On Thu, Oct 02, 2003 at 03:37:05PM +0100, BAZLEY, Sebastian wrote:
> 
> You could try temporarily replacing the HTTP Sampler(s) with a Java Sleep or
> Java Test sampler. This would allow you to check how fast Jmeter can run
> without using any network bandwidth or relying on a 
> remote server.

I might be on to something. I replaced all HTTP samplers with Java
samplers, and then ran the test plan on my workstation, and then on one
of the servers I'm using for the real tests (they're on the same subnet
as the web server, etc.).

On my workstation:

/usr/bin/time java -Xincgc -jar /usr/local/bin/ApacheJMeter.jar -t Test.jmx -p 
jmeter.properties -l Test.jtl -n &
while : ; do ps -ef|grep '[0-9] java'|egrep -v 'grep'|wc -l; done >threads.txt
cat threads.txt |xargs

29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29
29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29
29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29
29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29
29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29
29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29
29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29
29 29 29 29 29 29 29 29

And on the server:

1 1 1 1 1 1 13 1 1 1 1 4 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

Both machines are using the same Java and JMeter versions. The
workstation's CPU was maxed, while the server's was barely moving.
Something must be slowing down Java's threading, but what?

Thanks,
js.
-- 
Jean-Sebastien Morisset, Sr. UNIX Administrator <[EMAIL PROTECTED]>
Personal Home Page <http://jsmoriss.mvlan.net/>
JS & Melanie's Homebrewery <http://brewery.mvlan.net/>
Underwater and Travel Photographs <http://www.mvpix.com/>

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to