RIB failure in BGP

[M.Sultan]

Hi guys

Three questions:

what causes the RIB failure in BGP and what is the solution ?

As underlay protocol for full mesh connectivity we use EIGRP or OSPF, and overlay we use BGP, Does these two protocols yet need to be redistributed manually ?

R1 Loopback int 1

R2 Loopback int 2

R3 Loopback int 3

R4 Loopback int 4

However these above routers has networks in OSPF on physical interfaces but we established neighborship by adding only loopbacks into iBGP in inbound, in this case i only can ping loopback interfaces not the whole networks in neighbours router and vice versa. is this correct method ?

I know using loopback interface has facility of shifting path while a primary path goes down.

[MHM Cisco World]

https://www.cisco.com/c/en/us/support/docs/ip/border-gateway-protocol-bgp/213286-understand-bgp-rib-failure-and-the-bgp-s.html

[paul driver · Accepted solution]

Hello
It basically telling you the rtr has a better route already in it global route table GRT that is preferred over the same prefix in the bgp rib hence it becomes a bgp rib failure.

So an route learned dynamically via ibgp (AD200) and also via a IGP like OSPF (AD110) the ospf route is preferred due to it lower administrative distance, thus is entered into the GRT.

[M02@rt37 · Accepted solution]

Hello @M.Sultan,

BGP requires consistent routing information across its peers. If there is a mismatch in the advertised routes or attributes, it can lead to RIB failure. Also, incorrect configuration of BGP settings, such as neighbor relationships, route filters, or route redistribution, can cause RIB failures.

Solution to RIB failure?

--Verify the BGP configuration for any errors, including neighbor configurations, route filters, and redistribution settings. Ensure that the configuration is consistent across BGP peers.

--Ensure that the BGP neighbor relationships are established correctly and that the neighboring routers can exchange BGP updates.

--Verify that the routing information being advertised by BGP peers is consistent. Ensure that the advertised routes and attributes match on both sides.

If you have routers running EIGRP or OSPF and you establish iBGP neighborship using loopback interfaces, you can advertise the loopback networks into BGP without redistributing them into EIGRP or OSPF. This allows the loopback networks to be reachable within the BGP domain. Redistribution may be required if you want the networks learned via BGP to be reachable within the EIGRP or OSPF domain. However, if you are only concerned about the reachability of loopback networks within the BGP domain, you can advertise them directly into BGP without redistribution.

Regarding the use of loopback interfaces in BGP, it is a common practice to establish neighborship using loopback interfaces instead of physical interfaces. Using loopback interfaces provides several benefits, including increased stability and flexibility. By using loopback interfaces, you can ensure that BGP neighborship remains intact even if a physical interface goes down. It also allows for easy migration or reassignment of physical interfaces without affecting BGP neighborship.

[M.Sultan]

For example :  R1 has Interfaces range fastether 0/1-4 it mean it has four physical interfaces which all four added into OSPF routing domain as underlaying protocol, again we have interface loopback 1 established neighborship on iBGP.

Do we need to add add networks of physical interfaces on iBGP ?

Neighbour router only shows Loopback interface and one advertised network.

Also here for loopback interface we do not need redistribute, what about physical interfaces ?

[MHM Cisco World]

Do we need to add add networks of physical interfaces on iBGP ? NO need' since ospf is advertise these subnet in underlaying layer no need to add network to ibgp

Also here for loopback interface we do not need redistribute, what about physical interfaces ? If you use LO as update source of bgp then sure you need to redistrubte LO into ospf (not bgp' it no sense the LO using to establish bgp between two router using LO advertise by bgp itself)

[M02@rt37]

@M.Sultan,

If you have established iBGP neighborship on the loopback interface and want to advertise the networks learned via OSPF on the physical interfaces to other iBGP peers, you will indeed need to redistribute those networks into iBGP.

[MHM Cisco World]

why I add interface subnet to ibgp if we advertise by OSPF, that wrong 
his original post is ask about issue happened when we do that. 

[paul driver]

M02@rt37 

what wrong with a network statement?

[M02@rt37]

Hello @paul driver,

Nothing wrong with net. statement.

@MHM Cisco World, I understand that now he wants to know if it is possible to announce ospf networks learned from physical interfaces into iBGP based on loopbacks interfaces.

[paul driver]

hello M02@rt37 

scratch that last statement- if you need to advertise ospf from the lan into bgp then igp redistribution is the way to go -so apologies 

[M.Sultan]

Yes , do i need them to advertise or no ? i no why i can't ping the physical interfaces of ospf underlaying protocol from an eBGP router ?

[M02@rt37]

Hello @M.Sultan,

to sum up:

you have iBGP peering established on your topology, based on Loopbacks routed via OSPF.

What do you want to advertise ? OSPF' learned subnets ?

[M.Sultan]

Yes, But can I reach to the networks of ospf domain through eBGP peer from another AS however advertised loopback interfaces on eBGP ? 

I am confused !!!

[MHM Cisco World]

the OSPF (underlay) run ospf and exchange LO0 (using by iBGP as update source) between R1 and R2 
after we establish the iBGP (Overlay) we can now advertise the LO10 through iBGP ONLY