Re: [squid-users] Can't cache www.peelschools.org

On 18/03/2014 5:39 p.m., Alfred Ding wrote:
> Hi,
> Anyone had met this problem? Squid can't cache www.peelschools.org, but it can cache on MS ISA.
> My squid version is 3.1.19 on ubuntu 12.1 lts.
> Thanks much.

The URL with no path is a 302 redirection with no cache control headers.
You will need to upgrade to a later version of Squid with better
HTTP/1.1 caching compliance to store that one.

Following the redirect the main page URL all have Cache-Control:private
meaning "This response only allows a private cache to store it."
 NP: Squid is designed to be used as a *shared* cache.

The scripts on those pages all have serious errors in their Vary
headers. Which is resulting in random reply being either uncompressed or
gzip compressed when they should be the other way around. Some of the
page items are also exploded by over 300% when "compressed" by gzip.

Those scripts and other resources on the pages also all suffer from
missing mandatory headers in responses to revalidation requests.

So whats wrong you ask?
 Well, every object that might be cached is broken by one header or another.
 Combined with Squid-3.1 HTTP/1.1 compliance not being very high.

I recommend trying a newer Squid version, 3.4.4 is out nowdays with
about half again as much HTTP/1.1 capabilities over 3.1. That may help a
bit, but I would not hold out much hope for caching that site until its
server and/or ASP code is fixed.

Amos
Received on Tue Mar 18 2014 - 05:42:50 MDT