I am labbing up a an MPLS network in GNS3 and cannot figure out why BGP multipath is not working as expected. I don't even think it is a BGP issue at this point, but I am not sure. I am using all ASR9k, btw.
There are two PE routers at each site, and I am using BGP additional-paths on the route-reflectors to reflect all paths to a given VPNv4 prefix. Then the PE routers are configured with multipath ibgp 2 unequal-cost.
From site A, PE-A, vrf RED, I am trying to reach site B, vrf RED, prefix 192.168.11.0/24.
Site A, PE-A router's BGP tables shows this.
Site A, PE-A router's RIB shows this and this.
Site A, PE-A router's CEF table shows this.
When I traceroute, or attach a docker container behind the router at Site A and use MTR, I cannot get it to actually push real traffic via both site-B PE routers. It always routes via 1.1.1.2 (site B, PE-B). BGP has already decided to put both BGP routes in the RIB, and each RIB entry's next-hop is PE-A and PE-B at site B. So I would think that standard ECMP would take hold based on both BGP RIB entries, and CEF. But it doesn't seem to be.
Any help is greatly appriciated.
No comments:
Post a Comment