Problem of EIGRP

Unanswered Question
Nov 26th, 2007

Hi my seniors, I am new in networking field and preparing for lab have three router

r1,r2,r3. s0 of r1 is connected with s0 of r2. s1 of r2 is connected with s1 of r3.

r1=s0=, lo0=



r2=s0=, lo0=

r2=s1=, lo1=



r3=s1=, lo0=



network is working fine.but the lo0 of r3 is not being advertised by eigrp. I checked "sh ip route" on r1 and on r2 but this loopback was not there. when i changed the network command on r3 to without giving mask, i noted that now it is being advertised through r3 and on other router it is being shown,,,,I think the actual problem is in giving my network command, but I dont know what was the actual concern.

The other problem is about a message that I noted during my configuration. Message"neighbor, not on common subnet for s1" when i tried to configure the router with different addresses....thanks for reply....

I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 4.7 (3 ratings)
yclee_peter88 Mon, 11/26/2007 - 19:36

Hi, like you said the actual problem is in giving the network command, but you dont know what was the actual concern. Here are some of my suggestions for the CCNA exam and real life production network :

1) When you put in the EIGRP network statement, the router is looking for network range from to which is non-existence. The correct statement should be " " which will cover IP range from to

2) When you changed the network command on r3 to without giving mask, the router will auto summarize it to classful network of /16, before announcing it to other neighbors, the router make sure there is at least one subnet present in the routing table, which is in this case.

3) When configuring IP addressing for point-to-point serial connection, try to use the smallest subnet possible to save IP address space. Instead of using /24 (, it is recommend to use /30 as you only need two IP addresses for any point-to-point link.

4)When configuring for loopback address for router-id purpose, it is advisable to use a host mask of or /32, again to save IP address space. I can see you are using /8, /20, /28 for loopback which I strongly not recommended.

5) You can actually put your loopback in the EIGRP network statement this way :

Example - network

Just my 2cents view of your problem. Hope the above will clear your doubts.


Peter Lee


Kevin Dorrell Tue, 11/27/2007 - 00:08


1) The correct statement should be network The network statement takes a wildcard mask, not a subnet mask.

Similarly in point (5)

Kevin Dorrell


tahir1234 Tue, 11/27/2007 - 07:19

Hi Peter, Thanks for the reply. Got everything clear except one 1) why use mask in loopback, I mean what route will consider it ? 2) why usually we dont use subnet mask in network statement. This is not big for you but I am novice, and when one use actual routers ,we come to know about theory..


yclee_peter88 Tue, 11/27/2007 - 09:09

Hi, thanks for the amendments, Kevin. So similar to OSPF when configuring network command in EIGRP with the mask option, be sure to enter wildcard bits instead (example: network

Hi Tahir,

1) the ( in wildcard format) is called a "host mask" representing a subnet with one host(or one IP address). This host mask is usually use to configure a loopback address for the purpose of router ID. Because a router ID requires only one IP address, a host mask achieves just that.

In your home lab scenario, for example r2 will just pick the highest IP address in loopback lo1= as router ID by default. So u don't really need a host mask. But in a production network, this is the usual practice (using a host mask for the router ID for better network management).

In case u need to know more, what happens in real life network, we usually don't allow the router to choose its own router ID. The ID is manually configure using "router(config-router)#eigrp router-id x.x.x.x" which overrules all other default options.

2) If u don't use the mask option, eigrp will auto-summarize all the subnets to a major classful network. This is not a good idea in a discontinuous network where u need to advertise more specific subnets.

R1( <--------> R2 <--------> R3(

In the above scenario, R1 and R3 will both summarize their networks to if u leave the mask option. In addition, u also need to enter the "no auto-summary" command.

SSLIN is right, remember each interface of a router should be configured with a separate subnet or u will get "overlapping addresses" error.

Hope you find the above information helpful.


Peter Lee


tahir1234 Fri, 11/30/2007 - 18:38

Thanks Peter for this help. I have clear picture about this problem. Would you contribute your experience on "how to prepare the CCNA in a best way" Thanks again for contributing. Best regards.

CSCO10892433 Tue, 11/27/2007 - 07:14


I'm confused with this senario for two points:

1. The serial interfaces of all 3 routers are in tha same subnet: I don't think this is a good addressing scheme for two point-to-point connections. Routing protocol will unlikely work.

2. Especially in r2, s0 and s1 are in the same subnet. I don't believe a real router will allow you to do that. You will probably get a "overlapping addresses" error message and fail to assing an ip address to the second interface.

I suggest you reassigning the two point-to-point connections to two different subnets such as and, and then try your lab again.




This Discussion