[c-nsp] Fw: Re: [Disarmed] Re: RIB failure : Higher admin distance

[Randy]

--- On Fri, 1/15/10, Randy <randy_94108 at yahoo.com> wrote:


From: Randy <randy_94108 at yahoo.com>
Subject: Re: [Disarmed] Re: [c-nsp] RIB failure : Higher admin distance
To: "Andy Ashley" <lists at nexus6.co.za>
Date: Friday, January 15, 2010, 8:47 PM







Hi Andy:
...I am taking a closer look at your first post and going *wait a second..*
What you are seeing is what one would expect to see in Router A site A:
 
>From the "show commands" in your first post:
 
Router A learns site B's /24 via the gre tunnel as an iBGP route with an AD of 200.(as shown int your "sh ip bgp x.x.x.x/24" in question). Router A puts this route in it's BGP route table but does not advertise this route to any eBGP peer because iBGP routes are not injected into eBGP unless "redistribute internal" is explicitly configured.
Router A also learns site B's /24 via the private link as an OSPF route with an AD of 110(as shown in your "sh ip route x.x.x.x/24) and puts the route learned via ospf in this IP routing table and FIB since it has a better AD : 110 as opposed to 200.
As a result, the  same-/24  learned via iBGP that is in A's BGP route table; for obvious reasons suffers a RIB-failure because the same-route learned by A via OSPF with a better AD  is already installed in A's ip route tabel and FIB.
 
Having explained the *normal-behavior* you are seeing in router A, my question is:
 
1) Are you trying to announce site B's /24 from site A to your upstreams
OR
2) You are trying to announce your site-B /24 *from site B and that is failing.
 
If your are trying to announce site B's /24 from site A to it's upstreams you already have the "answer" to make that work! (deploy a lot of outbound filters before you redistribute iBGP into eBGP)
 
If on the otherhand siteB's /24 is not being announced *By-SiteB* to it's eBGP peer, I would have to look at the config in site B's rtr.
 
Regards,
./Randy