Disabling split horizon has more to do with the idea of whether R1 would
know about R2's information or not.  The successor list from their
perspective would still only be R3, so their behavior would not change.

If you did not have split horizon disabled, one spoke would not know of the
others routes.  Nor would R3 forward queries on to the same interface.

The example SHOULD be if R1 loses a route and sends a query to R2, with
split horizon working as it normally does, R2 would never receive a query
about it at all.

I don't think that R1 would ever receive a query back, but I honestly
haven't done a packet capture to try that out.  What I have seen in the past
though is that the query information is tied to successor's ID, so if it
receives a query, the router would know better than to send one back to the
exact same device.

HTH,

 
Scott Morris, CCIE4 (R&S/ISP-Dial/Security/Service Provider) #4713, JNCIE-M
#153, JNCIS-ER, CISSP, et al.
CCSI/JNCI-M/JNCI-ER
VP - Technical Training - IPexpert, Inc.
IPexpert Sr. Technical Instructor
 
A Cisco Learning Partner - We Accept Learning Credits!
 
[EMAIL PROTECTED]
 
Telephone: +1.810.326.1444 
Fax: +1.810.454.0130
http://www.ipexpert.com
 

-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Con Spathas
Sent: Thursday, October 18, 2007 10:40 PM
To: [email protected]
Subject: [OSL | CCIE_RS] EIGRP and Split Horizon...

Hi,

I've been going through this document but I'm a little curious about Queries
and the impact of Split Horizon on them:

http://www.cisco.com/warp/public/103/eigrp-toc.html#queries

In particular:

"Queries result in a split horizon only when a router receives a query or
update from the successor it is using for the destination in the query."

So this is the topology:


         |
 Subnet A|--- R1------\     |
         |             \    |----R4
                       R3---|
                       /    |
              R2------/     |
              | 
            -----

"R3 receives a query concerning SubnetA (which it reaches through R1) from
R4. If R3 does not have a successor for this destination because a link flap
or other temporary network condition, it sends a query to each of its
neighbors; in this case, R1, R2, and R4. If, however, R3 receives a query or
update (such as a metric change) from R1 for the destination SubnetA, it
does not send a query back to R1, because R1 is its successor to this
network. Instead, it only sends queries to R2 and R4."

So my question:
If you were to disable split horizon manually on the R3 interface connecting
to R1 - and R3 receives a query or an update from R1 for Subnet A - would R3
then actually send a query back to R1 (as well as R2 and R4 which it does
normally) or will disabling split horizon not change this behaviour? 

Are there any cases (other than the obvious NBMA ones) where disabling split
horizon would be required?

Thx.

Reply via email to