[c-nsp] IOS-XR BGP Aggregate-addresses and VRF Export route-policy behaviour

James Bensley jwbensley at gmail.com
Tue Oct 4 09:10:58 EDT 2016


On 1 October 2016 at 13:51, Mailing Lists <lists at coop3r.com> wrote:
> Need a second opinion on what should be happening with VRF export
> route-policies on IOS-XR when using BGP aggregate-address prefixes...
>
> Example scenario:
>
> 3 VRF's A,B and C
>
> A and B export specific prefixes to C using an export route-policy
>
> A contains an aggregate-address in BGP
>
> BGP aggregate gets tagged and exported to C even though the route-map
> specifies the specific longer prefix. This sort of behavior is repeatable
> on ASR9K running 5.2.4 and 5.3.3 SP3 as well as XRv running 5.3.3 which my
> example below is taken from
>
> Full Configuration from the XRv test example:
>
> http://pastebin.com/z0q3TTQB
>
> Routing table and BGP table from vrf-c and
>
> http://pastebin.com/nAc2vHza
>
> You can see the aggregate tagged with the RT that should only apply to the
> specific prefixes.
>
> Have I properly misunderstood something and this is intentional behavior,
> or is this very wrong?
>
> Thanks


Can you try to swap this...

if destination in (10.1.1.1/32, 10.2.2.2/32) then

For a prefix list, I wonder if there is a problem with the matching clasue?


Cheers,
James.


More information about the cisco-nsp mailing list