Stopping the test at the beginning of the test is only feasible when the disk space is very close to the threshold right when you start the test. However, most often the disk space constraint incident happens during long running tests(test that run over 1 hour). During a long running test, typically a system would have enough space at the beginning, but over the course of the test the available disk space shrinks. IMO, in order to make this a comprehensive solution, listeners which write should dynamically keep checking available disk space during runtime and stop writing results when the available space is close to the threshold.
At the start of the test, the default space threshold can probably be useful to warn the users before starting the test if the available space is close to the threshold. Thoughts? Thanks Chaitanya M Bhatt http://www.performancecompetence.com On Wed, May 13, 2015 at 2:05 PM, Philippe Mouawad < [email protected]> wrote: > My proposal is to stop the test at start of test. > Would your comments still apply ? > > Thanks > > On Wed, May 13, 2015 at 11:01 PM, chaitanya bhatt < > [email protected] > > wrote: > > > Instead of stopping the test, I would recommend that Jmeter stop > recording > > the results. There are many users who collect performance metrics from > APM > > tools like AppDynamics,Dynatrace, Wily etc. Proactively stopping the test > > would be a bad idea for such users. These users do not rely on > performance > > metrics from Jmeter's JTL files. > > > > Jmeter should ideally stop recording logs or results upon reaching the > disk > > space threshold thereby conserving space. The tool should just be > > generating load under this circumstance. > > > > Thoughts? > > > > Thanks > > Chaitanya M Bhatt > > http://www.performancecompetence.com > > > > On Wed, May 13, 2015 at 12:53 PM, Philippe Mouawad < > > [email protected]> wrote: > > > > > Hello, > > > > > > What do you think of the following feature. > > > > > > Sometimes it happened to me to have a test fail at end due to not > having > > > enough disk space , so I lost the end of my test results. > > > > > > > > > I was thinking of the following way to fix it : > > > > > > - Create a proprerty which contains the disk space required > estimation > > > - At start of test, listeners that need to write file check using > > File > > > or FileStore (JDK7) > > > - If value is lower than expected size, they trigger end of test > > > > > > > > > Do you think this is over engeneering ? or bad idea ? > > > > > > > > > Thanks > > > > > > Regards > > > > > > > > > -- > Cordialement. > Philippe Mouawad. >
