[kepler-dev] [Bug 5095] test kepler and wrp for memory leaks

[bugzilla-daemon at ecoinformatics.org]

http://bugzilla.ecoinformatics.org/show_bug.cgi?id=5095

--- Comment #31 from Aaron <aschultz at nceas.ucsb.edu> 2011-04-21 15:19:02 PDT ---
A standard test I've been using to measure leakiness in the main
KeplerGraphFrame class is to open 101 blank frames then close them all and
measure the difference in memory usage after garbage collections.  Here are
steps for performing this test.

1.) Start Kepler
2.) Start Java VisualVM
3.) Begin monitoring the running org.kepler.Kepler instance in VisualVM
4.) Perform GC on the Monitor tab in VisualVM
5.) Measure the lowest used heap amount just after performing the GC operation
by dragging the cursor over the Heap usage graph (rounded to nearest MB)
6.) Back in Kepler open 101 frames using Ctrl-N
7.) Perform GC on the Monitor tab in VisualVM
8.) Close 100 of the 101 open frames in kepler using the X button in the upper
right corner of the frame
9.) Select Help->About from the menu and click on the splash screen that comes
up
10.) press Ctrl-O and cancel the open frame (step 9 and 10 are just to clear
frame references from focus managers)
9.) Perform GC on the Monitor tab in VisualVM
10.) Measure the lowest used heap amount just after performing the GC operation
by dragging the cursor over the Heap usage graph (rounded to nearest MB)

I then take a snapshot of the graph and name it using the following naming
convention 
<Suite name>-<OS>-<test name>-<description>-<min>to<max>MB.jpg
where
<Suite name> The name of the suite of Kepler modules that was used for the test
(or Vergil if plain old ptolemy was used)
<OS> The operating system the test was run on, including 64bit or 32bit
designator
<test name> A descriptive name for the test, this is 101_blanks for the test
described here
<description> Information about the significance of this particular test.  Such
as a svn revision number or specific issue being tested
<min>to<max>MB are the two measurements of concern, for the 101_blanks test the
<min> comes from step 5 and the <max> comes from step 10

Attached are several of these graphs

-- 
Configure bugmail: http://bugzilla.ecoinformatics.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.