Using JMeter 2.13, Maven and Linux in all servers. Not running any tests from gui or command line, Maven is used!
Have a test plan with a lot of parameters that are set and passed from Maven execution in Jenkins to the JMeter test plan. Using "-J" in Maven to set the input parameters. Works fine in single JMeter environment.
If I start JMeter in slaves manually first, then start the Jenkins job the test plan in the specified remote servers are started properly. But I have observed that no parameters are transferred to the remote servers. So, I can't control the test plan execution. The temporary workaround is to hardcode some parameters in the test plan. But this is not acceptable solution!
I've looked around a lot. Have found these pages, JMeter distributed testing and command line parameters and this
But I'm running from Maven, not command line, so "-G" doesn't work!
Haven't done any RMI setup, but I don't think that's the problem. Probably more related to how parameters should be transferred to remote servers.
With this setting in pom.xml,
<propertiesJMeter>
<remote_hosts>10.71.98.54,10.71.98.82,10.71.98.81</remote_hosts>
</propertiesJMeter>
I did manage to get the basic slave connection to work; I see the printout of remote server IP addresses in the output/log and test plan is started and runs fine and logs seems to be okay too. But, problem is that parameters aren't transferred to remote servers!!
Some additional pom configuration:
<configuration>
<remoteConfig>
<startServersBeforeTests>true</startServersBeforeTests>
<stopServersAfterTests>true</stopServersAfterTests>
</remoteConfig>
<propertiesUser>
<THROUGHPUT>${throughput}</THROUGHPUT>
<NUMBER_OF_LOOPS>${number_of_loops}</NUMBER_OF_LOOPS>
<DURATION>${duration}</DURATION>
<NUMBER_OF_CLIENTS>${number_of_clients}</NUMBER_OF_CLIENTS>
</propertiesUser>
...
Please, can someone help?