Dave,
Thanks for the attachment and response. The I_SEE_YOU after the
REDIRECT_ASSIGN is crucial as that would confirm if the router has
indeed complied with the request by the lead cache in REDIRECT_ASSIGN.
So if you get a chance to look at that, let me know or send me a
longer trace of the WCCP packet capture - mainly to understand what's
in the WCCP packets that is tickling the router to evenly load balance
rather than load distribute according to the weights specified.
Secondly did you see the data traffic being redirected across the two
caches and what was the wccp status output ("sh ip wccp XX detail" )
at the router? Thanks again.
On 5/30/07, Dave <dcas@tpg.com.au> wrote:
> It is on a cisco 12.4(4t3 advanced enterprise)
>
> I have emailed you a packet trace.
>
> Thank you.
>
> varadha tech wrote:
> > Dave,
> > Is it possible for you to send a packet trace of the problem you see?
> > What is the router/IOS version that is used? I see a different problem
> > wherein the bucket assignment order is reversed at the router. Thanks.
> >
> > On 5/29/07, Dave <dcas@tpg.com.au> wrote:
> >> Hi,
> >>
> >> I have just upgraded to running WCCPv2 on squid 2.6 stable12. Everything
> >> is working apart from load being able to assign the weight of each proxy
> >> to the routers.
> >>
> >> In the squid.conf I have the following
> >>
> >> wccp2_forwarding_method 1
> >> wccp2_return_method 1
> >> wccp2_service standard 0
> >> wccp2_assignment_method 1
> >>
> >> On one server - wccp2_weight 200
> >> On the other - wccp2_weight 100
> >>
> >> But inside the router it still show 50% hash assignment.
> >>
> >> If I examine HERE_I_AM and I_SEE_YOU packets I can see the correct
> >> weight is being sent to the server as well.
> >>
> >> Any ideas ?
> >>
> >> Thank you
> >>
> >>
> >
> >
>
Received on Thu May 31 2007 - 09:36:39 MDT
This archive was generated by hypermail pre-2.1.9 : Fri Jun 01 2007 - 12:00:05 MDT