Red Hat Bugzilla – Bug 1028139
Metric Availability auto-refresh calls are not being cleaned up properly (and grow)
Last modified: 2014-01-02 15:38:48 EST
Description of problem:
Availability calls increase due to the number of resources clicked in the resource tree.
The additional of state (to remember open charts) to the view keeps the browser GC from cleaning up the multiple threads. When we change resources and create a new view, there is still a link the momento object that stores the opened charts and that linkage keeps the view from being GC'd. Without the link, the view is GC'd normally and the auto-refresh activity associated with the view is destroyed with view. But when the view is not destroyed and new ones are created each on has an auto-refresh associated with it.
For instance, clicking on 10 different resources leaves 10 auto-refresh threads.
Version-Release number of selected component (if applicable): 3.2.0 ER05
Steps to Reproduce:
1. Go to Monitoring --> Metrics tab
2. click in the resource tree to change resources
3. Go to firebug or browser console and examine how many times availability is called
Availability calls should not increase with changing resource links
Should only have 1 availability call no matter what
Workaround: refresh the browser
I had this fix in my branch mtho11/bz1014571 but that didn't make it in due to stability issues. So I'm extracting the relevant bits to place into master now.
Created factory methods for view creation that ensures singleton instance.
Committed to master: cee05ed
Committed to release/jon3.2.x as commit: 2c51610
Additional changes as I found it was not complete. Also removed some of the overly verbose logging to the browser console.
commit to master: e0cbede
commit to release/jon3.2.x: 979b91a
Moving to ON_QA as available for testing with new brew build.
Mass moving all of these from ER6 to target milestone ER07 since the ER6 build was bad and QE was halted for the same reason.
Build Number :