Re: [squid-users] 50 requests per second?

You're right, WCCP use under Squid is not well documented...but it also
is not very complicated.

If you're using Red Hat, you can grab one of my kernel packages here
which are already patched for WCCP (or wait a day and get the latest
2.4.9-37 kernel when I get it uploaded):

http://swelltech.com/support/updates/7.x/RPMS/

We'll be switching to Red Hat 7.3 in the next week or two, so 2.4.18
packages will come eventually.

Anyway, these kernels already have a patched ip_gre module, so to
configure the kernel side of things is as simple as:

modprobe ip_gre
ifconfig gre0 192.168.1.1

Then create a redirect rule, as usual, except the source for traffic can
be isolated to the gre0 interface. Note that the gre0 interface does
not need a routable IP--so give it a dummy. If you give it an IP on
your local net, and set it up using the Red Hat network scripts, it will
steal the route (since 'gre' comes after 'eth' alphabetically, the
startup scripts hit gre last and sets the route to gre0, which won't work).

Also add your router to the squid.conf wccp_router configuration directive.

Beyond that, normal transparent configuration applies.

See? Couldn't be easier. But could be better documented. I'll put it
on my to-do list (which is too long already).

Mark.H.Price@AOC.STATE.NC.US wrote:
> Yes, the Cisco 6509 will be our single point of failure. I think when you
> blow over 100k on a router and Cisco is only 20 minutes away, it shouldn't
> be a problem getting htem to fix/replace it :-) .. but, the 6509 is not my
> responsibility.
>
> WCCP would be great, but I couldn't get it to work. The wccp module for
> linux seems to be shoddy and not a clean simple thing to do. I need to be
> able to clearly document what I am doing in case I leave and someone else
> has to take over, or rebuild the box..
>
> any more suggestions would be welcomed. thanks
>
> Mark
> ------------------( Forwarded letter 1 follows )---------------------
> Date: Fri, 21 Jun 2002 20:37:39 -0500
> To: AaronS@et-n-m.com
> Cc: Mark.H.Price, squid-users@squid-cache.org
> From: joe@swelltech.com
> Subject: Re: [squid-users] 50 requests per second?
>
> Well...The Cisco doesn't have moving parts. That makes it a less likely
> point of failure than a Squid box.
>
> But since that's the case, I'm confused why Mark is asking about meshes
> and trees. Use WCCP, Mark. That's what it is for. But you don't need
> two boxes for effective failure prevention--WCCP will stop redirecting
> if the cache fails, and will instead route directly to the internet.
>
> Aaron Seelye wrote:
>
>>So your single point of failure is a Cisco 6509? :)
>>
>>Aaron
>>
>>
>>
>>>-----Original Message-----
>>>From: Mark.H.Price@AOC.STATE.NC.US
>>>[mailto:Mark.H.Price@AOC.STATE.NC.US]
>>>Sent: Friday, June 21, 2002 10:18 AM
>>>To: joe@swelltech.com
>>>Cc: squid-users@squid-cache.org
>>>Subject: Re: [squid-users] 50 requests per second?
>>>
>>>
>>>
>>>Yes, while it may not be necessary to use 2 load balanced boxes
>>>for serving that many, it would be frowned upon in my environment
>>>to have a single point of failure. Our network is focusing on a Cisco
>>>6509 we have that will be doing the load-balancing.
>>
> --
> Joe Cooper <joe@swelltech.com>
> Web caching appliances and support.
> http://www.swelltech.com
>

-- 
Joe Cooper <joe@swelltech.com>
Web caching appliances and support.
http://www.swelltech.com