Summary: | Summariser : The + (difference) reports show wrong elapsed time and throughput | ||
---|---|---|---|
Product: | JMeter - Now in Github | Reporter: | UbikLoadPack support <support> |
Component: | Main | Assignee: | JMeter issues mailing list <issues> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | p.mouawad |
Priority: | P2 | ||
Version: | 2.12 | ||
Target Milestone: | --- | ||
Hardware: | All | ||
OS: | All | ||
Attachments: | Test plan showing issue |
Description
UbikLoadPack support
2014-12-12 14:40:13 UTC
Hi, One additional info, the report really happens every 10 seconds but the in "XXXX" is wrong as long as throughput which uses it. Regards @ubikloadpack I cannot reproduce this. Does the test use any 3rd party add-ons? Can you provide a simple test case that shows the issue? Hi, When reported we were using jmeter plugins. But we made new tests without it and issue persists. I wonder if the problem is not the following: - What if a sampler takes more than the configured interval of Summariser to run or even worst what if TransactionController is used and contained such Samplers. I think current code of RunningSample would lead to what is happening as firstTime would be very far in the past.; I will try to build a simple test plan showing issue. Regards @ubikloadpack Created attachment 32324 [details]
Test plan showing issue
OK, I see the problem now. At present the times are taken from the samples; they should really be taken from the current time. This applies to both total types, though the problem is unlikely to be noticed with the overall total. At present, the overall total does not take into account that the start time might be later than the start of the actual test. Also that the end-time might be later than the end of the last sample (there may be initial and final time delays). Further notes: The DELTA RunningSample start time needs to be reset each time, but the TOTAL RunningSample start time needs to remain as the start of the test. For both, the end time needs to be the time when the samples are merged/printed. The RunningSample class was created a long while before the Summariser. It only now seems to be used for the Summariser. Previously it was used for the StatVisualizer; possibly other classes. The code may still be useful as it is, so probably best to create a new (local) class for handling the accumulation. URL: http://svn.apache.org/r1648146 Log: Summariser : The + (difference) reports show wrong elapsed time and throughput Bugzilla Id: 57346 Added: jmeter/trunk/src/core/org/apache/jmeter/reporters/SummariserRunningSample.java Modified: jmeter/trunk/src/core/org/apache/jmeter/reporters/Summariser.java jmeter/trunk/xdocs/changes.xml Date: Sun Dec 28 13:58:42 2014 New Revision: 1648204 URL: http://svn.apache.org/r1648204 Log: BUG 57346 Summariser : The + (difference) reports show wrong elapsed time and throughput Remove useless index field Add javadocs Add Apache License Header Bugzilla Id: 57346 Modified: jmeter/trunk/src/core/org/apache/jmeter/reporters/SummariserRunningSample.java Date: Sun Dec 28 13:59:38 2014 New Revision: 1648205 URL: http://svn.apache.org/r1648205 Log: BUG 57346 Summariser : The + (difference) reports show wrong elapsed time and throughput Impact removal of useless index field in SummariserRunningSample Add javadocs Rename some variables to make code clear Bugzilla Id: 57346 Modified: jmeter/trunk/src/core/org/apache/jmeter/reporters/Summariser.java Hi, We tested with nightly build and it is working fine. Thanks a lot for the fix. And happy new year 2015 to the Apache JMeter team. Regards @ubikloadpack This issue has been migrated to GitHub: https://github.com/apache/jmeter/issues/3498 |