> >This could be caused by BGP's automatic RT filter, which is not
examining
> the import route-maps when deciding whether or not to accept a vpnv4
iBGP
> update. So if there is no other VRF on PE#2 with "route-target import
1:222"
> or ".. 1:333", you need to enable
> -> by disabling "no bgp default route-target filter" the result shown
is
> PE#1 and PE#2 will directly install the route for each VRF match RD
value
> not by matching the route-target, because it ignore any route-target
> configured.
> please look for "Output#1" which is the same at PE#1 and PE#2. in this
> scenario I also tried using import map which is no impact. i just
importing
> based on same RD regarding route-target.

Hmm, I just tried this in the lab with an import-map, and it doesn't
work either. the routes matching the extcomm-list are not imported. Not
sure if this is a bug or lack of support of this in the import map. feel
free to work with TAC on this.

> >You still want to consider re-designing this, using distinct RTs to
> achieve this policy.
> -> Yes, we are planning too, but at the moment I am dealing with such
a mess
> :(
> 
> we are having a lot of PEs with multiple route-target, if cisco device
is
> importing routes based on single route-target then, there will be mess
> within routes received in Cisco device
> I also thinking about another workaround in case Cisco device unable
to
> match multiple route-target by filtering received route within vrf
based on
> prefix-list <but I haven't tried this>

prefix-list might work. There could be another hack/workaround:

ip vrf ABC
 rd 1:111
 route-target import 1:444 ! (define a new RT)
!
router bgp ..
 address-family vpnv4
  neighbor <your RR> route-map add-rt in

!
ip extcommunity-list 101 permit RT:1:222_RT:1:333_
!
route-map add-rt permit 10
 match extcommunity 101
 set extcommunity rt  1:444 additive
!
route-map add-rt permit 20
!

so you add the RT 1:444 to the vpnv4 updates carrying both 1:222 and
1:333, and then import this 

ugly, but working..

        oli

_______________________________________________
cisco-nsp mailing list  [email protected]
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/

Reply via email to