Re: [squid-users] doubt regarding delay_access

2005-02-20 Thread Henrik Nordstrom
On Mon, 14 Feb 2005, Dinil Divakaran wrote:
delay_access 2 allow allowed_hosts
delay_access 1 allow localhost
delay_access 1 allow all
delay_access is sorted per pool, always starting with pool 1, then pool 2 
etc... so the above configuration is equivalent to

  delay_access 1 allow localhost
  delay_access 1 allow all
  delay_access 2 allow allowed_hosts
so with the above configuration all requests will go into pool 1.
Regards
Henrik



Re: [squid-users] doubt regarding delay_access

2005-02-20 Thread Dinil Divakaran


delay_access 2 allow allowed_hosts
delay_access 1 allow localhost
delay_access 1 allow all
delay_access is sorted per pool, always starting with pool 1, then pool 2 
etc... so the above configuration is equivalent to

 delay_access 1 allow localhost
 delay_access 1 allow all
 delay_access 2 allow allowed_hosts
so with the above configuration all requests will go into pool 1.
Thanks Henrik.
This solved the confusion. Perhaps this explanation should be added 
to the squid.conf file to make it clearer.

One more doubt - how does squid manage the downloading rate from
any web server to the squid; is it just by introducing a delay in
the application level while receiving data ?


Re: [squid-users] doubt regarding delay_access

2005-02-20 Thread Henrik Nordstrom
On Sun, 20 Feb 2005, Dinil Divakaran wrote:
One more doubt - how does squid manage the downloading rate from
any web server to the squid; is it just by introducing a delay in
the application level while receiving data ?
When the delay pool is empty Squid simply stops reading data from the 
server until the pool has been refilled by the refill rate.

Regards
Henrik


Re: [squid-users] doubt regarding delay_access

2005-02-20 Thread Henrik Nordstrom
On Sun, 20 Feb 2005, Dinil Divakaran wrote:
This solved the confusion. Perhaps this explanation should be added to the 
squid.conf file to make it clearer.
Please file a bug report on it, preferably with a suggestion on wording 
(optimally as a patch to cf.data.pre)

Regards
Henrik