Overview:
In this lab you will see one of the problems that can sometimes appear on networks with multiple routers doing redistribution
Configure the routers to match the diagram in the left margin labeled topology. You will be using the range of class C addresses between 204.1.2.0-204.1.8. Never use a static route at any time on the routers to achieve the solution.
Keep a diagram of your network including all network addresses and routing domains.
Configuration:
1. Configure all interfaces with a single network from the range given.
2. Configure the loopback interfaces on R1, R3 and R4, but not on R2.
3. Configure routing domains as shown on the topology diagram. Notice that the Serial connection between R1 and R4 needs to support both RIP and IGRP.
4. Configure redistribution between the IGRP and RIP routing domains on R1 and on R4. When you are finished make sure that all routes can ping the loopback interface on R3.
1. You can pick any one of the class C address for any of the interfaces. Be sure to use the natural mask, 255.255.255.0.
2. There is no reason no to configure a loopback interface on R2, it just won’t add anything to this lab.
3. Here you need to make sure that what ever network you assign to the serial circuit between R1 and R4, is advertised via both RIP and IGRP. This happens on real networks rarely but it simulates something that you see often. On real work networks redistribution often takes place between multiple routers connecting the same networks/subnets. This is done for redundancy, but it creates the same problem that we will see in this lab.
4. Here is where the fun beings. First of all, make sure that the address of R3’s loopback interface is visible on R2. If it is not, you are not redistributing RIP back into IGRP. This is likely due to a problem that you will see time and time again with IGRP, EIGRP and RIP. When routes are redistributed into these protocols from other routing protocols, the metric that the other protocol use is almost always different. That leaves IGRP, EIGRP and RIP confused about what metric to use.
If you do not see the two networks attached to R2 on R3 you are not redistributing IGRP into RIP correctly. The problem has the same root as the RIP into IGRP problem above.
If you have RIP and IGRP redistributing correctly you will see that after a few minutes R1, R2 and R3 can no longer ping the loopback interface of R3. If you check the routing table on R1, you will see that to get to R3’s loopback you should go to either R2 or R4 (it will flip between the two). Neither of these is correct, however, because you should be going to R3. The reason is that the RIP route to R3’s loopback was redistributed into IGRP. Then IGRP, now aware of R3’s loopback network, began advertising a path for that network. Since IGRP has a lower administrative distance than RIP the RIP routers that originally told the IGRP domain about R3’s loopback, start using the IGRP route to R3’s loopback, instead of the RIP learned route. Trouble is the IGRP domain never actually touches R3 or it’s loopback so the traffic gets stuck in the IGRP routing domain. Some how you need to filter the routing updates so that the routers do not learn IGRP routes that they should be learning from other protocol. Be careful when you run into this problem that you filter routes coming in that should not be, not routes going out. If you filter routes going out, you may be filtering the only advertisement for that network.
Why don’t you see the same kind of problem with the IGRP learned routes being stuck in the RIP domain?…….
It’s because RIP has a higher administrative distance (lower preference by the router). When you see this kind of problem, the highest preference (lowest administrative distance) protocol will always be the one that the other network traffic gets “stuck” in. Some people refer to this is as a black hole, because all traffic on the network get sucked into the preferred routing domain and never leaves again.
Download The Advanced Redistribution Lab: {filelink=9}
January 5th, 2011 at 8:33 pm
There aren’t many web sites with data like this man! Bookmarked!