On Monday 13 May 2002 03:22, Alexander Dudko wrote:
> delay_pools 1
>
> delay_class 1 2
> delay_access 1 allow allowed_hosts
> delay_access 1 deny all
> delay_parametrs 1 16000/16000 400/400
This will give each user ca 3kbit, and not more than 128kbit in total
of all users.
> Did I make it correctly?
You probably want to leave the per-user limit somewhat higher with
some seconds bucket size to allow for burst, and the total limit
somewhat smaller to avoid saturating the link.
Note: You quite likely also need to tune down the TCP window size, at
least on Linux. 8KB is probably a reasonable value for what you are
doing here.
> As I know, squid will divide all bandwidth between all clients.
> But... what it takes place when some users are turned off, and
> don't use their part of traffic? Will another users's bandwidth
> increase more than 400, or no? And by what way?
The least available amount of both pool limits apply.
> And finally... I have squid-2.4STABLE6 on RedHat Linux.
> Maybe the new version of squid have new features to work with
> bandwidth?
2.4.STABLE6 is fine. There is no additional delay pool features in
later Squids (yet).
-- MARA Systems AB, Giving you basic free Squid support Your source of advanced web reverse proxying solutions http://www.marasystems.com/products/Received on Mon May 13 2002 - 01:57:05 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:08:05 MST