[c-nsp] Change VRF RD

Arie Vayner (avayner) avayner at cisco.com
Tue Sep 8 12:58:58 EDT 2009


Phil,

Why don't you just create a new VRF, with a new RD, then prepare all the
import/export policies (no need to change RT's). You can also prepare
any relevant PE-CE routing config in advance.
This can be done offline.

Then when you want to migrate the customer, you just change the "ip vrf
forwarding" command on the customer interface, and make the PE-CE
routing work (like wait for the routing protocols to come up).

You can do this migration per customer, and you do not have to do it per
the whole box, so you can basically control the scope of disruption very
easily.

Arie

-----Original Message-----
From: cisco-nsp-bounces at puck.nether.net
[mailto:cisco-nsp-bounces at puck.nether.net] On Behalf Of Phil Mayers
Sent: Tuesday, September 08, 2009 13:31
To: cisco-nsp at puck.nether.net
Subject: [c-nsp] Change VRF RD

All,

When we deployed our MPLS core, we made the minor mistake of setting the

RD the same on every router.

On the platform we're using (6500/sup720, 12.2(33)SXI) you don't seem to

be able to change the rd of a defined VRF; you have to "no rd" which 
promptly blows away the "router bgp / address family ipv4 vrf X" as well

as the route-target statements; and it won't let you enter a new rd for 
~60 seconds, claiming:

% Deletion of "rd" in progress; wait for it to complete

This is obviously hugely unhelpful for correcting the mistake.

Does anyone know if this is possible either without an outage or with a 
very short one? Or are we stuck with taking the outage?
_______________________________________________
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/


More information about the cisco-nsp mailing list