[squid-users] RE: EXTERNAL: Re: [squid-users] Problem accessing a particular site through squid

From: Bucci, David G <david.g.bucci_at_lmco.com>
Date: Wed, 15 Sep 2010 15:36:39 -0400

I just tried it on my Ubuntu 10.04 running the std repo squid 3.0.STABLE19-1. I see the same behavior, running Google Chrome (nightly experimental build).

Interestingly, through Squid, it hangs trying to retrieve something from http://directgov.stcllctrs.com ... which is in a <noscript></noscript> block. So something is causing Chrome to think it should execute noscripts (?)

The only other thing I notice is that your certificate is invalid for SSL to the smallsteps4life.direct.gov.uk ... the server name is a wildcarded *.workingwithprofero.com. Don't think that should be affecting Squid though.

When I save the html source before/after squid, they have a huge amount of difference ... oh, interesting, the noscript isn't correct, it's &lt;noscript&gt; instead of correctly using <> chars. That explains why chrome is trying to retrieve the URL.

I dunno ... nothing jumps out at me.

-----Original Message-----
From: Amos Jeffries [mailto:squid3_at_treenet.co.nz]
Sent: Wednesday, September 15, 2010 10:21 AM
To: squid-users_at_squid-cache.org
Subject: EXTERNAL: Re: [squid-users] Problem accessing a particular site through squid

On 15/09/10 23:24, Seb Harrington wrote:
>
> Hi everyone,
>
> I have a problem when accessing http://smallsteps4life.direct.gov.uk/
> through squid.
>
> When accessing the site directly the site is properly formatted, when
> accessing through squid the site appears 'unformatted', some of the
> images do not load and it looks as if the CSS has not been applied.
>
> I thought this behaviour was a little strange so I've tested it on two
> more instances of squid, one a default fresh install allowing everything
> through (the all acl).
>
> When accessing the site these are the logs:
> access.log: http://pastebin.com/HtkyfjUJ
> store.log: http://pastebin.com/0NjnDZzW
> cache.log: did not output anything useful or informative.
>
> I'm using the ubuntu version of squid3 (apt-get squid3) and I'm using
> ubuntu 10.04 Lucid Lynx.
>
> Squid version: Squid Cache: Version 3.0.STABLE19
><snip>
> Could someone please run that website through their version of squid for
> me and let me know if this is a squid issue, a website issue or a bug in
> the ubuntu packaged version of squid.
>
> Cheers,
>
> Seb
>

That trace from the "working" squid? There is zero CSS in it. Just
JavaScript files that generate page content on the fly. Most of the
content seems to be going through HTTPS which passes straight through Squid.

The all ACL working where regular config catches only some occasional
files makes me think either those files are on a domain being blocked,
or you have regex patterns that are catching more than you are aware of.

Amos

-- 
Please be using
   Current Stable Squid 2.7.STABLE9 or 3.1.8
   Beta testers wanted for 3.2.0.2
Received on Wed Sep 15 2010 - 19:37:00 MDT

This archive was generated by hypermail 2.2.0 : Thu Sep 16 2010 - 12:00:03 MDT