Re: [squid-users] access.log equivalent for server side

From: Henrik Nordstrom <hno@dont-contact.us>
Date: Wed, 20 Apr 2005 10:01:45 +0200 (CEST)

On Wed, 20 Apr 2005, Thien Vu wrote:

> That's what I'm thinking also, but we have nearly 300 acls and a
> similar number of http_access rules.

Ok.

What kinds of acls are these?

> Our setup is like this. We have a machine proxy.company.com that runs
> both a proxy and a web server. We monitor the web server by grabbing a
> file every minute. We monitor the proxy by grabbing that same file
> through the proxy every minute. The proxy is configured to not cache
> any results (other than dns and authentication credential and other
> adminitrative data, but specifically not web pages). Over a day, the
> direct requests to the web server for the file takes an average of 20
> milliseconds. On the other hand, the requests through the proxy
> average 888 milliseconds.

Sounds a bit excessive difference indeed.

Do you have a problem with CPU usage?

Is there any difference if you for a test cuts down on the number of
http_access rules?

Any warnings in cache.log?

Any swap activity on the system?

Are you also using a redirector helper?

What authentication method scheme is used?

Regards
Henrik
Received on Wed Apr 20 2005 - 02:01:52 MDT

This archive was generated by hypermail pre-2.1.9 : Sun May 01 2005 - 12:00:04 MDT