Re: [squid-users] deny cache of all .asp dynamic pages

> thx,
> the problem is my lines deny all .asp webpages, but howto configure to
> deny asp only from the one certain domain?

create another acl ... dstdomain line listing the domain. then combine
them together on the "cache deny" line. Preferrably with the faster
dstdomain ACL first/leftmost.

>
> btw. could you pls post lines howto use acl, and "refresh_pattern" for
> .asp?

refresh_pattern does not support ACLs. It's a global regex matcher.

>
> greetings
> Piotr
>
>
>
> On Sun, Feb 15, 2009 at 11:54 PM, Amos Jeffries <squid3_at_treenet.co.nz>
> wrote:
>>> hi
>>>
>>> I have Squid-3.0.PRE5 installed on debian etch and we have problems
>>
>> Please use the squid3 version from Lenny, its much more stable and less
>> buggy than the beta.
>>
>>> with some .asp webpages:
>>> DOUBLE click the Login button, page cannot be found etc.
>>>
>>> Howto configure squid.conf to deny cache of all dynamic .asp webpages?
>>>
>>> I've added to my squid.conf the following lines:
>>>
>>> hierarchy_stoplist \.asp ?
>>> acl QUERY1 urlpath_regex \.asp \?
>>> cache deny QUERY1
>>>
>>> I've found something like:
>>>
>>> acl ASP urlpath_regex .asp$
>>> cache deny ASP
>>>
>>> as well, but I'm not sure which lines are correct?
>>
>> They both are. Pick the one that you like the look of most.
>>
>>>
>>>
>>> should I put both to squid.conf?
>>>
>>
>> To meet your stated request either will do.
>>
>> However I'd urge you to consider carefully WHY, you need to do this.
>> If the nature of the .asp pages is breaking your Squid, then its
>> breaking
>> everyones elses' cache as well.
>>
>> Squid will cache what it is allowed to, the asp web app should be giving
>> it correct Cache-Control and Expires headers. If these are missing squid
>> will estimate based on last-change.
>>
>> For dynamic pages we currently recommend dropping the old default QUERY
>> acl, and using "refresh_pattern -i (/cgi-bin/\?) 0 0% 0" as the second
>> to
>> last pattern. This makes squid discard all dynamic objects without
>> working
>> expiry info immediately, while also letting those good sites that
>> provide
>> expiry to work.
>>
>> Amos
>>
>>
>
Received on Mon Feb 16 2009 - 00:02:06 MST