You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@jmeter.apache.org by Scott Eade <se...@backstagetech.com.au> on 2002/07/09 10:06:11 UTC

How many threads from a single machine?

Hi,

I am interested in obtaining some feedback as to the number of threads (test
sessions) I should expect to be able to execute from a single machine.
Obviously the intensity of the script and the type of machine used will have
a significant impact on this, so I guess some values obtained from
experience that include a little information about the scripts and the
machine configuration would be really helpful.

At the very worst, how about just some basic numbers so that I can get an
idea of the order of magnitude I should expect - i.e. 100, 1000, 2000?

Thanks,

Scott


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: How many threads from a single machine?

Posted by Mike Stover <ms...@apache.org>.
To answer your questions:

1. No, you should never run JMeter on the same machine as the server you are testing.  
When I say JMeter server, I mean an instance of JMeter running in "server mode".  In 
version 1.7.1, JMeter is run in server mode with "jmeter -s".  In 1.7, you use the jmeter-server 
script file.

2.  You can run JMeter servers on as many machines as you have/like.  One machine, 
running in normal GUI mode can control all of them.

3. You don't need to do any special to combine the results files.  Any listeners in your tests 
will receive the results from all the remote JMeter servers you run.

-Mike

On 10 Jul 2002 at 22:54, Scott Eade wrote:

> Hi Mike,
> 
> > From: "Mike Stover" <ms...@apache.org>
> > 
> > This sounds like more than one machine can handle, regardless.  Have you
> > thought about using JMeter's distributed capabilities?  It allows you to run a
> > JMeter GUI on one machine that can control multiple JMeter servers running on
> > other machines.  
> 
> Yes, I believe I will need to do this.  I'm thinking of obtaining a new 2GHz
> P4 machine as the main testing machine and then seeing what resources I can
> pilfer of other machines that happen to be available at the time.
> 
> Could you possibly provide some feedback on the questions included I the
> paragraph below?  Most importantly I need to know if the results files will
> all come back to the controlling machine.
> 
> >> From the various comments received I am thinking that multiple client
> >> machines are going to be necessary.  The Remote Testing documentation
> >> appears to suggest that the jmeter testing engine would be running on the
> >> application server (I assume in my case this would be the web server), but
> >> wouldn't this be using resources that should be kept available for the
> >> application?  I assume remote testing can also support multiple testing
> >> engines that execute their scripts against a common application server - is
> >> this correct?  How would I go about combining the results from the testing
> >> executed on multiple clients - would it be a simple matter of merging the
> >> results files?  Has anyone tried this?
> 
> Thanks,
> 
> Scott
> -- 
> Scott Eade
> Backstage Technologies Pty. Ltd.
> 
> 
> 
> --
> To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
> For additional commands, e-mail: <ma...@jakarta.apache.org>
> 



--
Michael Stover
mstover1@apache.org
Yahoo IM: mstover_ya
ICQ: 152975688

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: How many threads from a single machine?

Posted by Scott Eade <se...@backstagetech.com.au>.
Hi Mike,

> From: "Mike Stover" <ms...@apache.org>
> 
> This sounds like more than one machine can handle, regardless.  Have you
> thought about using JMeter's distributed capabilities?  It allows you to run a
> JMeter GUI on one machine that can control multiple JMeter servers running on
> other machines.  

Yes, I believe I will need to do this.  I'm thinking of obtaining a new 2GHz
P4 machine as the main testing machine and then seeing what resources I can
pilfer of other machines that happen to be available at the time.

Could you possibly provide some feedback on the questions included I the
paragraph below?  Most importantly I need to know if the results files will
all come back to the controlling machine.

>> From the various comments received I am thinking that multiple client
>> machines are going to be necessary.  The Remote Testing documentation
>> appears to suggest that the jmeter testing engine would be running on the
>> application server (I assume in my case this would be the web server), but
>> wouldn't this be using resources that should be kept available for the
>> application?  I assume remote testing can also support multiple testing
>> engines that execute their scripts against a common application server - is
>> this correct?  How would I go about combining the results from the testing
>> executed on multiple clients - would it be a simple matter of merging the
>> results files?  Has anyone tried this?

Thanks,

Scott
-- 
Scott Eade
Backstage Technologies Pty. Ltd.



--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: How many threads from a single machine?

Posted by Mike Stover <ms...@apache.org>.
This sounds like more than one machine can handle, regardless.  Have you 
thought about using JMeter's distributed capabilities?  It allows you to run a JMeter 
GUI on one machine that can control multiple JMeter servers running on other 
machines.  

-Mike

On 10 Jul 2002 at 12:48, Scott Eade wrote:

> Thanks for all the replies - very helpful.
> 
> Here is a little more information about what I am proposing to do.
> 
> I need to test a site with something like 7,500 sessions being initiated per
> hour.  The sessions are going to be relatively long, say 15 minutes for
> 5,000 and over an hour for the remaining 2,500.  The shorter sessions will
> have a 45 second pause between each page request and the longer running
> sessions a combination of 30 second and 5 minute pauses.
> 
> As suggested by Michael and Olav I will keep an eye on the CPU usage of the
> machines running the test sessions to ensure I am not getting any adverse
> effects relating to the overloading of these machines.  At least a second
> machine running a smaller number of threads to monitor the relative response
> time is a good idea, thanks Michael.
> 
> We have decided to ignore network effects for our current testing by using
> machines connected to the same switch as the web server.
> 
> From the various comments received I am thinking that multiple client
> machines are going to be necessary.  The Remote Testing documentation
> appears to suggest that the jmeter testing engine would be running on the
> application server (I assume in my case this would be the web server), but
> wouldn't this be using resources that should be kept available for the
> application?  I assume remote testing can also support multiple testing
> engines that execute their scripts against a common application server - is
> this correct?  How would I go about combining the results from the testing
> executed on multiple clients - would it be a simple matter of merging the
> results files?  Has anyone tried this?
> 
> Thanks again,
> 
> Scott
> 
> > From: Scott Eade <se...@backstagetech.com.au>
> > Reply-To: "JMeter Users List" <jm...@jakarta.apache.org>
> > Date: Tue, 09 Jul 2002 18:06:11 +1000
> > To: JMeter Users List <jm...@jakarta.apache.org>
> > Subject: How many threads from a single machine?
> > 
> > Hi,
> > 
> > I am interested in obtaining some feedback as to the number of threads (test
> > sessions) I should expect to be able to execute from a single machine.
> > Obviously the intensity of the script and the type of machine used will have
> > a significant impact on this, so I guess some values obtained from
> > experience that include a little information about the scripts and the
> > machine configuration would be really helpful.
> > 
> > At the very worst, how about just some basic numbers so that I can get an
> > idea of the order of magnitude I should expect - i.e. 100, 1000, 2000?
> > 
> > Thanks,
> > 
> > Scott
> > 
> > 
> > --
> > To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
> > For additional commands, e-mail: <ma...@jakarta.apache.org>
> > 
> 
> 
> --
> To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
> For additional commands, e-mail: <ma...@jakarta.apache.org>
> 



--
Michael Stover
mstover1@apache.org
Yahoo IM: mstover_ya
ICQ: 152975688

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: How many threads from a single machine?

Posted by Scott Eade <se...@backstagetech.com.au>.
Thanks for all the replies - very helpful.

Here is a little more information about what I am proposing to do.

I need to test a site with something like 7,500 sessions being initiated per
hour.  The sessions are going to be relatively long, say 15 minutes for
5,000 and over an hour for the remaining 2,500.  The shorter sessions will
have a 45 second pause between each page request and the longer running
sessions a combination of 30 second and 5 minute pauses.

As suggested by Michael and Olav I will keep an eye on the CPU usage of the
machines running the test sessions to ensure I am not getting any adverse
effects relating to the overloading of these machines.  At least a second
machine running a smaller number of threads to monitor the relative response
time is a good idea, thanks Michael.

We have decided to ignore network effects for our current testing by using
machines connected to the same switch as the web server.

>From the various comments received I am thinking that multiple client
machines are going to be necessary.  The Remote Testing documentation
appears to suggest that the jmeter testing engine would be running on the
application server (I assume in my case this would be the web server), but
wouldn't this be using resources that should be kept available for the
application?  I assume remote testing can also support multiple testing
engines that execute their scripts against a common application server - is
this correct?  How would I go about combining the results from the testing
executed on multiple clients - would it be a simple matter of merging the
results files?  Has anyone tried this?

Thanks again,

Scott

> From: Scott Eade <se...@backstagetech.com.au>
> Reply-To: "JMeter Users List" <jm...@jakarta.apache.org>
> Date: Tue, 09 Jul 2002 18:06:11 +1000
> To: JMeter Users List <jm...@jakarta.apache.org>
> Subject: How many threads from a single machine?
> 
> Hi,
> 
> I am interested in obtaining some feedback as to the number of threads (test
> sessions) I should expect to be able to execute from a single machine.
> Obviously the intensity of the script and the type of machine used will have
> a significant impact on this, so I guess some values obtained from
> experience that include a little information about the scripts and the
> machine configuration would be really helpful.
> 
> At the very worst, how about just some basic numbers so that I can get an
> idea of the order of magnitude I should expect - i.e. 100, 1000, 2000?
> 
> Thanks,
> 
> Scott
> 
> 
> --
> To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
> For additional commands, e-mail: <ma...@jakarta.apache.org>
> 


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>