Rational Performance Tester Out Of Memory Error
Contents |
test on agent machine 1233998; RPT; out of memory; agent; performance tester Technote (troubleshooting) Problem(Abstract) This technote explains how to resolve an error message is received stating that the agent machines failed to create a thread: retval -1073741830, errno 11 are out of memory, when running a test with a large number of users
Java.lang.outofmemoryerror: Failed To Create A Thread
through agent machines in IBM Rational Performance Tester (RPT). Resolving the problem On remote computers running Microsoft Windows XP, Windows 2003 Server, Red Hat Enterprise Linux, or SuSE Linux, the RPT_DEFAULT_MEMORY_SIZE property is automatically set to 75% of total memory size after the first schedule run. Although this value is arbitrary, it is intended to provide the best execution environment for your computer. On other remote computers, this property is not automatically set. Therefore, you are somewhat more likely to receive an "out of memory" error. If you receive an out-of-memory error when you run a test or schedule, increase the memory allocation for that remote computer. To do this, set the RPT_VMARGS property, which overrides RPT_DEFAULT_MEMORY_SIZE. To increase the memory allocation on a remote computer: In the Test Navigator (from your local computer), expand the project until you find the deployment location that you want to change. Deployment locations are represented by the icon. Right-click the deployment location, and then click Open. In the deployment location, click the Attributes tab, and then click Add. In the New Property window: In the Property Name field, enter RPT_VMARGS. In the Operator field, confirm that the operator is =. In the Property Value field, enter -Xmxnnnnm, where nnnn is the amount of memory, in megabytes, and then click OK. The following New Property window sets maximum heap to 1900 megabytes: NOTE: Do not set the maximum heap higher than the physical memory that is available on the agent machine. A good rule of thumb is to set the maximum heap to 50% - 75% of the total memory on the agent machine. If you have increased the available memory and you still receive out-of-memory errors, add more remote computers for your user groups. Document information More support for: Rational Performance Tester Agent Controller Software version: 6.1.0.0, 6.1.0.1, 6.1.1.0, 6.1.2.0 Operating system(s): Windows Reference #: 1233998 Modified date: 27 October 2010 Site availability Site assistance Contact and feedback Need support? Submit feedback to IBM Support 1-800-IBM-7378 (USA) Directory of worldwide contacts Contact Privacy Terms of use Accessibility
localhost RPT; memory; allocation; JVM; Playback; Execute; localhost; agent Technote (FAQ) Question How much memory is allocated to the JVM when executing a Test or Schedule on the localhost in IBM Rational Performance Tester? Answer Prior to the release of Rational Performance Tester 8.x the default memory allocated http://www-01.ibm.com/support/docview.wss?uid=swg21233998 when running back a test or schedule locally was 256 MB. This paradigm has been changed in version 8.x. The playback JVM's maximum heap size when running a Test is set to 1/2 of installed memory (i.e. on a computer with 2 GB http://www-01.ibm.com/support/docview.wss?uid=swg21404126 of RAM, the heap would be set to 1 GB). For schedule playback, the maximum heap size is set to 3/4 of installed memory. Just as in prior versions, if you run Schedule using a remote agent controller (for example a new location) you can set a specific memory value using the VMARGS property for the remote agent. See the help topic Increasing memory allocation in the Rational Performance Test Information Center. Document information More support for: Rational Performance Tester Agent Controller Software version: 8.0, 8.0.0.1, 8.0.0.2, 8.0.0.3, 8.1, 8.1.0.1 Operating system(s): Windows Reference #: 1404126 Modified date: 27 October 2010 Site availability Site assistance Contact and feedback Need support? Submit feedback to IBM Support 1-800-IBM-7378 (USA) Directory of worldwide contacts Contact Privacy Terms of use Accessibility
How to find the memory leakage by using graphs in RPT? 6 replies Latest Post - 2009-11-16T09:21:21Z by E.D. Display:ConversationsBy Date 1-7 of 7 Previous Next arunthomas 270002BGHF 52 Posts Pinned topic How to find the memory leakage by https://www.ibm.com/developerworks/forums/thread.jspa?threadID=279100 using graphs in RPT? 2009-10-12T04:53:20Z | Tags: Answered question This question has been answered. Unanswered question This question has not been answered yet. Hi All, I am using RPT 8.0 for performance testing. I am feeling that there is some memory leakage in our application. I need to find which transaction causes the memory leakage. Which graphs/counters I should use to find the exacta reason for failed to memory leakage. Please share your experiences Thanks in advance. Arun Log in to reply. Updated on 2009-11-16T09:21:21Z at 2009-11-16T09:21:21Z by E.D. jamesb 120000QREP 21 Posts Re: How to find the memory leakage by using graphs in RPT? 2009-10-12T08:30:32Z This is the accepted answer. This is the accepted answer. Is this a java application? In RPT you can monitor the remote servers memory usage with rstatd failed to create or windows monitors but of course if your heap size stays the same you will probably see nothing useful. A soak test ran from RPT will determine if you eventually get out of memory errors and do have a memory leak. I'd turn on verbose:gc in the JVM before running the test so you can determine if it is a native memory leak or heap leak and some info on how bad it is. Remember there will be a continuous small increase in memory useage as the JVM JITs frequently used methods - this will level off and stop in a soak test. In short others may have tips and things to try in RPT but I would not use it to determine if there is a leak beyond running the soak test. Finding the exact reason for a leak is a difficult activity and you need good knowledge of java and the code running - verbose:gc will give you the graphs to analyse to see if there is a leak - look for PMAT in alphaworks or use the IBM support tool garbage collector analyser. Maybe cutting down your test to identify what transactions/pages in yo
be down. Please try the request again. Your cache administrator is webmaster. Generated Tue, 25 Oct 2016 22:02:53 GMT by s_wx1085 (squid/3.5.20)