Redistribution between RIP2 and EIGRP

[Filip Knezevic]

Hello, I'm also studying for my CCNA and Im doing some labs in Packet tracer. I have RIP2 network and EIGRP network, all computers in those networks can ping each other, so the routing is good. Now I wanted to connect this two networks so I created router in the middle (CORE router) that should redistribute traffic between networks. I enabled RIp2 and EIGRP on Core router and I followed instructions for redistribution, so I typed this commands:

router eigrp 1

redistribute rip metric 1024 0 255 1 1500

network 10.0.0.0

no auto-summary

!

router rip

redistribute eigrp 1 metric 1

network 10.0.0.0

network 172.30.0.0

network 192.168.13.0

no auto-summary

!

All computers in RIP 2 network can see Core router, all computers in EIGRP network can see Core router, but they cant see each other. 

When I type show ip route it looks like EIGRP network can see RIP2 network, but not vice versa.

So it looks like EIGRP router and Core router learned all routes, but RIP2 router didnt. Obviously I messed something with redistribution.

Please help. Thank you in advance.

In attachment you can see how my lab looks like.

[Mark Malone]

Hey

metrics look wrong in eigrp should be redistribute rip 10000 100 255 1 1500 as the default but you can take the metrics of whatever type of interface your using either and I would increase the metric in the rip redistribute as well redistribute eigrp 1 metric 10, see if that helps

[Filip Knezevic]

Thank you for fast response, Mark. I'll give it a try now :)

[Filip Knezevic]

Hi, Peter, thank you for joining the discussion.

If metrics are ok, then obviously something else is a problem. Ill try to change delay later and give it a try. 
Regarding metrics for RIP I also understood, reading online, that it is better if the RIP metrics are set to lower value than to higher value, because of that infinite loop after 16. 

Still, I have to point that I'm still learning, and I'm using this concepts without expert understanding of this topics, I'm currently really just trying to grasp all this information.

On Core router there indeed are two serial interfaces connecting routers R3 and R4. Maybe I'm making a mistake, but when I was configuring rip interface on Core router, I did network command for all networks connected to R3:

Router#sh ip int br

Interface IP-Address OK? Method Status Protocol

FastEthernet0/0 172.30.100.1 YES manual up up

FastEthernet0/1 192.168.13.1 YES manual up up

Serial0/0/0 10.0.5.1 YES manual up up

Serial0/0/1 209.165.200.234 YES manual up up

Loopback0 172.30.200.1 YES manual up up

Vlan1 unassigned YES unset administratively down down

So, i guess that in RIP interface i typed something like this: 

router (config-router)#

ne 172.30.100.0

ne 192.168.13.0

ne 10.0.5.0

ne 209.165.200.232

no auto sum

Now, correct me if I'm wrong, when I execute no auto sum, routers can make difference between, say, networks  ne 209.165.200.232/30 and 209.165.200.228/30. In both networks (RIP2 and EIGRP), I have 10.x.x.x networks, thus that overlap that you noticed (at least I think so). If you need screenshots of any settings, I'll be happy to provide them. 

P. S. I couldn't figure how to reply to the whole thread, so I can reply only to individual posts, don't know why.

[Filip Knezevic]

Peter, I gave it a try with those metrics, but it still isn't working. Confusing thing is that it looks like router auto corrects my settings. Although I typed 

redistribute rip metric 100000 1 255 1 1500, in show running config it is 

router eigrp 1

redistribute rip metric 100000 0 255 1 1500

network 10.0.0.0

no auto-summary

!

[Peter Paluch]

Mark,

I beg to differ - please allow me to explain.

For a single-point redistribution, there are no "right" metrics apart from a single consideration: Set them to a sufficiently low value so that they can grow.

Filip's EIGRP metrics of Bw=1024, Dly=0, Rly=255, Load=1, MTU=1500 are perfectly correct. Perhaps the Delay of 0 is unusual and Packet Tracer might not be working correctly with it, but on real-life routers, this configuration will be accepted and will be working. Because of relatively low Bw, the resulting metric will be high but not that high to be considered an infinity. I am usually recommending EIGRP redistribution metrics of Bw=100000, Dly=1, Rly=255, Load=1, MTU=1500, but once again, these are only recommendations, not required values, and Filip's values are in every way correct.

The same goes for Filip's redistribution into RIP - the seed metric of 1 is just correct.

In addition, recommending to increase the RIP metric from 1 to 10 is not going to help, quite the contrary: The metric will be closer to the "infinity" of 16. If a route is not accepted with the metric of 1, it is not going to be accepted with a higher metric, anyway.

Let's keep in mind that this is Packet Tracer, so its behavior may differ unintentionally from real life devices. My suggestions are:

1. Restart the entire simulation
2. Change the Delay metric in the EIGRP redistribution from 0 to 1 - it is possible that Packet Tracer is having issues understanding the Delay of 0.

Filip, I have one question regarding your setup: On the Core router, in the EIGRP configuration, there are three networks - one of them overlaps RIP (10.0.0.0), and then two others. What do these three networks stand for? According to the screenshot, the Core router has only two interfaces.

Best regards,
Peter

[Mark Malone]

Ok that's interesting because I was always taught to take the metric off the actual interface that is local on the central router or router performing redistribution  , if there Ethernet take it from the below as an example for your k1 metrics or if the links were serial interface use the metrics attached to that interface.

  MTU 1500 bytes, BW 1000000 Kbit, DLY 10 usec,
     reliability 255/255, txload 1/255, rxload 1/255

[Peter Paluch]

Hi Mark,

I was always taught to take the metric off the actual interface that is local on the central router or router performing redistribution

Yes, that is a best practice recommendation that makes sure that the metrics of redistributed routes are realistic and generally correspond to the scales of metrics in your network overall. Even more correct would be taking the metrics of the interface belonging to the source routing protocol to make sure that the metrics of redistributed routes resemble directly connected networks on that interface. However, as far as validity of the configuration goes, Filip's configuration was correct even though his EIGRP metrics were, say, less-than-typical. That is why I opined that changing the metrics probably won't help.

Best regards,
Peter

[Mark Malone]

Ok great thanks for that further explanation

[Filip Knezevic]

Hi, Peter, thank you for joining the discussion.

If metrics are ok, then obviously something else is a problem. Ill try to change delay later and give it a try. 
Regarding metrics for RIP I also understood, reading online, that it is better if the RIP metrics are set to lower value than to higher value, because of that infinite loop after 16. 

Still, I have to point that I'm still learning, and I'm using this concepts without expert understanding of this topics, I'm currently really just trying to grasp all this information.

On Core router there indeed are two serial interfaces connecting routers R3 and R4. Maybe I'm making a mistake, but when I was configuring rip interface on Core router, I did network command for all networks connected to R3:

Router#sh ip int br

Interface IP-Address OK? Method Status Protocol

FastEthernet0/0 172.30.100.1 YES manual up up

FastEthernet0/1 192.168.13.1 YES manual up up

Serial0/0/0 10.0.5.1 YES manual up up

Serial0/0/1 209.165.200.234 YES manual up up

Loopback0 172.30.200.1 YES manual up up

Vlan1 unassigned YES unset administratively down down

So, i guess that in RIP interface i typed something like this: 

router (config-router)#

ne 172.30.100.0

ne 192.168.13.0

ne 10.0.5.0

ne 209.165.200.232

no auto sum

Now, correct me if I'm wrong, when I execute no auto sum, routers can make difference between, say, networks  ne 209.165.200.232/30 and 209.165.200.228/30. In both networks (RIP2 and EIGRP), I have 10.x.x.x networks, thus that overlap that you noticed (at least I think so). If you need screenshots of any settings, I'll be happy to provide them. 

P. S. I couldn't figure how to reply to the whole thread, so I can reply only to individual posts, don't know why.

[Filip Knezevic]

Mark, unfortunately, that didn't work. I did do redistribute rip 10000 100 255 1 1500 command, after that I executed show run, and I get this:

router eigrp 1

redistribute rip metric 10000 10 255 1 1500

network 10.0.0.0

no auto-summary

From some reason it is set to 10 although I typed 100. 

Also, routers R3 (RIP2 router) and R4 (EIGRP router) still can't ping each other.

In attachment, I posted screenshots of show ip route commands on routers R3 and R4. R4 knows everything about Rip2 network, but it looks like R3 is clueless, so my guess is that I did something wrong in distributing eigrp in rip2.

[JDGranger · Accepted solution]

router rip
 version 2


Try this.

[Filip Knezevic]

Indeed, I forgot to turn on RIP version 2. Works like a charm now. Thank you very much.

[JDGranger]

Awesome, we have all made these mistakes believe me. Good luck with the studies!