On 28 February 2014 20:26, Jared Griffith <[email protected]> wrote: > So even though I'm recording the data results, I'm going to get different > results based on whether or not I'm using the GUI?
The results should be the same. However GUI uses more resources than non-GUI and for some tests this can affect JMeter performance. > The docs mention that > is not the case and the results should be the same. The main difference is > that I'm seeing the data results in GUI mode, which results in a .jtl file > of about 2.5 Mb (xml) as opposed to the CLI mode where the response data is > not recorded (even thought it's the same exact jmx file), and the resulting > .jtl file is all of about 40 K (xml). That seems strange. What command did you use to run JMeter in non-GUI mode? The only difference related to result logging I can think of is that some of the non-GUI scripts automatically add a results file. The configuration for this file is taken from jmeter.properties. In the GUI case, results are not saved unless you configure a Listener with a file name. The Listener takes its defaults from jmeter.properties as well, but can be ovverridden in the Config. If you run a test in non-GUI mode which has a Listener configured, you should get the same results as in GUI mode. You may additionally get another results file, depending on the JMeter command line parameters. And that may have different settings. I suppose if the GUI listener uses the same file name as the non-GUI result file, the contents of the file may not be what you expect ... > As for the differences between the 2 machines, one's a server, so a bit > more powerful (Xeon vs Core2Duo), and more RAM on the server. The > properties files for Jmeter are exactly the same. But I'm getting more > results from my less powerful workstation than I am on the server. > > > On Fri, Feb 28, 2014 at 12:17 PM, UBIK LOAD PACK Support < > [email protected]> wrote: > >> Hello, >> Regarding difference between GUI and Non GUI, this is absolutely regular. >> GUI is for creating Test Plan, debugging, recording. >> Non GUI is for load testing. >> >> See: >> - http://www.ubik-ingenierie.com/blog/jmeter_performance_tuning_tips/ >> - http://jmeter.apache.org/usermanual/best-practices.html >> >> >> Regarding differences between 2 machines in NON GUI mode, check system >> configuration and machine configuration (is it the same ?). >> >> >> -- >> >> Regards >> Ubik Load Pack <http://ubikloadpack.com> Team >> Follow us on Twitter <http://twitter.com/ubikloadpack> >> >> >> Cordialement >> L'équipe Ubik Load Pack <http://ubikloadpack.com> >> Suivez-nous sur Twitter <http://twitter.com/ubikloadpack> >> >> >> On Fri, Feb 28, 2014 at 9:14 PM, Jared Griffith >> <[email protected]>wrote: >> >> > Hello, I'm pretty new to jmeter, and I'm seeing some irregularities in >> test >> > results. >> > >> > I am seeing differing results whether or not I use the gui or command >> line >> > from the same machine, and I'm also seeing differing results from the >> > command line on 2 separate machines. I'm using Linux all the way around, >> > Java 1.7.0_40, Jmeter 2.11 on both systems. The only difference between >> > the systems is that one is a CentOS server (no gui, no browser, etc) and >> > one is an Ubuntu workstation. But even on the workstation, I get >> different >> > results between the gui and cli for the same tests. >> > >> > Is this normal or is there something that I am missing? >> > >> > -- >> > >> > Jared Griffith >> > Linux Administrator, PICS Auditing, LLC >> > P: (949) 936-4574 >> > C: (909) 653-7814 >> > >> > <http://www.picsauditing.com> >> > >> > 17701 Cowan #140 | Irvine, CA | 92614 >> > >> > Join PICS on LinkedIn and Twitter! >> > >> > <https://twitter.com/PICSAuditingLLC> >> > >> >> >> >> <http://twitter.com/ubikloadpack> >> > > > > -- > > Jared Griffith > Linux Administrator, PICS Auditing, LLC > P: (949) 936-4574 > C: (909) 653-7814 > > <http://www.picsauditing.com> > > 17701 Cowan #140 | Irvine, CA | 92614 > > Join PICS on LinkedIn and Twitter! > > <https://twitter.com/PICSAuditingLLC> --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
