From: Simon Grace (SimonG@pcsystems.gr)
Date: Wed Oct 17 2007 - 06:17:17 ART
How would prefix lists make a difference?
I just tried it on a router and same result.
Route-map prefix perm 10
Match ip address prefix 10 20
This all makes sense though, think about when you use a route-map to
redistribute connected interfaces.
It will list all the interfaces in one line and that works fine.
Now....how do you make it work so that BOTH routes HAVE to be preseant
for it to advertise the default.
I've tried two permit statements in the route map, one matching one
prefix and one permit statement matchin the other.
Same thing, as long as one of the routes is active the default-route
will be advertised.
So, I've tried it with ACL's PREFIX-list, both match statements being on
the same line under one permit statement of the route map and two permit
statements each one matching one of the acl's or prefix lists.
I'm having a good think about this, if anyone has some other ideas it
would be good to hear them!
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
Farhan Anwar
Sent: Tuesday, October 16, 2007 4:03 PM
To: Koen Zeilstra
Cc: ccielab@groupstudy.com
Subject: Re: conditional default advertising in BGP with multiple routes
to monitor
use prefix-lists instead of ACLs.
On 10/16/07, Koen Zeilstra <koen@koenzeilstra.com> wrote:
>
> Hi Group,
>
> Please observe the following (I hope) interesting task.
>
> Advertise a default route to a BGP neighbor conditionally. Watch two
> routes in your routing table. When both routes are present the default
> route should be announced to the neighbor.
>
> I was thinking towards the following direction.
>
> neighbor 10.0.0.1 default-originate route-map checkroutes
> !
> route-map checkroutes
> match ip address 10
> match ip address 20
> !
> access-list 10 permit 1.1.1.1
> access-list 20 permit 2.2.2.2
> !
>
>
> However this route-map results in:
>
> route-map checkroutes
> match ip addres 10 20
> !
>
> Which is an "OR" function afterall and not the "AND" we are looking
for.
>
> Any bright ideas on this particular assignment?
>
> thanks in advance,
>
> Koen
>
>
This archive was generated by hypermail 2.1.4 : Fri Nov 16 2007 - 13:11:15 ART