I double checked my script and I definitely had no listeners running. The script itself was about 1.7 megs in size, which doesn't seem that unreasonable compared to other tests that we have created. I had about 6 or 7 regular expression extractors, but I can't imagine that that would have a huge impact on performance.
GOOD NEWS! After tweaking some settings I was able to get it to run and I have successfully kicked off a load test with 75 users without running out of memory on the Jmeter side. Since I had already increased the Heap size to 1024, the next thing I tried was changing the EVACUATION variable to EVACUATION="-XX:MaxLiveObjectEvacuationRatio=40%" (originally this was 20%) Somehow this seemed to have done the trick, as Jmeter no longer runs out of memory resources. I would love some clarification as to what this variable (and any of the other variables Scott mentioned) actually does if anyone is more familiar with those settings. Thanks to all for your help and suggestions, and thank you Peter for your very detailed explanation of the effect of HTTPS on performance! Elli -----Original Message----- From: Scott Pack [mailto:[EMAIL PROTECTED] Sent: Wednesday, May 11, 2005 12:09 PM To: JMeter Users List; Peter Lin Subject: RE: Out of memory error HTTPS testability issues aside, the question remains for me - is there a failure in JMeter to properly maintain/release objects, Graph Results in particular, or is this simply a response data issue? How do the NEW, SURVIVOR, TENURING, EVACUATION and PERM options affect memory management and garbage collection? It's not particularly clear, at least to me. Scott -----Original Message----- From: Peter Lin [mailto:[EMAIL PROTECTED] Sent: Wednesday, May 11, 2005 9:58 AM To: JMeter Users List Subject: Re: Out of memory error You're right. it could be a problem with his test plan, but without seeing the test plan it's hard to know. Even if his test plan is correct, testing SSL/HTTPS is still very hard to do. Unless the hardware resources are available, it's just very hard to simulate HTTPS traffic effectively and accurately. peter --------------------------------------------------------------------- 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]
