Mailing-List: contact users-help@httpd.apache.org; run by ezmlm
Precedence: bulk
Reply-To: users@httpd.apache.org
Sender: obo@bourse.ch
Message-ID: <3CD25E70.9671070F@bourse.ch>
Date: Fri, 03 May 2002 11:54:56 +0200
From: Owen Boyle <obo@bourse.ch>
MIME-Version: 1.0
To: users@httpd.apache.org
Subject: Re: Apache hangs up for more than 100 concurrent users.
References: <00e601c1ef5b$05cd7d60$ef02050a@mahindrabt.com>
 <3CCD0BDF.7843FBD8@bourse.ch> <009101c1f1c1$da9710c0$ef02050a@mahindrabt.com>
 <3CD134BB.79690CC4@bourse.ch> <000001c1f266$5d349e40$ef02050a@mahindrabt.com>
 <3CD24788.621F9B45@bourse.ch> <001d01c1f284$1dfe6260$ef02050a@mahindrabt.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Tushar Swarnakar wrote:
> 
> The server hangs forever and no requests are ever served again, even
> after I stop my performance tester. The server does not response any more.

Very strange. I can think of two things:

- You have dicovered a bug in an ancient version of apache (1.3.9).
Upgrade to the current version (1.3.24) and try again.

- Your performance tester is buggy. What is it? Did you write it
yourself? Is it "ab"? Is it running on the same machine as apache (i.e.
loading localhost)? If so, are you sure it is droppping connections
after it is stopped (is a zombie hanging around, blocking sockets?) What
do you have to do to get running again - reboot? Or is it sufficient to
restart apache?

As an experiment, can you test the theory that the problem only occurs
if you emulate N clients where MaxClients = 2N, i.e. set:

MaxClients = 100, test 50 clients
MaxClients = 50, test 25 clients
MaxClients = 30, test 15 clients
etc.

see if there is any logic to it..

Rgds,

Owen Boyle.

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@httpd.apache.org
For additional commands, e-mail: users-help@httpd.apache.org