Re: [squid-users] acl maxconn per file or url

From: Chudy Fernandez <chudy_fernandez_at_yahoo.com>
Date: Tue, 30 Jun 2009 17:56:34 -0700 (PDT)

Theoretically suppose to work. although it limits 4 connection but the download doesn't finish.
I wonder how site's replied that gave 4 connections when I'm requesting 16.
http code 416? or 408...

----- Original Message ----
From: Amos Jeffries <squid3_at_treenet.co.nz>
To: Henrik Nordstrom <henrik_at_henriknordstrom.net>
Cc: Chudy Fernandez <chudy_fernandez_at_yahoo.com>; squid-users_at_squid-cache.org
Sent: Wednesday, July 1, 2009 8:25:52 AM
Subject: Re: [squid-users] acl maxconn per file or url

On Tue, 30 Jun 2009 18:35:09 +0200, Henrik Nordstrom
<henrik_at_henriknordstrom.net> wrote:
> tis 2009-06-30 klockan 15:13 +1200 skrev Amos Jeffries:
>> On Mon, 29 Jun 2009 18:50:30 -0700 (PDT), Chudy Fernandez
>> <chudy_fernandez_at_yahoo.com> wrote:
>> > I think this well help
>> >
>> > acl maxcon maxconn 4
>> > acl partial rep_header Content-Range .*
>> > http_reply_access deny partial maxcon
>> >
>>
>> I wonder....
>>
>> What this _does_ is cause replies to be sent back to the client with all
>> range encoding and wrapping, but without the range position information
>> or
>> other critical details in the Content-Range: header.
>
> No it doesn't. It prevents replies with Content-Range if there is more
> than 4 concurrent connections from the same IP (with no regard to what
> those connections is being used for).
>
> It's http_reply_access, not http_header_access...
>

Doh! Thanks Henrik.

>> This does not prevent Squid from fetching the multiple requests for
>> ranges
>> in the first place, nor save any bandwidth used by Squid doing so.
>
> But it does cause Squid to abort the partial requests once headers have
> been received.
>
> Regards
> Henrik

Amos

      
Received on Wed Jul 01 2009 - 00:56:41 MDT

This archive was generated by hypermail 2.2.0 : Wed Jul 01 2009 - 12:00:04 MDT