From: Scott Morris (swm@emanon.com)
Date: Tue Aug 02 2005 - 21:21:44 GMT-3
Do you have a misconfiguration there? Normal route-caching works on flows.
So one flow (similar src/dst) will go the same direction all the time. CEF
pre-processes L2 next hop information so that you have ALL the info there
all the time, which makes a balancing situation take place.
So if you have misconfigured something (remember, there aren't secondary
addresses per se with IPv6, but you can have multiple "ipv6 address"
commands) So if you put one in the wrong place and thought you were
overwriting it like we do with the "ip address" command (there can be only
one primary) then you have a multiple route concept happening. Multiple
next hops would lead the router to assume multi-path operation. Only one
works, the other doesn't. Hence your skipping there.
I'm late to the conversation, so forgive me if this has been hashed out
already. But I'd do a quick "wr t" and look for extra addresses (that's
"show run" in old-guy speak!)
:)
Scott
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
gladston@br.ibm.com
Sent: Tuesday, August 02, 2005 3:36 PM
To: ccielab@groupstudy.com
Subject: Re: Re: RE: Failing Ping from Spoke to Spoke
Hi,
Disabling IPv6 CEF makes the problem disappear:
Rack2R2(config-subif)#no ipv cef
Rack2R2(config)#do sh ipv6 cef
%IPv6 CEF not running
Rack2R2(config)# ipv cef
Rack2R5#pi 2001:148:5:6::1
!!!!!
Rack2R2(config-subif)#ipv cef
Rack2R5#pi 2001:148:5:6::1
!.!.!
Success rate is 60 percent (3/5), round-trip min/avg/max = 8/9/12 ms
Strange, isn't it?
The test above is from another configuration that I did just to check if the
problem was or not related to Hub/Spoke. This is the topology:
R5-----(s0/0.235)-R2-(e0/0.26)-----R6
This archive was generated by hypermail 2.1.4 : Sun Sep 04 2005 - 17:01:18 GMT-3