Fwd: [squid-users] result codes and status code

i've asked to be removed countless times. here's another message that
i didn't want. it's really not that hard to remove somebody is it...

This is a forwarded message
From: Henrik Nordstrom <hno@squid-cache.org>
To: Werner.Rost@zf.com
Date: Friday, January 30, 2004, 4:13:02 AM
Subject: [squid-users] result codes and status code

===8<==============Original message text===============
On Thu, 29 Jan 2004 Werner.Rost@zf.com wrote:

> for accounting purpose I need some further information.
>
> A local proxy (with redirector squidguard) uses a parent proxy. I need all
> data records form the local access.log which are not seen by the parent
> proxy.
>
> I guess I must examine result codes and status codes. Records with following
> codes are local and do not reach the parent proxy:
>
> TCP_HIT
> TCP_IMS_HIT
> TCP_MEM_HIT
> TCP_DENIED/407 Proxy Authentication Required
> TCP_DENIED/403 Forbidden ( by squidguard )
> TCP_DENIED/400 Bad Request

All TCP_DENIED/XXX are local regardless of the status code. There is also
a couple of TCP_MISS/.. codes in case of errors depending on your Squid
version but I have tried to clean these up to give TCP_DENIED in current
2.5.STABLE.

A better criteria for determininig if the request is serviced locally or
via the network is to look at the hierarchy field.

But then you also have TCP_REFRESH_HIT which is a little complex to
account.. the object data is serviced locally but there was also a IMS
query over the network to verify if the object is up to date. To promote
the effect of having a cache I would recommend this to be accounted as
locally services even if the hierarchy field indicates the network was
used and simply ignore the fact that the network was used for a small IMS
query to verify the cached object.

Regards
Henrik

===8<===========End of original message text===========

-- 
Best regards,
 mortbox                            mailto:mortbox@gamebox.net