This problem has nothing to do with Route Reflection and is simply the
typical behavior for IBGP. Next_Hop attributes are not changed throughout
the AS. Your indication of the two methods of handling next hop resolution
are accurate, and which you use tends to be a point of preference. I find
that most will use the next_hop_self knob however as it does reduce the size
of the link state database, and also gives you a Next_Hop address that is
more usable to other protocols such as MPLS should you decide to use them in
the future.
Pete
*********** REPLY SEPARATOR ***********
On 8/5/2001 at 11:02 PM Howard wrote:
>Hi,
>
>There are two eBGP border routers, each configured to run as route
>reflector. One problem I noticed is that the routes from external AS
>will be passed to the RR client without the any change to nexthop by the
>border router. Thus the nexthop of this routes remain as the interface
>addr of the neighbor's BGP border router (external AS). As a result, RR
>clients are not able to reach sites that are external, as they do not
>know how to reach the IP addr (nexthop) of neighbor's BGP border router.
>
>Found two ways to resolve this problem:
>
>1. Configure passive OSPF for the interface that is facing the neighbor
>BGP router. Thus the IP addr of the neighbor's BGP border router is
>learnt thro' OSPF and pass on to the RR clients.
>
>or
>
>2. Configure nexthop self for the IP addr range being used betw the
>border router and the neighbor's BGP border router. Thus the IP addr of
>the neighbor's BGP router is learnt thro' BGP.
>
>Method 1 is simpler to configure, while method 2 is self-documenting.
>Any comments on these 2 methods? Or is there a better way to solve the
>problem? Is it a good idea to configure BGP border router as a route
>reflector in the first place? I have about 20+ BGP routers in my AS
>currently.
>
>Thanks!
>
>
>
>Nicolas McCartney wrote:
>>
>> I suppose this kind if thing depends on how many peers and BGP sessions
>the
>> BR will have configured on it. What is your definition of a BR - purely
>an
>> eBGP termination point?
>Moo Whoo wrote:
>>
>> Is there any good reason why border router should NOT be configured as
>> route reflector?
>>
>> Thanks.
Message Posted at:
http://www.groupstudy.com/form/read.php?f=7&i=15013&t=14762
--------------------------------------------------
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
