RE: [squid-users] Delay Pools

From: Robert Trouchet <[email protected]>
Date: Wed, 15 Dec 2004 16:23:26 +0800

> 3. How should I look at the cachemgr.cgi file. It is not viewable with
> Notepad.

You must place this in your web server's cgi-bin folder and setup access
in squid - see the docs on this. It doesn't need to be on the same server,
but can be (different port for web server of course). Apparently you can
use squidclient to retrieve this info, but I have not tried this yet.

--> I don't have a web server.
 
> 4. My network is a class A (10.a.b.c). I have created a dedicated
> account just for testing. The restrictions need to be on a user by
> user basis as users can log onto any computer on the network.

Ok, but it may be simpler to troubleshoot if you add an ACL for a specific
IP and use that for now. Once it's working then move on to group auth
ACLs.

--> I have set an account for one IP address (my own) and I have seen how
slow the slow
Delay pool really is. It works for that user but not for my user
authenticated by the NT group.

I would recommend you consider Class 2 or 3 ACLs, with limits set
accordingly so that one user can't sap all the bandwidth for the group.

Below is my current setup, which is working well, but still subject to
finetuning. This is supporting approx 325 concurrent users on 7 vlans,
and seems to work nicely. It is set to use at most 80% of our bandwidth
for the entire pool, and so far we've seen at most about 65%. Before this
we would see peaks of 98% daily.

--> I haven't had problems with a single user hogging the connection so I
haven't bothered with class 3 pools.

--> I am looking at the use of delay pools as a punitive measure to restrict
users who still need Internet access rather than a way of parcelling out
access fairly.
Received on Wed Dec 15 2004 - 01:23:52 MST

This archive was generated by hypermail pre-2.1.9 : Sat Jan 01 2005 - 12:00:02 MST