[EDGENT-396] JobMonitorApp restarts job 3 times more than it should

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: Apache Edgent 1.2.0
• Component/s: Runtime
• Assignee: Dale LaBossiere
• Reporter: Dale LaBossiere
• Created: 14/Mar/17 19:02
• Updated: 02/Jan/18 19:07
• Resolved: 15/Mar/17 13:24

Description

N.B. I don't think this bug affects IotProvider since I'm pretty sure that while IotProvider includes JobMonitorApp, it doesn't register a JobRegistryService so the monitor does nothing. JIRA forthcoming for that.

JobMonitorAppTest exercises the app but it doesn't perform any validation that restarts were actually happening. Adding instrumentation / validation highlights that 3x the number of rebuilds/restarts are happening.

appOne: buildCnt: 7 injectedFailureCnt: 2
appTwo: buildCnt: 10 injectedFailureCnt: 3

Further investigation identifies the JobMonitorApp's job event filtering as the problem. Each "failed" job ends up with 3 events that pass through the filter

RUNNING, RUNNING, UNHEALTHY
RUNNING, CLOSED, UNHEALTHY
CLOSED, CLOSED, UNHEALTHY

... or something like that