cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
634
Views
8
Helpful
8
Replies

Routing & subnet in EIGRP

a.kiprawih
Level 7
Level 7

Hi All,

I have many routers (> 30) that currently uses EIGRP for routing. I am using 10.10.0.0 segment (example only) , and subnets exists everywhere. Topology is hub-spoke.

So, in my router, the eigrp config look like below:

router eigrp 100

network 10.10.0.0 -----> for LAN segment

network 172.16.0.0 ---> for serial/WAN segment

no auto-summary

eigrp log-neighbor-changes

Example:

Site-A & Router A carry 10.10.1.0 - 10.10.10.0

Site-B & Router B use 10.10.12.0/24

Site-C & Router C use 10.10.20.0/24, HUB ROUTER with serial links to RouterA, B, D, E and many more....

Site-D & Router D use 10.10.50.0/24

Site-E & Router E use 10.10.51.0/24

Questions:

If my WAN link from RouterC to RouterB is down, the eigrp routing table will be updated and all other routers in the network should be are aware of this.

If I perform 'sh ip route 10.10.12.0' in RouterC, will I get "% Subnet not in table" or will it point to RouterA (since RouterA has/uses) lowest IP subnets?

The reason is , currently in RouterC, when the link to RouterB is down, I am seeing route to 10.10.12.0 is available via RouterA instead of getting "subnet not in table".

Since routing table is already updated and route to 10.10.12.0 is not longer available/reachable (deleted), what is the correct "sh ip route 10.10.12.0" results?

Is there such thing that eigrp will point the route (to 10.10.12.0/24) to RouterA since this router carry/has lower 10.10.1.0-10.10.10.0 subnets?

TQ in adv.

AK

8 Replies 8

jackyoung
Level 6
Level 6

The correct result should be "subnet unavailable" or just show the "10.0.0.0/8" that falls into the class A of 10.0.0.0/8 subnet.

I suspect the problem may be somehere use static route to point to router A as next-hop of router B. Or there is backdoor to learn this route.

Could you post the router A, B & C EIGRP config. for reference ?

Can you issue the clear ip eigrp and check the routing table again ?

Hope this helps.

I've tested the 'sh ip route x.x.x.x' where x.x.x.x is unused subnet under 10.10.0.0. The result is "% Subnet not in table".

FYI, we do not use static route, all pure via eigrp. Router config is straight forward

As for now, I can't clear the eigrp route to avoid any disruption to the production network. But I do plan to do so after office hour, but only for subnet 10.10.12.0/24.

But, is there any (strange) possibility that RouterC will always point to RouterA due to lower IP subnet being used there? It sounds impossible.

Thanks

AK

If there is one subnet w/ shorter mask that include the 10.10.12.0/24 subnet then the 10.10.12.0/24 traffic will flow to this route. But the routing table will not show the 10.10.12.0/24 route. Therefore, it should not display the router A as next-hop. As you said, it is a specific route w/ 24 mask, it is not normal. And I believe there may be some issue in router A to advertise this route.

Can you provide the router A config ?

Or it may be an IOS bug.....

Hope this helps.

Edison Ortiz
Hall of Fame
Hall of Fame

AMRIH,

You mentioned you have a hub & spoke configuration and then you mentioned about WAN links from remote site to another remote site in addition to the hub site - I believe this would be a mesh configuration.

On a mesh configuration, what you are seeing is correct. If the WAN Link between B and C goes down, there is still a connection between A and B and that's the reason you are seeing that network via A as a transit network.

EIGRP will announce its connected network as well as any networks participating in the same AS process.

HTH,

Could you please confirm it is purely hub-and-spoke design or mesh design ?

If it is mesh design, please provide the network diagram then we have the clear picture.

Hi,

Given the fact, you mentioned your network is hub-and-spoke and the hub router is routing to 10.10.12.0 subnet via RouterA, when the link to RouterB is down, it's possible that you may have configured a subnet mask on RouterA that includes 10.10.12.0/xx subnet as well. Obviously, when the more specific subnet 10.10.12.0/24 becomes unreachable, via RouterB, the less specific route via RouterA becomes the feasible EIGRP route for RouterC to get to 10.10.12.0 network. This is the normal behavior.

If you are still having problems can you post the 'show ip route 10.10.12.0' from RouterC.

HTH

Sundar

Hi All,

Thank you for helping me out. It is a router problem (possibly bugs..) - it advertises subnet that belongs to other router. We'll test it with different IOS ver.

Cheers!

AK

Please feel free to let us know the result.

Review Cisco Networking products for a $25 gift card