On 08/05/2009, drubix <andrew.schr...@gmail.com> wrote: > > Sorry for the double post, edits don't get sent to the mailing list. > > I've removed to the Response Assertion in case that was the cause of the > large memory use (it's the only "non-standard" component I'm using) but that > doesn't change the memory usage. > > I've also tried running the test on a Windows box (which is running Sun's > JDK instead of OpenJDK) with approximately the same results. > > Is this memory usage really standard? If that's the case, with a more > complex test plan and a higher number of users, I'd be looking at a server > farm just to simulate a couple of dozen users.
I've certainly run HTTP tests with lots of users and not had such a problem. Which samplers are you using? > Thanks, > > Drew > > > > drubix wrote: > > > > Hi Sebb, > > > > I do actually, just ran it when I got into work with a single instance and > > its memory usage started at about 22MB and seemed to be growing at about > > 1MB every 2 seconds or so until a plateau at about 132MB after 15mins. > > This test had a single thread and was launched using the command: > > > > jmeter -n -t SimpleTest.jmx -l out.jtl > > > > It consists of mostly HTTPClient Samplers, a loop controller and a > > Response Assertion (but no listeners). Does this memory usage seem > > unusually high? > > > > Thanks, > > > > Drew > > > > > > sebb-2-2 wrote: > >> > >> On 07/05/2009, drubix <andrew.schr...@gmail.com> wrote: > >>> > >>> Hi, > >>> > >>> I'm running a distributed test all on one system (to take advantage of > >>> IP > >>> aliasing) with one thread per instance of JMeter. When I watch the > >>> processes with top on my RH Linux box, the memory usage of each JMeter > >>> instance seems to be rising dramatically over time. The same number of > >>> threads running for 5 mins uses around 1.5GB of memory total by the end > >>> of > >>> the run but when running for 30 mins ends up using closer to 4GB. > >>> > >>> I'm using the HTTPClient for all calls so that should be avoiding the > >>> HTTPS > >>> memory leak (only using HTTPS during the initial log on step anyway) > >>> and I > >>> have no listeners running (I'm running in non-GUI mode using the -l > >>> switch). > >>> I'm also using the default mode so that each sample is reported to the > >>> master - the slave instances themselves shouldn't really be storing any > >>> more > >>> information than just the cookies, should they? > >>> > >>> Is there any other reason why the memory growth would be as it is? > >> > >> Sorry, no idea how memory usage behaves in client server mode. > >> > >> Do you see the same memory increase if you use independent non-GUI > >> JMeter instances? > >> > >>> Thanks, > >>> > >>> Drew > >>> > >>> -- > >>> View this message in context: > >>> http://www.nabble.com/Increasing-Memory-Usage-tp23421516p23421516.html > >>> Sent from the JMeter - User mailing list archive at Nabble.com. > >>> > >>> > >>> --------------------------------------------------------------------- > >>> To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org > >>> For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org > >>> > >>> > >> > >> --------------------------------------------------------------------- > >> To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org > >> For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org > >> > >> > >> > > > > > > -- > > View this message in context: > http://www.nabble.com/Increasing-Memory-Usage-tp23421516p23438414.html > > Sent from the JMeter - User mailing list archive at Nabble.com. > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org > For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org > > --------------------------------------------------------------------- To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org
