Description
To support fault tolerant, we need to track task states and ensure the correct/allowed state transitions.
Here are the task states:
TaskNew,
TaskSubmitting,
TaskRunning,
TaskCompeleted,
TaskWaitingForClose,
TaskClosedByDriver,
TaskFailedByEvaluatorFailure,
TaskFailedByGroupCommunication,
TaskFailedAppError,
TaskFailedSystemError
Here are the events taht trigger the state transition:
SubmittingTask,
RunningTask,
CompleteTask,
WaitingTaskToClose,
CloseTask,
FailTaskAppError,
FailTaskSystemError,
FaileTaskEvaluatorError,
FailTaskCommuError
The state transition diagram is defined the the Fault Tolerant design doc. It is attached.