lenboo commented on issue #2112: [BUG] many dependent tasks depend on a node, task status bug URL: https://github.com/apache/incubator-dolphinscheduler/issues/2112#issuecomment-596397760 > Before submitting the failure task, it has been verified whether it is a retry task. > So, in the process of submitting the task, there is no need to check the failure status again. > > In business design, the retry task is processed first, then the failure strategy is processed. > > MasterExecThread.java > task.taskCanRetry() > > ``` > // node fails, retry first, and then execute the failure process > if(task.getState().typeIsFailure()){ > if(task.getState() == ExecutionStatus.NEED_FAULT_TOLERANCE){ > this.recoverToleranceFaultTaskList.add(task); > } > if(task.taskCanRetry()){ > addTaskToStandByList(task); > }else{ > completeTaskList.put(task.getName(), task); > if( task.getTaskType().equals(TaskType.CONDITIONS.toString()) || > haveConditionsAfterNode(task.getName())) { > submitPostNode(task.getName()); > ``` yes, you are right. there is a check when add retry tasks to standby list , but before submitting to run, we should check the failure strategy again when there have some other tasks failed. because other tasks would be failed between standby and submitted.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
