Re: [squid-users] About Squid Delay Pools, help needed

From: <[email protected]>
Date: Mon, 4 Jul 2005 18:50:21 -0300 (BRT)

Hi Abu Khaled,

    I am trying to understand delay_pools, but I am a little stucked, even
after reading Squid FAQs and Visolve instructions.
    Could you please help me?

>> acl client1 src 66.67.68.69/255.255.255.255
>> delay_pools 1
>> delay_class 1 2
>> delay_parameters 1 -1/-1 4000/500000
>> delay_access 1 allow client1
>> delay_access deny all
>>
>> So second question is: what am I missing?
>
> Once the clients bucket fils up with 500KB the speed is droped to
> 32Kbps no mater what file size he downloads/browses next and stays
> that way until the bucket replenishes (stops downloading/browsing for
> a while).

    So, the maximum will act like a kind of "bytes quota"?
    I mean, if the client reachs 4 Mbits, it will be dropped to 32Kbits/s,
until it stops browsing/downloading and have its bucket replenished?
How long (rate?) it will take to be replenished?
    If it is right, and I try:

    delay_parameters 1 -1/-1 8000/32000 ( I am using this line )

    it means that, if my client just browse a big portal index (bigger
than 256Kbits) it will be dropped to 64Kbits/s?

    I want to limit clients that keeps "downloading" (any kind of file),
including sites with a lot of images and similars, without disturbing
the ones that are doing their researchs without dragging my bandwidth.
    This rule applies to everybody. There is no way to separate users in
groups.
    Are my ideias about delay pools correct or I am "lost"?
    Really thank you for your attention

Regards,

Freitas
Received on Mon Jul 04 2005 - 15:43:44 MDT

This archive was generated by hypermail pre-2.1.9 : Mon Aug 01 2005 - 12:00:02 MDT