Re: [squid-users] Problem with HTTP Headers

 On Sun, 13 Nov 2011 19:14:48 +0200, Ghassan Gharabli wrote:
> Dear Amos,
>
> After allowing access "Head" method in Squid Config
>
> I deleted www.facebook.com from cache andthen I tried executing
>
> squidclient -m head http://www.facebook.com
>
> Results :
>
> HTTP/1.0 302 Moved Temporarily
> Location: http://www.facebook.com/common/browser.php
> P3P: CP="Facebook does not have a P3P policy. Learn why here:
> http://fb.me/p3p"
> Set-Cookie: datr=hfW_TtrAQmi_2SxwAUY4EjPH; expires=Tue, 12-Nov-2013
> 16:51:17 GMT
> ; path=/; domain=.facebook.com; httponly
> Content-Type: text/html; charset=utf-8
> X-FB-Server: 10.53.10.59
> X-Cnection: close
> Content-Length: 0
> Date: Sun, 13 Nov 2011 16:51:17 GMT
> X-Cache: MISS from Peer6.skydsl.net
> X-Cache-Lookup: MISS from Peer6.skydsl.net:3128
> Connection: close
>
> I am not seeing any pragma or cache-control and expires! but redbot
> shows the correct info there!.

 Ah, your squidclient is not sending a user-agent header. You will need
 to add -H "user-Agent: foo"

>
> BTW .. I am also using store_url but im sure nothing is bad there . I
> am only playing with Dynamic URL regarding to Pictures and Videos
> extensions so I have only one thing left for me to try which is
> unlike
> to do it ..
>
> acl facebookPages urlpath_regex -i /(\?.*|$)
>
> First does this rule affect store_url?

 This is just a pattern definition. It only has effect where and when
 the ACL is used. The config I gave you only used it in the "cache deny"
 access line.

 That said, "cache deny" prevents things going to the cache, where
 storeurl* happens.

>
> For example when we have url like
>
> http://www.example.com/1.gif?v=1244&y=n
>
> I can see that urlpath_regex requires Full URL which means this rule
> matches :
>
> http://www.example.com/stat?date=11

 The pattern begins with '/' and the "cache" access line I gave you
 included another ACL. Which tested the domain name was *.facebook.com.

 It will match things like:
   http://www.facebook.com/?v=1244&y=n

 but *not* match things like:
   http://www.example.com/1.gif?v=1244&y=n

>
> I will try to ignore this rule and let me focus on facebook problem
> since we have more than 60% traffic on Facebook.
>

 As you wish. I added that line because I noticed the front page for FB
 you wanted to non-cache has the URL path starting with the two
 characters "/?" instead of .html or .php.

 Amos
Received on Sun Nov 13 2011 - 22:57:42 MST