RE: [squid-users] min-fresh / max-stale not working?

From: Markus Karg <karg_at_quipsy.de>
Date: Thu, 4 Sep 2008 08:24:23 +0200

I see. So that means that SQUID has much too less contributors and a
unstable financial foundation? What is the amount of people and money
that you would need to make the HTTP/1.1 support perfect?

> -----Original Message-----
> From: adrian.chadd_at_gmail.com [mailto:adrian.chadd_at_gmail.com] On Behalf
> Of Adrian Chadd
> Sent: Donnerstag, 4. September 2008 02:27
> To: Markus Karg
> Cc: Amos Jeffries; squid-users_at_squid-cache.org
> Subject: Re: [squid-users] min-fresh / max-stale not working?
>
> When someone contributes the work or funds development.
>
>
>
> Adrian
>
> 2008/9/4 Markus Karg <karg_at_quipsy.de>:
> > Is there a plan when HTTP/1.1 completely will be supported in all
> sides?
> > I mean, I hardly can't believe it -- HTTP/1.1 was specified in 2008.
> Why
> > waiting so long?
> >
> > Thanks
> > Markus
> >
> >> -----Original Message-----
> >> From: Amos Jeffries [mailto:squid3_at_treenet.co.nz]
> >> Sent: Mittwoch, 3. September 2008 15:40
> >> To: Markus Karg
> >> Cc: squid-users_at_squid-cache.org
> >> Subject: Re: [squid-users] min-fresh / max-stale not working?
> >>
> >> Markus Karg wrote:
> >> > Sorry it was a typo. The test was done mit SQUID-2.7-STABLE4
> >> actually.
> >> > The HTTP/1.1-Support is only experimental???
> >>
> >> Brand new in 2.7 and some bugs still being found.
> >> It's also only on one side of Squid, the one which links to Servers
> >> IIRC, so the client-facing code is still HTTP/1.0-only.
> >>
> >> Amos
> >>
> >> >
> >> >> -----Original Message-----
> >> >> From: Amos Jeffries [mailto:squid3_at_treenet.co.nz]
> >> >> Sent: Mittwoch, 3. September 2008 07:14
> >> >> To: Markus Karg
> >> >> Cc: squid-users_at_squid-cache.org
> >> >> Subject: Re: [squid-users] min-fresh / max-stale not working?
> >> >>
> >> >>> Dear SQUID Community,
> >> >>>
> >> >>> it seems as if SQUID is not dealing correctly with "min-fresh"
> and
> >> >>> "max-stale":
> >> >>>
> >> >>> Currently we are evaluating the use of SQUID-2.6-STABLE4. It
all
> >> >> seems
> >> >>> to work pretty well, but just "min-fresh" and "max-stale" is
not
> >> >>> working. Our client agent wants to guarantee to get data that
is
> >> >> fresh
> >> >>> for a specific amount of time. So we provide "min-fresh=3500"
> and
> >> >>> "max-stale=0". To verify SQUID's behaviour we have programmed
an
> >> >> origin
> >> >>> server the always responds with some static headers and entity
> >> data,
> >> >> and
> >> >>> a client that requests exactly that information, via SQUID as a
> >> >> proxy.
> >> >>> The client uses the Cache-Control header with a min-fresh=3500
> and
> >> >>> max-stale=0 value, and the server is always sending data with a
> >> >>> max-age=3600 value. But the client gets from SQUID a 200 OK
> >> response
> >> >>> having max-age=3600 and Age=502! So, the current age of 502
plus
> >> the
> >> >>> desired min-fresh of 3500 is 4002, minus the max-stale of 0
> still
> >> is
> >> >>> 4002, what is much more than the max-age of 3600 -- so the
> request
> >> >>> cannot be satisfied without a warning, since the response will
> not
> >> > be
> >> >>> fresh long enough! So we expect to get at least a Warning
> header.
> >> > But
> >> >>> there is none! It looks like SQUID just ignores the min-
> fresh=3500
> >> >> and
> >> >>> max-stale=0 headers!
> >> >>>
> >> >>> The HTTP/1.1 specification says:
> >> >>> 13.1.2 Warnings
> >> >>> Whenever a cache returns a response that is neither first-hand
> nor
> >> >>> "fresh enough" (in the sense of condition 2 in section 13.1.1),
> it
> >> >> MUST
> >> >>> attach a warning to that effect, using a Warning
general-header.
> >> >>> also it says:
> >> >>> 13.1.1 Cache Correctness
> >> >>> If a stored response is not "fresh enough" by the most
> restrictive
> >> >>> freshness requirement of both the client and the origin server,
> in
> >> >>> carefully considered circumstances the cache MAY still return
> the
> >> >>> response with the appropriate Warning header.
> >> >>>
> >> >>> In the default case, this means it meets the least restrictive
> >> >> freshness
> >> >>> requirement of the client, origin server, and cache (see
section
> >> >> 14.9)
> >> >>> So for me it looks as if SQUID is buggy, since it does not add
> the
> >> >>> mandatory Warning header. Can that be true? Or do I have to
> enable
> >> >> some
> >> >>> switch like "HTTP/1.1-Compliance = YES"?
> >> >> Squid 2.6 is HTTP/1.0 only. For any HTTP/1.1 stuff you will
need
> >> > Squid
> >> >> 2.7 and its experimental support.
> >> >>
> >> >> As for the cache controls, someone more knowledgeable will
> > hopefully
> >> >> speak
> >> >> up.
> >> >>
> >> >> Amos
> >> >
> >>
> >>
> >> --
> >> Please use Squid 2.7.STABLE4 or 3.0.STABLE8
> >
> >
Received on Thu Sep 04 2008 - 06:24:38 MDT

This archive was generated by hypermail 2.2.0 : Thu Sep 04 2008 - 12:00:02 MDT