RE: [squid-users] Can we use "no-cache" or "max-age=0" to refresh the cached objects

From: Christian Tzolov <Christian.Tzolov_at_tomtom.com>
Date: Thu, 2 Oct 2008 17:29:44 +0200

Thanks for the useful explanation Henrik!

Cheers,
 Chris

Christian Tzolov | Senior Software Developer - Content & Services |
TomTom | christian.tzolov_at_tomtom.com | +31 (0)207575451 |
Oosterdoksstraat 114, 1011 DK, Amsterdam
 

> -----Original Message-----
> From: Henrik Nordstrom [mailto:henrik_at_henriknordstrom.net]
> Sent: Thursday, October 02, 2008 12:52 PM
> To: Christian Tzolov
> Cc: Amos Jeffries; squid-users_at_squid-cache.org
> Subject: RE: [squid-users] Can we use "no-cache" or "max-age=0" to
refresh
> the cached objects
>
> On tor, 2008-10-02 at 10:57 +0200, Christian Tzolov wrote:
>
> > I am scared by the "I believe" part :).
> >
> > 1. Can we relay on Squid to always update its cached content if the
> > response is newer (e.g. response has new Expires date and no other
> > validators)?
>
> Yes.
>
> > 2. Squid does not change/optimize its behavior under high load in
way
> > that could affect assumption (1)?
>
> Under very high load you may in worst case end up with no on-disk copy
> of the cached object if an overload condition does not allow writing
> objects out to disk..
>
> > 3. If (1) holds is this a (HTTP) standard behavior or Squid
> > implementation?
>
> Yes. Or to be exact the standard requires the old copy to be
invalidated
> if different from the new. There is no requirement to cache the new
> representation.
>
> Note however that there is a significant difference between no-cache
and
> max-age=0:
>
> * no-cache forces a new retreival, The response is not allowed to be
> satisfied from cache.
>
> * max-age=0 forces a cache validation and allows the old version to
be
> returned if the server says it's still fresh.
>
> Regards
> Henrik
Received on Thu Oct 02 2008 - 15:36:04 MDT

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