[c-nsp] BGP re-announcement question

Arie Vayner (avayner) avayner at cisco.com
Mon Jul 29 19:06:33 EDT 2013

The best route is through your upstream (I guess), so you are not advertising it back...
You could increase the local-pref for routes you receive from your customers as compared to routes you receive from your upstreams. In this way you would always prefer the local path to your customer (not sure they would like you to do that...)

You could also allow them to control it from their end by implementing a set of communities they can signal to you to change the way local-pref is applied on your end of the link. Something like this:


-----Original Message-----
From: cisco-nsp [mailto:cisco-nsp-bounces at puck.nether.net] On Behalf Of Adam Greene
Sent: Monday, July 29, 2013 15:53 PM
To: cisco-nsp at puck.nether.net
Subject: [c-nsp] BGP re-announcement question



We are receiving a heavily prepended route announcement from a customer and are trying to re-announce it to our upstream provider, without success. 


We learn the route on router A via eBGP. Router A announces the route to Router B via iBGP. Router B is then supposed to announce it to Router C, our upstream provider, via eBGP, but it is not. 


Here is what Router B is seeing:


7206VXR#sh ip bgp

BGP routing table entry for, version 262804866

Paths: (2 available, best #2, table Default-IP-Routing-Table)


  27241 27241 27241 27241 27241 27241, (received & used) (metric 1) from (

      Origin IGP, metric 300, localpref 100, valid, internal

  12271 7843 3356 46887 27241, (received & used) from (

      Origin IGP, localpref 100, valid, external, best

      Community: 7843:2161 7843:2303 7843:2313


As you can see, the eBGP route Router B is receiving from Router C
(upstream) is considered the best route. Is that why we are not reannouncing it back upstream? 


Prefix-lists and route-maps have all been checked, including regular expressions; we allow out ^27241(_27241)*$ and a "sh ip bgp regexp ^27241(_27241)*$" provides this output: 


   Network          Next Hop            Metric LocPrf Weight Path

* i198.11.15.0            300    100      0 27241 27241
27241 27241 27241 27241 i


I do wonder a bit about the "metric 300" (I think my customer might be trying to weight things) but I'm not sure that's coming into play here or not.


This was working fine recently, with no changes on our end, but the customer was announcing without any prepends, and I'm not sure if the origin was "?"
before, or "i" like it is now. 




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