[c-nsp] vs route leaking

Saku Ytti saku at ytti.fi
Sat Jun 8 11:31:51 EDT 2024


On Sat, 8 Jun 2024 at 18:26, Arne Larsen via cisco-nsp
<cisco-nsp at puck.nether.net> wrote:

> Yes, it'd with route-target I'm trying to get it to work, and what I'm
> trying to get rid off is the default route from the IOT vrf to be
> imported into the SHARED vrf.

Ok so the problem is not sharing routes between VRF, problem is
sharing selectively routes between VRF?

In the example the problem is that VRF_SHARED_SERVICE gets default
route from VN_IOT.

You could accomplish this two ways

a) VRF_SHARED_SERVICE has import policy, which drops the default route
for  64515:136
b) VN_IOT has export policy, which doesn't set 64515:95 on default route


I think a) is more robust, you'd probably just deny importing any
default route at all, if you know you're going to have the 64515:95
default route you want. So no matter what happens in the other VRFs,
you'd never end up importing their default.

Like

vrf definition VRF_SHARED_SERVICE
  address-family ipv4
      import map FOO

route-map FOO deny 100
   match ip address prefix-list DEFAULT
route-map FOO permit 200


>
> Here are the vrf definition.:
>
>
> vrf definition VRF_SHARED_SERVICE
>   rd 192.168.101.110:95
>   !
>   address-family ipv4
>    route-target export 64515:95
>    route-target import 64515:95
>    route-target import 64515:10
>    route-target import 64515:136
>    route-target import 64515:112
>    route-target import 64515:101
>   exit-address-family
>
>
>
> vrf definition VN_IOT
>   rd 192.168.101.110:136
>   !
>   address-family ipv4
>    route-target export 64515:136
>    route-target import 64515:136
>    route-target import 64515:95
>   exit-address-family
>
>
> /Arne
>
>
>
> On 08/06/2024 12.25, James Bensley wrote:
> > Hi Arne,
> >
> > The normal way to do this is with route targets but you didn't mention route targets in your email. Are you importing the export RTs from VRF1 and VRF2 in to VRF3?
> >
> > You also mentioned route-maps. Are you already importing the export RTs and trying to filter which routes are imported to only be the default route?
> >
> > You didn't post any config, it always helps people to help you if you can show what you have tried already.
> >
> > Cheers,
> > James.
> >
> >
> >
> > -------- Ursprüngliche Nachricht --------
> > Am 08.06.24 08:04 um Arne Larsen via cisco-nsp schrieb <cisco-nsp at puck.nether.net>:
> >
> >>   Hi all
> >>
> >>   I’m struggling with an 9606 Cisco router and route leaking between vrf’s.
> >>
> >>   I have 2 vrf’s with a default route that needs to imported into a 3.
> >>
> >>   The default route from the one vrf’s is direct connected on the box,
> >>   andthe other is via mBGP.
> >>
> >>   I’ve tried several forms for import maps base on community, prefix, acl
> >>   and so on, but I always ends up with pulling my legs.
> >>
> >>   The 3 vrf is for shared services, so I import more the the 2 vrf’s with
> >>   the default route.
> >>
> >>   Can someone give me a hint how to get this to work.
> >>
> >>   The 2 vrf’s with the def route has community xxxxx:112 and xxxxx:114.
> >>   I need to import all other routes from all other vrf’s including the 2
> >>   with the def route.
> >>
> >>   Hope someone can help me out here
> >>
> >>   Regards Arne
> >>   _______________________________________________
> >>   cisco-nsp mailing list  cisco-nsp at puck.nether.net
> >>   https://puck.nether.net/mailman/listinfo/cisco-nsp
> >>   archive at http://puck.nether.net/pipermail/cisco-nsp/
> >>
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/



-- 
  ++ytti


More information about the cisco-nsp mailing list