Mailing-List: contact jmeter-user-help@jakarta.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "JMeter Users List" <jmeter-user@jakarta.apache.org>
Message-ID: 
 <E25EEB2EF1D8F949BEAD22AFF6B9298E01BA25ED@sc4mail.corp.snowball.com>
From: Jason Fisher <jfisher@ign.com>
To: "'JMeter Users List'" <jmeter-user@jakarta.apache.org>
Subject: RE: Extending JMeter with a new generative controller
Date: Fri, 27 Sep 2002 11:14:43 -0700
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01C26651.C12BEA20"

------_=_NextPart_001_01C26651.C12BEA20
Content-Type: text/plain;
	charset="iso-8859-1"

> And I think Eric Siegerman has a good suggestion for you.

For my current purpose I just want to simulate a real load on a real server.
I am trying to reproduce basically the same state that occurs under actual
conditions, so that I can probe the server and the code in question while
it's happening.  
It *would* be nice to use real loads that toppled a server for future
regression test, but only *after* you solve how to keep it from toppling
under such a load, right?

Incidentally, Mike, it would seem to me that the interaction from the gui
down to the underlying Sampler object involves many objects.   I want to
intercept the HTTP Request controller behavior at a couple of points.
First, in the gui I want to take in a file location for the apache log to
use, as well as the domain of the server to test, but I no longer need the
'path' info.  Second I want the path info to no longer come from the gui,
but rather to come from a new line of the log for each request that is made.
I am currently trying to trace the code from gui to sampler, but there is a
lot there and it's all new to me.  I am positive that it would be faster to
simply pick your brain for the answers, but I realize that you are, of
course, a busy man.....

Thanks,
Jason


------_=_NextPart_001_01C26651.C12BEA20--