Examples with HeartbeatServices org.apache.flink.runtime.heartbeat.HeartbeatServices used on opensource projects

Example 6 with HeartbeatServices

use of org.apache.flink.runtime.heartbeat.HeartbeatServices in project flink by apache.

the class JobMasterTest method testRequestPartitionState.

/**
 * Tests the {@link JobMaster#requestPartitionState(IntermediateDataSetID, ResultPartitionID)}
 * call for a finished result partition.
 */
@Test
public void testRequestPartitionState() throws Exception {
    final JobGraph producerConsumerJobGraph = producerConsumerJobGraph();
    final JobMaster jobMaster = new JobMasterBuilder(producerConsumerJobGraph, rpcService).withConfiguration(configuration).withHighAvailabilityServices(haServices).withHeartbeatServices(heartbeatServices).createJobMaster();
    jobMaster.start();
    try {
        final CompletableFuture<TaskDeploymentDescriptor> tddFuture = new CompletableFuture<>();
        final TestingTaskExecutorGateway testingTaskExecutorGateway = new TestingTaskExecutorGatewayBuilder().setSubmitTaskConsumer((taskDeploymentDescriptor, jobMasterId) -> {
            tddFuture.complete(taskDeploymentDescriptor);
            return CompletableFuture.completedFuture(Acknowledge.get());
        }).createTestingTaskExecutorGateway();
        final LocalUnresolvedTaskManagerLocation taskManagerLocation = new LocalUnresolvedTaskManagerLocation();
        final JobMasterGateway jobMasterGateway = jobMaster.getSelfGateway(JobMasterGateway.class);
        final Collection<SlotOffer> slotOffers = registerSlotsAtJobMaster(1, jobMasterGateway, producerConsumerJobGraph.getJobID(), testingTaskExecutorGateway, taskManagerLocation);
        assertThat(slotOffers, hasSize(1));
        // obtain tdd for the result partition ids
        final TaskDeploymentDescriptor tdd = tddFuture.get();
        assertThat(tdd.getProducedPartitions(), hasSize(1));
        final ResultPartitionDeploymentDescriptor partition = tdd.getProducedPartitions().iterator().next();
        final ExecutionAttemptID executionAttemptId = tdd.getExecutionAttemptId();
        final ExecutionAttemptID copiedExecutionAttemptId = new ExecutionAttemptID(executionAttemptId);
        // finish the producer task
        jobMasterGateway.updateTaskExecutionState(new TaskExecutionState(executionAttemptId, ExecutionState.FINISHED)).get();
        // request the state of the result partition of the producer
        final ResultPartitionID partitionId = new ResultPartitionID(partition.getPartitionId(), copiedExecutionAttemptId);
        CompletableFuture<ExecutionState> partitionStateFuture = jobMasterGateway.requestPartitionState(partition.getResultId(), partitionId);
        assertThat(partitionStateFuture.get(), equalTo(ExecutionState.FINISHED));
        // ask for unknown result partition
        partitionStateFuture = jobMasterGateway.requestPartitionState(partition.getResultId(), new ResultPartitionID());
        try {
            partitionStateFuture.get();
            fail("Expected failure.");
        } catch (ExecutionException e) {
            assertThat(ExceptionUtils.findThrowable(e, IllegalArgumentException.class).isPresent(), is(true));
        }
        // ask for wrong intermediate data set id
        partitionStateFuture = jobMasterGateway.requestPartitionState(new IntermediateDataSetID(), partitionId);
        try {
            partitionStateFuture.get();
            fail("Expected failure.");
        } catch (ExecutionException e) {
            assertThat(ExceptionUtils.findThrowable(e, IllegalArgumentException.class).isPresent(), is(true));
        }
        // ask for "old" execution
        partitionStateFuture = jobMasterGateway.requestPartitionState(partition.getResultId(), new ResultPartitionID(partition.getPartitionId(), new ExecutionAttemptID()));
        try {
            partitionStateFuture.get();
            fail("Expected failure.");
        } catch (ExecutionException e) {
            assertThat(ExceptionUtils.findThrowable(e, PartitionProducerDisposedException.class).isPresent(), is(true));
        }
    } finally {
        RpcUtils.terminateRpcEndpoint(jobMaster, testingTimeout);
    }
}

Example 7 with HeartbeatServices

use of org.apache.flink.runtime.heartbeat.HeartbeatServices in project flink by apache.

the class JobMasterTest method testTaskExecutorNotReleasedOnFailedAllocationIfPartitionIsAllocated.

@Test
public void testTaskExecutorNotReleasedOnFailedAllocationIfPartitionIsAllocated() throws Exception {
    final JobManagerSharedServices jobManagerSharedServices = new TestingJobManagerSharedServicesBuilder().build();
    final JobGraph jobGraph = JobGraphTestUtils.singleNoOpJobGraph();
    final LocalUnresolvedTaskManagerLocation taskManagerUnresolvedLocation = new LocalUnresolvedTaskManagerLocation();
    final AtomicBoolean isTrackingPartitions = new AtomicBoolean(true);
    final TestingJobMasterPartitionTracker partitionTracker = new TestingJobMasterPartitionTracker();
    partitionTracker.setIsTrackingPartitionsForFunction(ignored -> isTrackingPartitions.get());
    final JobMaster jobMaster = new JobMasterBuilder(jobGraph, rpcService).withConfiguration(configuration).withHighAvailabilityServices(haServices).withJobManagerSharedServices(jobManagerSharedServices).withHeartbeatServices(heartbeatServices).withPartitionTrackerFactory(ignored -> partitionTracker).createJobMaster();
    final CompletableFuture<JobID> disconnectTaskExecutorFuture = new CompletableFuture<>();
    final CompletableFuture<AllocationID> freedSlotFuture = new CompletableFuture<>();
    final TestingTaskExecutorGateway testingTaskExecutorGateway = new TestingTaskExecutorGatewayBuilder().setFreeSlotFunction((allocationID, throwable) -> {
        freedSlotFuture.complete(allocationID);
        return CompletableFuture.completedFuture(Acknowledge.get());
    }).setDisconnectJobManagerConsumer((jobID, throwable) -> disconnectTaskExecutorFuture.complete(jobID)).createTestingTaskExecutorGateway();
    try {
        jobMaster.start();
        final JobMasterGateway jobMasterGateway = jobMaster.getSelfGateway(JobMasterGateway.class);
        final Collection<SlotOffer> slotOffers = registerSlotsAtJobMaster(1, jobMasterGateway, jobGraph.getJobID(), testingTaskExecutorGateway, taskManagerUnresolvedLocation);
        // check that we accepted the offered slot
        assertThat(slotOffers, hasSize(1));
        final AllocationID allocationId = slotOffers.iterator().next().getAllocationId();
        jobMasterGateway.failSlot(taskManagerUnresolvedLocation.getResourceID(), allocationId, new FlinkException("Fail allocation test exception"));
        // we should free the slot, but not disconnect from the TaskExecutor as we still have an
        // allocated partition
        assertThat(freedSlotFuture.get(), equalTo(allocationId));
        // trigger some request to guarantee ensure the slotAllocationFailure processing if
        // complete
        jobMasterGateway.requestJobStatus(Time.seconds(5)).get();
        assertThat(disconnectTaskExecutorFuture.isDone(), is(false));
    } finally {
        RpcUtils.terminateRpcEndpoint(jobMaster, testingTimeout);
    }
}

Example 8 with HeartbeatServices

use of org.apache.flink.runtime.heartbeat.HeartbeatServices in project flink by apache.

the class TaskExecutorITCase method testSlotAllocation.

@Test
public void testSlotAllocation() throws Exception {
    TestingFatalErrorHandler testingFatalErrorHandler = new TestingFatalErrorHandler();
    TestingHighAvailabilityServices testingHAServices = new TestingHighAvailabilityServices();
    final Configuration configuration = new Configuration();
    final ScheduledExecutorService scheduledExecutorService = new ScheduledThreadPoolExecutor(1);
    final ResourceID taskManagerResourceId = new ResourceID("foobar");
    final UUID rmLeaderId = UUID.randomUUID();
    final TestingLeaderElectionService rmLeaderElectionService = new TestingLeaderElectionService();
    final TestingLeaderRetrievalService rmLeaderRetrievalService = new TestingLeaderRetrievalService();
    final String rmAddress = "rm";
    final String jmAddress = "jm";
    final UUID jmLeaderId = UUID.randomUUID();
    final JobID jobId = new JobID();
    final ResourceProfile resourceProfile = new ResourceProfile(1.0, 1);
    testingHAServices.setResourceManagerLeaderElectionService(rmLeaderElectionService);
    testingHAServices.setResourceManagerLeaderRetriever(rmLeaderRetrievalService);
    testingHAServices.setJobMasterLeaderRetriever(jobId, new TestingLeaderRetrievalService(jmAddress, jmLeaderId));
    TestingSerialRpcService rpcService = new TestingSerialRpcService();
    ResourceManagerConfiguration resourceManagerConfiguration = new ResourceManagerConfiguration(Time.milliseconds(500L), Time.milliseconds(500L), Time.minutes(5L));
    SlotManagerFactory slotManagerFactory = new DefaultSlotManager.Factory();
    JobLeaderIdService jobLeaderIdService = new JobLeaderIdService(testingHAServices, rpcService.getScheduledExecutor(), resourceManagerConfiguration.getJobTimeout());
    MetricRegistry metricRegistry = mock(MetricRegistry.class);
    HeartbeatServices heartbeatServices = mock(HeartbeatServices.class, RETURNS_MOCKS);
    final TaskManagerConfiguration taskManagerConfiguration = TaskManagerConfiguration.fromConfiguration(configuration);
    final TaskManagerLocation taskManagerLocation = new TaskManagerLocation(taskManagerResourceId, InetAddress.getLocalHost(), 1234);
    final MemoryManager memoryManager = mock(MemoryManager.class);
    final IOManager ioManager = mock(IOManager.class);
    final NetworkEnvironment networkEnvironment = mock(NetworkEnvironment.class);
    final TaskManagerMetricGroup taskManagerMetricGroup = mock(TaskManagerMetricGroup.class);
    final BroadcastVariableManager broadcastVariableManager = mock(BroadcastVariableManager.class);
    final FileCache fileCache = mock(FileCache.class);
    final TaskSlotTable taskSlotTable = new TaskSlotTable(Arrays.asList(resourceProfile), new TimerService<AllocationID>(scheduledExecutorService, 100L));
    final JobManagerTable jobManagerTable = new JobManagerTable();
    final JobLeaderService jobLeaderService = new JobLeaderService(taskManagerLocation);
    ResourceManager<ResourceID> resourceManager = new StandaloneResourceManager(rpcService, resourceManagerConfiguration, testingHAServices, slotManagerFactory, metricRegistry, jobLeaderIdService, testingFatalErrorHandler);
    TaskExecutor taskExecutor = new TaskExecutor(taskManagerConfiguration, taskManagerLocation, rpcService, memoryManager, ioManager, networkEnvironment, testingHAServices, heartbeatServices, metricRegistry, taskManagerMetricGroup, broadcastVariableManager, fileCache, taskSlotTable, jobManagerTable, jobLeaderService, testingFatalErrorHandler);
    JobMasterGateway jmGateway = mock(JobMasterGateway.class);
    when(jmGateway.registerTaskManager(any(String.class), any(TaskManagerLocation.class), eq(jmLeaderId), any(Time.class))).thenReturn(FlinkCompletableFuture.<RegistrationResponse>completed(new JMTMRegistrationSuccess(taskManagerResourceId, 1234)));
    when(jmGateway.getHostname()).thenReturn(jmAddress);
    rpcService.registerGateway(rmAddress, resourceManager.getSelf());
    rpcService.registerGateway(jmAddress, jmGateway);
    final AllocationID allocationId = new AllocationID();
    final SlotRequest slotRequest = new SlotRequest(jobId, allocationId, resourceProfile);
    final SlotOffer slotOffer = new SlotOffer(allocationId, 0, resourceProfile);
    try {
        resourceManager.start();
        taskExecutor.start();
        // notify the RM that it is the leader
        rmLeaderElectionService.isLeader(rmLeaderId);
        // notify the TM about the new RM leader
        rmLeaderRetrievalService.notifyListener(rmAddress, rmLeaderId);
        Future<RegistrationResponse> registrationResponseFuture = resourceManager.registerJobManager(rmLeaderId, jmLeaderId, jmAddress, jobId);
        RegistrationResponse registrationResponse = registrationResponseFuture.get();
        assertTrue(registrationResponse instanceof JobMasterRegistrationSuccess);
        resourceManager.requestSlot(jmLeaderId, rmLeaderId, slotRequest);
        verify(jmGateway).offerSlots(eq(taskManagerResourceId), (Iterable<SlotOffer>) argThat(Matchers.contains(slotOffer)), eq(jmLeaderId), any(Time.class));
    } finally {
        if (testingFatalErrorHandler.hasExceptionOccurred()) {
            testingFatalErrorHandler.rethrowError();
        }
    }
}

Example 9 with HeartbeatServices

use of org.apache.flink.runtime.heartbeat.HeartbeatServices in project flink by apache.

the class JobMasterTest method testHeartbeatTimeoutWithTaskManager.

@Test
public void testHeartbeatTimeoutWithTaskManager() throws Exception {
    final TestingHighAvailabilityServices haServices = new TestingHighAvailabilityServices();
    final TestingLeaderRetrievalService rmLeaderRetrievalService = new TestingLeaderRetrievalService();
    haServices.setResourceManagerLeaderRetriever(rmLeaderRetrievalService);
    haServices.setCheckpointRecoveryFactory(mock(CheckpointRecoveryFactory.class));
    final TestingFatalErrorHandler testingFatalErrorHandler = new TestingFatalErrorHandler();
    final String jobManagerAddress = "jm";
    final UUID jmLeaderId = UUID.randomUUID();
    final ResourceID jmResourceId = new ResourceID(jobManagerAddress);
    final String taskManagerAddress = "tm";
    final ResourceID tmResourceId = new ResourceID(taskManagerAddress);
    final TaskManagerLocation taskManagerLocation = new TaskManagerLocation(tmResourceId, InetAddress.getLoopbackAddress(), 1234);
    final TaskExecutorGateway taskExecutorGateway = mock(TaskExecutorGateway.class);
    final TestingSerialRpcService rpc = new TestingSerialRpcService();
    rpc.registerGateway(taskManagerAddress, taskExecutorGateway);
    final long heartbeatInterval = 1L;
    final long heartbeatTimeout = 5L;
    final ScheduledExecutor scheduledExecutor = mock(ScheduledExecutor.class);
    final HeartbeatServices heartbeatServices = new TestingHeartbeatServices(heartbeatInterval, heartbeatTimeout, scheduledExecutor);
    final JobGraph jobGraph = new JobGraph();
    try {
        final JobMaster jobMaster = new JobMaster(jmResourceId, jobGraph, new Configuration(), rpc, haServices, heartbeatServices, Executors.newScheduledThreadPool(1), mock(BlobLibraryCacheManager.class), mock(RestartStrategyFactory.class), Time.of(10, TimeUnit.SECONDS), null, mock(OnCompletionActions.class), testingFatalErrorHandler, new FlinkUserCodeClassLoader(new URL[0]));
        // also start the heartbeat manager in job manager
        jobMaster.start(jmLeaderId);
        // register task manager will trigger monitoring heartbeat target, schedule heartbeat request in interval time
        jobMaster.registerTaskManager(taskManagerAddress, taskManagerLocation, jmLeaderId);
        ArgumentCaptor<Runnable> heartbeatRunnableCaptor = ArgumentCaptor.forClass(Runnable.class);
        verify(scheduledExecutor, times(1)).scheduleAtFixedRate(heartbeatRunnableCaptor.capture(), eq(0L), eq(heartbeatInterval), eq(TimeUnit.MILLISECONDS));
        Runnable heartbeatRunnable = heartbeatRunnableCaptor.getValue();
        ArgumentCaptor<Runnable> timeoutRunnableCaptor = ArgumentCaptor.forClass(Runnable.class);
        verify(scheduledExecutor).schedule(timeoutRunnableCaptor.capture(), eq(heartbeatTimeout), eq(TimeUnit.MILLISECONDS));
        Runnable timeoutRunnable = timeoutRunnableCaptor.getValue();
        // run the first heartbeat request
        heartbeatRunnable.run();
        verify(taskExecutorGateway, times(1)).heartbeatFromJobManager(eq(jmResourceId));
        // run the timeout runnable to simulate a heartbeat timeout
        timeoutRunnable.run();
        verify(taskExecutorGateway).disconnectJobManager(eq(jobGraph.getJobID()), any(TimeoutException.class));
        // check if a concurrent error occurred
        testingFatalErrorHandler.rethrowError();
    } finally {
        rpc.stopService();
    }
}

Example 10 with HeartbeatServices

use of org.apache.flink.runtime.heartbeat.HeartbeatServices in project flink by apache.

the class JobMasterTest method testTaskManagerRegistrationTriggersHeartbeating.

@Test
public void testTaskManagerRegistrationTriggersHeartbeating() throws Exception {
    final CompletableFuture<ResourceID> heartbeatResourceIdFuture = new CompletableFuture<>();
    final UnresolvedTaskManagerLocation unresolvedTaskManagerLocation = new LocalUnresolvedTaskManagerLocation();
    final TestingTaskExecutorGateway taskExecutorGateway = new TestingTaskExecutorGatewayBuilder().setHeartbeatJobManagerFunction((taskManagerId, ignored) -> {
        heartbeatResourceIdFuture.complete(taskManagerId);
        return FutureUtils.completedVoidFuture();
    }).createTestingTaskExecutorGateway();
    rpcService.registerGateway(taskExecutorGateway.getAddress(), taskExecutorGateway);
    final JobMaster jobMaster = new JobMasterBuilder(jobGraph, rpcService).withResourceId(jmResourceId).withConfiguration(configuration).withHighAvailabilityServices(haServices).withHeartbeatServices(new HeartbeatServices(1L, 10000L)).createJobMaster();
    jobMaster.start();
    try {
        final JobMasterGateway jobMasterGateway = jobMaster.getSelfGateway(JobMasterGateway.class);
        // register task manager will trigger monitor heartbeat target, schedule heartbeat
        // request at interval time
        CompletableFuture<RegistrationResponse> registrationResponse = jobMasterGateway.registerTaskManager(jobGraph.getJobID(), TaskManagerRegistrationInformation.create(taskExecutorGateway.getAddress(), unresolvedTaskManagerLocation, TestingUtils.zeroUUID()), testingTimeout);
        // wait for the completion of the registration
        registrationResponse.get();
        assertThat(heartbeatResourceIdFuture.join(), anyOf(nullValue(), equalTo(jmResourceId)));
    } finally {
        RpcUtils.terminateRpcEndpoint(jobMaster, testingTimeout);
    }
}