From: Jason1 (jason1@xxxxxxxxxx)
Date: Sun May 06 2001 - 12:54:57 GMT-3
Actually in looking at it again. I do see a problem.
A route-reflector client can peer with a route-reflector in different
cluster (in the senses that it will work) but it is against the basic rules
and it will result in duplicate copies of the same route since the same
route is received BUT the cluster-ID is different from the local cluster-id
and as such will be accepted.
----- Original Message -----
From: "Jason1" <jason1@v-labs.net>
To: <ccielab@groupstudy.com>
Sent: Sunday, May 06, 2001 1:05 PM
Subject: Re: Routing TCP/IP Volume II, Jeff Doyle -- Possible mistake?
> I do not see a problem here.
>
> 1. A route reflector client cannot peer with any other iBGP router outside
> of it's cluster , it can however peer with any route-reflector or eBGP
> router.
>
> 2. A route-reflector client can peer with a route-reflector in different
> cluster because it doesn't have to have any knowledge of the cluster it
> belongs to. Part of route-reflector configuration is that the client is
> config as per normal. Since there exists a cluster-ID for each of the
> cluster, there shouldn't be any issue with "belonging" to multiple
cluster.
>
>
> ----- Original Message -----
> From: <Ccieyet2b@aol.com>
> To: <myccie@hotmail.com>; <ccielab@groupstudy.com>
> Sent: Sunday, May 06, 2001 8:45 AM
> Subject: Re: Routing TCP/IP Volume II, Jeff Doyle -- Possible mistake?
>
>
> > Hi all,
> >
> > Like yourselves, I was waiting with great anticipation for this book to
> > become available and finally last week I picked it up. And, so far, I'm
> very
> > pleased.
> >
> > However, I think I may have come across a significant error (or, at
least,
> an
> > inconsistency). I don't have a practice lab yet so I can't test this,
but
> > here's the possible error.
> >
> > In chapter 3, in the discussion of Router Reflectors, Jeff says, (I'm
> > paraphrasing), "A route reflector client can't peer with any internal
BGP
> > router outside of it's cluster" which makes alot of sense. But, then,
a
> > little later, he goes on to say that a route reflector client can peer
> with
> > multiple route reflectors and the route reflectors do not have to be
part
> of
> > the same cluster.
> >
> > This implies that a route reflector client can belong to two different
> > clusters simultaneously. Is this possible? If so, is this useful?
> Sounds
> > to me that even if this is possible, it something that should be
avoided.
> > What do you think?
> >
> > Jim
> > **Please read:http://www.groupstudy.com/list/posting.html
> **Please read:http://www.groupstudy.com/list/posting.html
**Please read:http://www.groupstudy.com/list/posting.html
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 10:30:35 GMT-3