Redistribution of static routes into EIGRP

[pkhatri]

Hi all,

I'm currently testing the feature of EIGRP whereby static routes are automatically redistributed into EIGRP if a 'network' statement covering the static route exists. Note that I'm NOT using 'redistribute static' under the EIGRP process.

On one router, running 12.2(15)T9, the static route referenced by the network statement is installed into the topology table only if it points to an interface, and not a next-hop IP address, which I believe is the correct behaviour.

On another router, running 12.1(17), the static route referenced by the network statement is installed into the topology table even if it points to a next-hop IP address (as long as the address is reachable). Is this a bug ?

Cheers,

Paresh.

[Kevin Dorrell]

When I first started reading your posting, I was surprised, because I thought that only BGP would take static routes on board in reaction to a network statement. But then when you said about taking only those statics that point to an interface, it made more sense. Effectively, EIGRP takes the directly connected routes that correspond to network statements, including those directly connected routes that are defined via a static, as well as the ip addresses of the interfaces themselves.

The behaviour you describe in you third paragraph does indeed sound like a bug ... but ... I wonder ... it points to an IP address, but is that IP address directly connected? On your 12.1(17) router, when you do a detailed show ip route a.b.c.d for the route in question, what does it look like? Is it any different to the one you are testing on your 12.2(15)T9 ?

Kevin Dorrell

Luxembourg

[pkhatri]

Thanks for your response, Kevin.

The IP next-hop address pointed to in both cases is identical - and this address is the far-side of a directly connected link.

Cheers,

Paresh.

[Kevin Dorrell]

Paresh,

OK, it sounds like a bug or a change-in-specification then. I had a look in the bug database, but I didn't find anything.

Kevin Dorrell

Luxembourg

[pkhatri]

Thanks Kevin,

What I really wanna know now is which version has the bug :-)

Paresh.