6

We have a customer who we are advertising full routes to, and he is sending prefixes to us.

However, the customer wants his ASN to show in the as path for his /24. Right now his /24 is terminating at our ASN, and I'm not sure what configuration parameter needs to be set to allow his ASN to show in the as path.

Any thoughts?

Edit (added parts of config):

 neighbor x.x.x.x route-map in RX:xxxxx
 neighbor x.x.x.x route-map out SSSSSS

#sh route-map SSSSSS
route-map SSSSSS permit 10
 match as-path  10

ip as-path access-list 10 seq 5 permit .*

I have the customers /24 added to the upstream's prefix list. Do I not do this?

Mike Pennington
  • 29,876
  • 11
  • 78
  • 152
Network Samurai
  • 1,378
  • 13
  • 25
  • 1
    The part about his prefix showing in the path is a little unclear; however, why are you originating his prefix, when he runs BGP with you? – Mike Pennington Sep 06 '13 at 02:36
  • Hey Mike, can you elaborate a bit more as the client thinks that he should still be able to see his asn in the path. Thanks Edit - the problem is that the customers ASN does not show up in the path and the customer wants it to. – Network Samurai Sep 06 '13 at 02:39
  • Why isn't the customer announcing his route to begin with? – Mike Pennington Sep 06 '13 at 02:42
  • This is what the customer got back to me with, "But for my route, you're originating the route, instead of announcing my origination." What do you think? – Network Samurai Sep 06 '13 at 02:47
  • You probably can't use [local-as](http://www.cisco.com/en/US/docs/ios/iproute_bgp/command/reference/irg_bgp3.html#wp1107625) since that affects all prefixes you announce, and not just his. The solution in this case seems to be for him to announce his own route. Does the site where the prefix belongs have a BGP-capable router? – Mike Pennington Sep 06 '13 at 02:50
  • 3
    To me this reads like customer IS announcing the route, but @NetworkNinja is announcing it as well. Configuration would bring clarity. Typical BCP configuration at provider end should simply be prefix match in client facing route-map setting magic community, then at upstream interfaces leak out prefixes having this magic community. – ytti Sep 06 '13 at 06:26
  • Why are you originating customer's route? – Marco Marzetti Sep 06 '13 at 06:42
  • Where does the customer expect to see his /24 and is this PI or PA space? – generalnetworkerror Sep 06 '13 at 07:48
  • I added some config to my original reply. – Network Samurai Sep 06 '13 at 12:41

1 Answers1

7

I believe I found the problem. It looks like another tech had "nailed" down his /24 to our routers to null0 which was why it was terminating at our ASN. Removing the null0 route resolved the issue it seems.

YLearn
  • 27,141
  • 5
  • 59
  • 128
Network Samurai
  • 1,378
  • 13
  • 25
  • That's sometimes done at the edge of the network to keep announcing the prefix even if the edge router stops receiving it. This prevents flapping of routes but there must have been a redistribute static that picked it up into BGP instead of announcing it from the customer via BGP. – Daniel Dib Sep 06 '13 at 13:51
  • @Daniel, we do have "redistribute static" in our bgp config. Craig, it won't let me. Says I have to wait 2 days before accepting it since it's my own answer. – Network Samurai Sep 06 '13 at 13:58
  • 2
    @NetworkNinja, someone should set the localpref of the redistributed null0 route very low... that way, the customer's route is automatically chosen if it is present in the RIB. Also, you should be able to accept now. – Mike Pennington Sep 09 '13 at 07:57