>" on BGP table because BGP also advertised /32 along with /30 but not sure.

My question is
Why /32 automatically showing on routing table (someone said it is because of CEF and normal)?
Can I eliminate /32 on routing table?
How can I eliminate "r>" RIB failure on BGP table (apply route-map and no advertise /32?)?

interface GigabitEthernet0/0
description To XXXX G0/0 (AS65530)
ip address 10.9.0.2 255.255.255.252
duplex auto
speed auto
media-type rj45

router bgp 65531
bgp log-neighbor-changes
neighbor 10.9.0.1 remote-as 65530

Router# show ip route
C 10.9.0.0/30 is directly connected, GigabitEthernet0/0
L 10.9.0.1/32 is directly connected, GigabitEthernet0/0

Router# show ip bgp
r> 10.9.0.0/30 10.9.0.1 0 0 65530 ?

[M02@rt37 · Accepted solution]

Hello @hanadul1234 

The /32 routes automatically appearing in your routing table are a result of cisco's implementation of connected and local routes. When an IP address is assigned to an interface, cisco IOS creates a /32 route to efficiently handle packets destined for that specific IP address, which is a behavior tied to cisco express forwarding (CEF). This allows the router to optimize traffic management to its own interfaces. Therefore, the /32 routes are essential for the router's operation and cannot be removed from the routing table. They serve a critical function in ensuring that packets destined for the router's own addresses are correctly and efficiently routed.

To resolve the "r>" RIB failure in the BGP table, which occurs because BGP is trying to install routes that are already more specifically covered by the /32 local routes, you can prevent BGP from advertising these /32 prefixes. This can be done using a route-map to filter out the /32 prefixes before they are advertised to the BGP peer. By creating a prefix-list to match the /32 routes and then applying this list in a route-map to deny the advertisement of these routes, you ensure that only the desired /30 route is advertised. Youwill eliminate the RIB failure in the BGP table, as BGP will no longer attempt to advertise routes that conflict with the existing more specific local routes...

[MHM Cisco World]

Your bgp neighbor advertise link between two routers via bgp' 

Note:- bgp never advertise/32 it adverise "c" not "L". 

Your router since it have better lower AD (direct connect) than neighbor advertise bgp prefix it mark this prefix in bgp as "r"

This need to solve in your neighbor not in your router' it can make issue sometimes.

Use filter in your router not solve problem' check neighbor router use redistrubte connect with route-map and deny "c" not "L"

MHM

[Giuseppe Larosa]

Hello @hanadul1234 ,

>> I suspect I got "r>" on BGP table because BGP also advertised /32 along with /30 but not sure

This is not a real issue because the subnet is direclty connected if the eBGP neighbor has other connected prefixes to advertise they will be advertised .

the code "r" just means a better routing source provides the same prefix.

You can avoid this kind of small issue by using network commands under BGP avoiding to advertise the subnet used for eBGP peering

Hope to help

Giuseppe