From: Brian Dennis (bdennis@internetworkexpert.com)
Date: Mon Nov 24 2003 - 16:11:43 GMT-3
EIGRP uses router-IDs to protect against routing loops for external routes.
If a router receives an external route that has the same router-ID as its
own, it will reject the route.
Brian Dennis, CCIE #2210 (R&S/ISP-Dial/Security)
bdennis@internetworkexpert.com
Toll Free: 877-224-8987
Direct: 775-745-6404 (Outside the US and Canada)
Internetwork Expert, Inc.
http://www.InternetworkExpert.com
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
Andrew Moriarty
Sent: Monday, November 24, 2003 8:32 AM
To: jhays@jtan.com; Ronald.v.Dommelen@Nobel.NL; ccielab@groupstudy.com
Subject: RE: Lab questions
Cool!
I haven't come across either of those scenario's before, but they make
perfect sense. So when would I want to change the EIGRP router id?
am
>From: "Jonathan V Hays" <jhays@jtan.com>
>To: "'Andrew Moriarty'" <amgroupstudy@hotmail.com>,
><Ronald.v.Dommelen@Nobel.NL>, <ccielab@groupstudy.com>
>Subject: RE: Lab questions
>Date: Sat, 22 Nov 2003 10:33:10 -0500
>
>-----Original Message-----
>From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
>Andrew Moriarty
>Sent: Saturday, November 22, 2003 6:34 AM
>To: Ronald.v.Dommelen@Nobel.NL; ccielab@groupstudy.com
>Subject: Re: Lab questions
>
>
>A few points:
>
>BGP and OSPF need a router-id, and both support a router-id command.
>OSPF
>uses it (sometimes) to choose the DR for a link. BGP uses it
>(eventually) to
>break ties in best-path selection. Unless you are changing this
>behaviour,
>why would you want to change the router id? It does show up in various
>databases, so you might want to change it for better readability of a
>display. I wouldn't configure a router-id command unless it was required
>by
>a question; and even then I'd ask the procter.
>
>I don't think EIGRP uses router-id at all- I may be wrong on that, and
>if I
>am someone please correct me.
>
>I always use the exact address/inverse mask combination, or the smallest
>I
>can.
>
>net 172.16.0.3 0.0.0.0 area 0
>
>rather than
>
>172.16.0.3 0.0.0.3 area 0
>
>In eigrp I always advertise only one interface at a time, using the
>inverse
>mask as well. This way I know exactly what I am advertising. The
>passive
>interface command won't hurt in that situation, but it won't help
>either. I
>believe both are correct.
>
>am
>
>========
>Andrew,
>
>Here are a few points to consider.
>
>1. When to change the BGP router ID.
>
>In your CCIE studies you will eventually come across the problem that in
>some situations the BGP and OSPF router IDs must be identical.
>Specifically there is a sort of corollary to the BGP synchronization
>rule:
>
>When the IGP in a BGP AS is OSPF, then the iBGP peer's router ID and the
>OSPF ASBR's router ID must match before the iBGP peer will advertise a
>given prefix.
>
>If you are experiencing this problem one possible fix would be to change
>the BGP router ID.
>
>2. EIGRP router ID.
>
>R1(config)#router eigrp 1
>R1(config-router)#eigrp r?
>router-id
>
>R1(config-router)#eigrp r
>R1(config-router)#eigrp router-id 137.10.1.1
>R1(config-router)#
>
>3. Use of exact mask in OSPF network statement.
>
>I prefer using the interface mask. It's a good habit and I can think of
>at least one specific instance where using 0.0.0.0 as a wildcard mask
>will result in undesired OSPF behavior, namely the Type 5 LSA forwarding
>address. See the link below for details.
>
>http://www.groupstudy.com/archives/ccielab/200310/msg01043.html
>
>HTH,
>
>Jonathan
>
>
This archive was generated by hypermail 2.1.4 : Fri Dec 12 2003 - 12:29:16 GMT-3