RE: conditional default advertising in BGP with multiple routes

From: Balmik Soin (balmik@staff.iinet.net.au)
Date: Wed Oct 17 2007 - 23:15:43 ART


Unless I'm reading it wrong, it looks like the config is wrong, which is
causing confusion.

Your Loopback2 is 2.2.2.0/24, but access-list 10 is matching 2.0.0.0

Thus, the conditions for access-list 10 are never met. Only access-list
20 is met, and when you shutdown loopback20, it no longer is, causing
the route to be withdrawn.

If you want to test this, either change Loopback2 to be a 2.0.0.0
subnet, change access-list 10 to match 2.2.2.0, or just try shutting
down Loopback2 instead of Loopback20 and see what happens.

 - Balmik.

> -----Original Message-----
> From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf
Of
> Bob Sinclair
> Sent: Thursday, 18 October 2007 5:42 AM
> To: Saul Arjona
> Cc: ccielab@groupstudy.com
> Subject: Re: conditional default advertising in BGP with multiple
routes
> to monitor
>
> Saul Arjona wrote:
> >> Bob,
> >>
> >
> >
> >
> > I labbed it up and it is not working as a functional AND but as a
> functional
> > OR as Koen mentioned. So as long as one of the routes is still in
the
> > routing table we will still advertise the default route, it will
stop
> > advertising it only if both routes are out of the routing table.
> >
> > regards,
> > -Saul
> >
> > You might want to lab this up. When I do a conditional default
using
> >
> >> RIP or BGP the router in fact implements the match as an AND
function.
> >> In other words, BOTH must be in the table. Here is an example
with
> BGP:
> >>
> >> interface Loopback2
> >> ip address 2.2.2.2 255.255.255.0
> >> !
> >> interface Loopback20
> >> ip address 20.0.0.1 255.0.0.0
> >> shutdown
> >> router bgp 200
> >> no synchronization
> >> bgp log-neighbor-changes
> >> neighbor 172.16.25.5 remote-as 500
> >> neighbor 172.16.25.5 default-originate route-map CONDITIONAL
> >> no auto-summary
> >> access-list 10 permit 2.0.0.0
> >> access-list 20 permit 20.0.0.0
> >> !
> >> route-map CONDITIONAL permit 10
> >> match ip address 10 20
> >>
> >>
> Hi Saul,
>
> It for this reason that I always tell students "believe none of what
you
> hear and only half of what you see". The result of any test will
> depend on the complete context and exact procedures. Above, I have
> shown you what I consider to be the relevant config on the sending
> router. Below you will see the debug of what happens when ONLY ONE
of
> the loopbacks is shut:
>
> R2#sh ip int brief | excl unass
> Interface IP-Address OK? Method
> Status Protocol
> Serial0/0 172.16.123.2 YES NVRAM administratively
> down down
> Serial1/0 172.16.25.2 YES NVRAM
> up up
> Loopback2 2.2.2.2 YES manual
> up up
> Loopback20 20.0.0.1 YES manual
> up up
>
>
> R2(config)#int loop20
> R2(config-if)#shut
> R2(config-if)#
> *Oct 16 04:21:10.321: BGP(0): 172.16.25.5 send unreachable 0.0.0.0/0
> <<<<<<<NOTE
> *Oct 16 04:21:10.321: BGP(0): 172.16.25.5 enqueued default-originate
> update
> R2(config-if)#
> *Oct 16 04:21:12.317: %LINK-5-CHANGED: Interface Loopback20, changed
> state to administratively down
> *Oct 16 04:21:13.317: %LINEPROTO-5-UPDOWN: Line protocol on Interface
> Loopback20, changed state to down
> R2(config-if)#no shut
> R2(config-if)#
> *Oct 16 04:21:17.413: BGP(0): 172.16.25.5 enqueued default-originate
> update <<<<< NOTE
> R2(config-if)#
> *Oct 16 04:21:19.405: %LINK-3-UPDOWN: Interface Loopback20, changed
> state to up
> *Oct 16 04:21:20.405: %LINEPROTO-5-UPDOWN: Line protocol on Interface
> Loopback20, changed state to up
> R2(config-if)#
>
>
> It sure looks to me like when I shut ONE of the loopbacks, withdrawing
> ONLY ONE of the required routes, BGP withdraws the default. And when
I
> no-shut that interface it is re-advertised. To me, this pretty
clearly
> says that the router is implementing the route-map as an AND.
>
> What am I missing??
>
> Maybe you could share some of the context that led to your conclusion.
>
> --
>
>
> Bob Sinclair CCIE 10427 CCSI 30427
> www.netmasterclass.net
>
>



This archive was generated by hypermail 2.1.4 : Fri Nov 16 2007 - 13:11:15 ART