In NewAdaptiveFlowStatsCollector.java in the runnables for CalAndShortFlowsTask, MidFlows and LongFlowsTask, callCount*FlowsTask are compared with ENTIRE_POLL_TIMES.
ENTIRE_POLL_TIMES in comments is mentioned to be variable greater or equal than common multople of MID_POLL_TIMES and LONG_POLL_TIMES, so it is not necessarily a multiple of MID_POLL_TIMES and LONG_POLL_TIMES.
The comparison in the runnables should be "greater than or equal to" rather than "equal to".
The current implementation should work well, but if the values are changed then it will not. Hence, this change is required.
|14016,4||[ONOS-6480] Bug fix in NewAdaptiveFlowStatsCollector.java||master||onos||Status: MERGED||+2||+1|