Eigrp Convergance on C3560V2

[MANSOORQ123]

Dear Support Team

Please check the diagram for the exact topology.

if the BGP is down on Primary Link WAN Router , Core Switch (C3560V2 / C3560G running IP service Image 12.2(55) SE7) installs the route from the Second WAN Router, but when the Primary Link is back again, Routes on the core switch from primary WAN router are not automatically installed, although there is a huge metric difference between them.

to avail the primary link, we have to do clear ip bgp * on Secondary WAN Router.

This is causing a lot of issues....

how we can force/ instruct the switch to immediately install the routes from the Primary WAN Router, as soon as it is up.

Thanks in Advance..

Ahad

[daniel.dib]

There is no iBGP between the routers? Is there a reason you are not running iBGP between the Cat3560 and the routers? Which AS are you peering with towards the providers?

Daniel Dib
CCIE #37149

[MANSOORQ123]

Hello Daniel

Yes there is no IBGP between the 2 routers and that is not needed as well, because the L3 Switch has to do the route selection….

Router A (High Bw Link) is peering with AS 35819.

Router B (Low Bw Link) is peering with AS 65000.

Thanks for your time to look into the problem.

[Jon Marshall]

So when you do a "sh ip route" on the core switch it is still pointing to router with low bandwidth (note your diagram says router B has high bandwidth but your description says router A has the high bandwidth - which is it ?).

What does a "sh ip eigrp topology" show. Are the routes being received the same ie. same prefix/subnet mask.

Finally are you using route tagging for the mutual redistribution ?

Jon

[MANSOORQ123]

Hello Jon

Yes Router B is connected to High Bw Link.

Even after the Primary WAN Link comes back, Core Switch still shows that all routes learned from Router A (Low Bw Link).

Sh ip eigrp topology on Core Switch shows exactly the same routes received from Router B, but as such they are not installed in Routing Table, unless BGP is cleared on Router A(Low Bw Link).

Is there is any command or timer , which forces eigrp to recheck its topology and accordingly update Routing Table.

Thanks

[Jon Marshall]

Is there is any command or timer , which forces eigrp to recheck its topology and accordingly update Routing Table.

Shouldn''t need to because once EIGRP receives the new routes it should do that check anyway ie. is the new route a better metric than the one currently being used.

There is definitely a better metric for the new routes received ?

Presumably it is up and running properly now ? Can you pick a network being advertised into your site and post -

"sh ip bgp " from both BGP routers

"sh ip route " from both BGP routers and the 3560

"sh ip eigrp topology " from boh BGP routers and the 3560

and can you provide IPs for the BGP  interfaces - WAN (EBGP peering) and LAN (to 3560)  and also the IPs of the 2 routed ports on the 3560 ?

Jon