Ouch!!!

In my scenario here that is a very drastic step!

Before doing so could you please confirm my belief that this stub config is keeping my 4500 from propgating routes downstream.

Thanks for your reply.

I'm unable to tell you 100% that eigrp stub is culpable, but it could.

See if the following document helps:

http://cisco.com/en/US/products/ps6599/products_white_paper0900aecd8023df6f.shtml

To work around the stuck statement, you can try 'no router eigrp xxx' and then "router eigrp xxx" with all the necessary commans. Make sure you have connectivity by some mean when doing this!!!

Chris

Do I understand your post correctly that 4500s on one side have eigrp stub and on the other side they do not?

On the side where eigrp stub is configured, are the 4500s using Sup II-plus ? I believe that it is a restriction of the Sup II-plus that it can only run eigrp with the stub implementation. This link is one of the places where I found that mentioned:

http://www.cisco.com/en/US/products/hw/switches/ps4324/products_data_sheet09186a0080197424.html

HTH

Rick

Chris,

Rick's response may have answered one part of the question about not being able to make the 4500 a non-stub EIGRP device.

As far as why the 4500 is not learning the routes we need more information to determine the reason for it. EIGRP stub would prevent the 4500 from advertising all but connected/summary routes it shouldn't prevent the 4500 from learning the routes from it's neighbor. Check the EIGRP configuration on the device that's connected to the 4500 switch to see if there's any distribute-list or routes being summarized out to the 4500 in question.

If you are still having problems it would help if you can post the configuration of all the devices and attach a diagram, if one available.

HTH

Sundar

All,

I really appreciate the replies and suggestions. I have not yet read the links which were provided but will do so soon. To answer some of the questions posed though:

The 4500's that have this issue have Sup Engine V's 10GE.

The other side of the link actually are 4948's not 4500's and that side is fine.

The 4500's themselves do not have an issue learning the remote routes. The problem is that they are apparently not advertising them to the "downstream" 2811 router, which to my understanding is what the questionable config is meant to do. They are only "apparently" advertising the locally connected subnets.

This is all a private network (dark fiber and private T1). All of the address space is in the 10.X.X.X space so I don't believe summarization is the problem but i did turn off auto-sum just in case.

There are no distribute-lists involved

That makes sense. Since the 4500 is configured as a stub it would only advertise the connected/summary routes to the 2811.

You may be able to workaround this by using a new feature called 'eigrp stub leak-map' and that would allow the stub router to leak some routes. It's relatively a new feature and if your switch supports it, you may need a new code.

There's not a whole lot of documentation available on this topic but here's one. I had tested this is my lab a while ago and this feature did work indeed!!

www.cisco.com/application/vnd.ms-powerpoint/en/us/guest/tech/tk207/c1482/cdccont_0900aecd801e4aab.ppt

HTH

Sundar