I'm trying to export routes from VRF.inet.0 to inet.0 using rib-groups. Routes are placed into VRF.inet.0 from another PE by MP-BGP (family inet-vpn) session. That's rib group configuration:

rib-groups {
    export-rib VRF.inet.0;
    import-rib [ VRF.inet.0 inet.0 ];

Then I apply rib-group into routing-instances protocols bgp configuration:

    instance-type vrf;
    interface em0.0;
    route-distinguisher 200:99;
    vrf-target target:200:99;
    protocols {
        bgp {
            family inet {
                unicast {
                    rib-group VRF-GRT;

But this is not work for me. I have no VRF routes into inet.0. Is that a right way to accomplish this task?


This is actually not possible using rib-groups - if you're receiving the routes from another PE, then the NRLI will be carried as type inet-vpn instead of inet-unicast which you need in order to be installed in inet.0 (even though they are ultimately installed as inet-unicast in the VRF.inet.0).

To achieve the same functionality, you might need to come at it from a different angle:

  • On the far-end PE (the one that learnt the route in the first place) leak the route (using rib-groups) into inet.0 on the far-end PE.
  • Make sure your BGP link from the far-end PE and the near-end PE is configured for both inet-vpn and inet-unicast - this will carry the route from the far-end PE to the near-end PE and it will be installed into inet.0
  • Allow next-hop resolution for inet.0 to consult your VRF as well:

set routing-options resolution rib inet.0 resolution-ribs [ VRF.inet.0 inet.0 ]

Obviously this solution has a few drawbacks, but it is probably the simplest way.

Another solution would be to configure logical-tunnel interfaces between your VRF and inet.0 on the near-end PE, form an adjacency and export the route via BGP.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.