Details
-
Bug
-
Status: Closed
-
Major
-
Resolution: Won't Fix
-
2.0.2
-
None
-
None
Description
The asynchronous Work instances - ie those spawned by WorkManager.startWork/scheduleWork - go off and do whatever processing they do. But there seems a general lack of synchronization of these Work instances with ServiceMix operations like a container shutdown.
I expect the result is that the rug is pulled from underneath the Work instances probably causing them all to fall over.
I don't know what is the solution, but perhaps the Work instances ought to have associated WorkListeners so that any shutdown/stop type processing ServiceMix wants to do can correctly coordinate with the WORK_COMPLETED events.
This might also be a case for changing all the WM.scheduleWork calls to WM.startWork so you at least know the work is accepted in the first place.
See also IRC (12/12)