Re: MPLS VPN Route Target Rewrite

Tom,
just to easy things for you, and not to leave a mess in the list archives!

First, as I don't have the topology, what I believe it is:
(please use a monospaced type like courier for the ASCII "art")

VPN_A(100:1) VPN_A(100:1)
                                     | <100:55*
172.16.7.0/24 (S) | 100:55> 172.16.5.0/24
155.1.67.0/24 (C) <100:1> | <100:1> 150.1.58.0/24
                                     |
                    [R6] --MP-BGP-- [R5]
                                     |
150.1.76/0/24 <100:2> | <100:2> 150.1.5.0/24
192.168.7.0/24 <100:55 |
                                     |
VPN_B(100:2) VPN_B(100:2)

How to read: R5 and R6 habe a MP-BGP adjacency and have both VPN_A
(upper networks) and VPN_B (lower networks) VRFs.
In the midle, <100:X> are RTs, with signs showing import or export
direction into vpnv4 "bag" of routes. The one with an asterisk is
controlled by an export map. Inside parenthesis are RDs.

If this diagram is wrong, the following is wrong too! (GIGO principle :)

Now, your "interesting" route was 172.16.5.0/24.
This is local to R5 in vrf VPN_A.

Because it matches the export map, it is tagged with RT 100:55 and
because it originates at R5's VPN_A VRF, it gets RD 100:1.

Goes to R6 via MP-BGP.

It is imported by VPN_B VRF (because of RT 100:55) into vpnv4 (RD 100:2)
and unicast vrf VPN_B.

So you see the route twice at vpnv4, one with RD 100:1 (default for
VPN_A_at_R6 but really the RD of VPN_A_at_R5) and one with RD:100:2 (via
VPN_B_at_R6).

Notice that even though the RD corresponds to that of VPN_A_at_R6, the
route only has RT 100:55, and that is not being imported by VPN_A_at_R6.

Hope this is clear.
-Carlos

-- 
Carlos G Mendioroz  <tron_at_huapi.ba.ar>  LW7 EQI  Argentina
Blogs and organic groups at http://www.ccie.net