GJL commented on a change in pull request #9860: [FLINK-14331][runtime] Reset vertices right after they transition to terminated states URL: https://github.com/apache/flink/pull/9860#discussion_r333521037
########## File path: flink-runtime/src/test/java/org/apache/flink/runtime/scheduler/DefaultSchedulerTest.java ########## @@ -254,6 +262,58 @@ public void failJobIfNotEnoughResources() throws Exception { "Could not allocate the required slot within slot request timeout.").isPresent()); } + @Test + public void vertexIsResetBeforeRestarted() throws Exception { + final JobGraph jobGraph = singleNonParallelJobVertexJobGraph(); + + final TestSchedulingStrategy.Factory schedulingStrategyFactory = new TestSchedulingStrategy.Factory(); + final DefaultScheduler scheduler = createScheduler(jobGraph, schedulingStrategyFactory); + final TestSchedulingStrategy schedulingStrategy = schedulingStrategyFactory.getLastCreatedSchedulingStrategy(); + final SchedulingTopology topology = schedulingStrategy.getSchedulingTopology(); + + startScheduling(scheduler); + + final SchedulingExecutionVertex vertex = Iterables.getOnlyElement(topology.getVertices()); + final ExecutionAttemptID attemptId = scheduler.getExecutionVertex(vertex.getId()).getCurrentExecutionAttempt().getAttemptId(); + + schedulingStrategy.schedule(Collections.singleton(vertex.getId())); + + scheduler.updateTaskExecutionState(new TaskExecutionState(jobGraph.getJobID(), attemptId, ExecutionState.FAILED)); + + taskRestartExecutor.triggerScheduledTasks(); + + Set<ExecutionVertexID> verticesToRestart = schedulingStrategy.getReceivedVerticesToRestart(); + assertEquals(1, verticesToRestart.size()); + assertEquals(ExecutionState.CREATED, topology.getVertexOrThrow(verticesToRestart.iterator().next()).getState()); + } + + @Test + public void scheduleOnlyIfVertexIsCreated() throws Exception { + final JobGraph jobGraph = singleNonParallelJobVertexJobGraph(); + + final TestSchedulingStrategy.Factory schedulingStrategyFactory = new TestSchedulingStrategy.Factory(); + final DefaultScheduler scheduler = createScheduler(jobGraph, schedulingStrategyFactory); + final TestSchedulingStrategy schedulingStrategy = schedulingStrategyFactory.getLastCreatedSchedulingStrategy(); + final SchedulingTopology topology = schedulingStrategy.getSchedulingTopology(); + + startScheduling(scheduler); + + final SchedulingExecutionVertex vertex = Iterables.getOnlyElement(topology.getVertices()); + + // The first time scheduling of a vertex will result in one deployment + schedulingStrategy.schedule(Collections.singleton(vertex.getId())); + testTaskManagerGateway.getDeployedExecutionVertices(1, TIMEOUT_MS); + + // The scheduling of a non-CREATED vertex will result in no deployment + schedulingStrategy.schedule(Collections.singleton(vertex.getId())); + try { + testTaskManagerGateway.getDeployedExecutionVertices(1, TIMEOUT_MS); Review comment: It's probably better to use the changes from #9872 later ---------------------------------------------------------------- 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: us...@infra.apache.org With regards, Apache Git Services