no_cache still caching?

Hello Squid users,

Because some of our clients don't really trust Squid to not return stale
objects, they want to have the option of disabling all caching. Because
they are behind a firewall they still have to go through Squid. I want
to run two Squid processes on the firewall, one that does caching, and
one that doesn't.

I am using Squid 1.1.22, and I have configured the no_cache directive like
this:
        acl nocache proto HTTP
        no_cache deny nocache

After doing this I have stopped the Squid process, emptied the cache with
squid -z and restarted Squid. In the access log file I find the following:

905338421.773 1403 194.151.66.136 TCP_MISS/200 3941 GET
http://www.apache.org/
 - DIRECT/www.apache.org text/html
905338423.378 738 194.151.66.136 TCP_MISS/200 2647 GET
http://www.apache.org/
images/apache_pb.gif - DIRECT/www.apache.org image/gif
905338423.807 1382 194.151.66.136 TCP_MISS/200 5649 GET
http://www.apache.org/
images/apachecon.gif - DIRECT/www.apache.org image/gif
905338429.780 7409 194.151.66.136 TCP_MISS/200 23762 GET
http://www.apache.org
/images/apache_logo.gif - DIRECT/www.apache.org image/gif
905338446.766 18 194.151.66.136 TCP_HIT/200 3941 GET http://www.apache.org/
- NONE/- text/html

Unless I misunderstand the log file the last line shows that Squid is still
caching, even with the no_cache directive.

Before use the proto HTTP acl I have also tried using src 0.0.0.0/0.0.0.0
as acl for the no_cache directive. Also I am quite sure that I am using
the right Squid proxy (the one that should not cache), this one is logging
to a different logfile and running on a different port.

I would appreciate any suggestions you may have.

Best regards,
Edwin de Graaf
Received on Wed Sep 09 1998 - 02:38:16 MDT