From: Győri Gábor (Gabor.Gyori@lnx.hu)
Date: Wed Apr 28 2004 - 04:10:50 GMT-3
An alternative solution is to increase the metric (cost, delay/bandwidh) of redistributed rouutes in redistribution higher then the metric of link to the neighbor router in particular routing protocol.
Gabor
-----Original Message-----
From: Edwards, Andrew M [mailto:andrew.m.edwards@boeing.com]
Sent: Tuesday, April 27, 2004 10:20 PM
To: ccielab@groupstudy.com
Subject: Re: generic redistribution question
All,
In the lab examples I've done so far I think I'm reading too much into redistribution such that I might actually loose points because of "over configuration". Looking for some feedback on this...
For example,
Single router is the border router for an OSPF and EIGRP routing domain. Requirement is to redistribute OSPF to EIGRP such that all routes are visible on all routers.
The approach I generally take and make brain not swirl too much is to build route-maps such that I tag redistributed routes from one routing domain into another, and then deny those tagged routes on the redistribution to the original routing domain.
More specifically,
1. OSPF to EIGRP with a tag of 222. EIGRP to OSPF deny tag 222. 2. EIGRP to OSPF with a tag of 333. OSPF to EIGRP deny tag 333
Makes sense to me, but from an "over configuration" standpoint, would it be more correct to just simply redistribute the routes and then let AD work on my behalf where applicable?
This all stems from my attempt to build a model for redistribution... cause sometimes the redistribution gets a little too much to keep track of in my opinion.
Andy
This archive was generated by hypermail 2.1.4 : Mon May 03 2004 - 19:48:57 GMT-3