On 14/09/2007, Ronan Klyne <[EMAIL PROTECTED]> wrote: > sebb wrote: > > On 13/09/2007, Ronan Klyne <[EMAIL PROTECTED]> wrote: > >> Bj wrote: > >>> Very hungry threads ! > >>> There's obvously a memory leak somewhere. > >>> How long it's running with 5 threads ? > >>> > >> About 10 minutes. > > > > Is that several loops? > > > > If so, that suggests a leak. > > > > Perhaps try adding some timers to the test plan to make it easier to > > see when the memory increases. > > > > If it still runs about as many samples before running out of memory, > > that would tend to confirm a memory leak from one of the test > > elements. > > Yes, it's several loops. Using 4 threads only, I start using 830Mb of > ram, which then climbs by 20Mb every 2-3 minutes. > Any advice on how to find this leak? I don't see what information the > timers will provide for me - what should I do with them?
The timers are useful to slow down the test, so you can see which parts of the test cause the memory usage to increase. Otherwise, try disabling parts of the test to see which has most effect. > Also, advice on how to reduce the initial memory requirement woud be > useful... http://jakarta.apache.org/jmeter/usermanual/best-practices.html > # r > > -- > Ronan Klyne > Business Collaborator Developer > Tel: +44 (0)870 163 2555 > [EMAIL PROTECTED] > www.groupbc.com > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
