[c-nsp] Preventing IS-IS adjacencies over 1Q trunk

[Oliver Boehmer (oboehmer)]

Bruce Pinsky <> wrote on Thursday, August 16, 2007 10:45 PM:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Justin Shore wrote:
>> I have a situation with IS-IS that's perplexing me.  I have a pair of
>> core routers connected via an etherchannel.  The etherchannel is
>> also a 1Q trunk.  The core routers have both routed connections to
>> other routers as well as common VLANs w/ HSRP for non-router
>> devices.  My intention was to build and IS-IS adjacency over the
>> native VLAN on the trunk.  However I noticed that both routers are
>> building adjacencies over 3 different VLANs that are permitted over
>> that trunk.  All 3 SVIs have IS-IS enabled on them. 
>>
>> I'm assuming that I probably shouldn't have IS-IS enabled on the SVIs
>> but is redistributing connected interfaces that appropriate solution?
>> I'm trying to eliminate as much redistribution as possible.  Is
>> there a better way to handle this?  I do need L2 between the chassis
>> for the non-routed links. 

passive-interface VlanX 
passive-interface VlanY

But nothing necessarily wrong with redistributing the connected
(possibly along with a route-map filter), both options would produce
basically the same result.

> 
> If your IOS version supports it, you could use the
> "advertise-passive-only" option in IS-IS.

Bruce, what do you mean? This reduces the ISIS IP prefixes to a minimum,
but not sure if it addresses Justin's goal..

	oli