Examples with TMSlotRequestRejected org.apache.flink.runtime.resourcemanager.messages.taskexecutor.TMSlotRequestRejected used on opensource projects

Example 1 with TMSlotRequestRejected

use of org.apache.flink.runtime.resourcemanager.messages.taskexecutor.TMSlotRequestRejected in project flink by apache.

the class SlotManager method sendSlotRequest.

private void sendSlotRequest(final ResourceSlot freeSlot, final SlotRequest slotRequest) {
    final AllocationID allocationID = slotRequest.getAllocationId();
    final TaskExecutorRegistration registration = freeSlot.getTaskExecutorRegistration();
    final Future<TMSlotRequestReply> slotRequestReplyFuture = registration.getTaskExecutorGateway().requestSlot(freeSlot.getSlotId(), slotRequest.getJobId(), allocationID, // TODO: set proper JM address
    "foobar", rmServices.getLeaderID(), timeout);
    slotRequestReplyFuture.handleAsync(new BiFunction<TMSlotRequestReply, Throwable, Void>() {

        @Override
        public Void apply(TMSlotRequestReply slotRequestReply, Throwable throwable) {
            TaskExecutorRegistration current = taskManagers.get(slotRequestReply.getResourceID());
            if (current != null && current.getInstanceID().equals(slotRequestReply.getInstanceID())) {
                if (throwable != null || slotRequestReply instanceof TMSlotRequestRejected) {
                    handleSlotRequestFailedAtTaskManager(slotRequest, freeSlot.getSlotId());
                } else {
                    LOG.debug("Successfully registered slot {} ", freeSlot.getSlotId());
                }
            } else {
                LOG.debug("Discarding message from obsolete TaskExecutor with InstanceID {}", slotRequestReply.getInstanceID());
            }
            return null;
        }
    }, rmServices.getMainThreadExecutor());
}

Example 2 with TMSlotRequestRejected

use of org.apache.flink.runtime.resourcemanager.messages.taskexecutor.TMSlotRequestRejected in project flink by apache.

the class TaskExecutorTest method testRejectAllocationRequestsForOutOfSyncSlots.

/**
	 * Tests that all allocation requests for slots are ignored if the slot has been reported as
	 * free by the TaskExecutor but this report hasn't been confirmed by the ResourceManager.
	 *
	 * This is essential for the correctness of the state of the ResourceManager.
	 */
@Ignore
@Test
public void testRejectAllocationRequestsForOutOfSyncSlots() throws Exception {
    final ResourceID resourceID = ResourceID.generate();
    final String address1 = "/resource/manager/address/one";
    final UUID leaderId = UUID.randomUUID();
    final JobID jobId = new JobID();
    final String jobManagerAddress = "foobar";
    final TestingSerialRpcService rpc = new TestingSerialRpcService();
    try {
        // register the mock resource manager gateways
        ResourceManagerGateway rmGateway1 = mock(ResourceManagerGateway.class);
        rpc.registerGateway(address1, rmGateway1);
        TestingLeaderRetrievalService testLeaderService = new TestingLeaderRetrievalService();
        TestingHighAvailabilityServices haServices = new TestingHighAvailabilityServices();
        haServices.setResourceManagerLeaderRetriever(testLeaderService);
        TaskManagerConfiguration taskManagerServicesConfiguration = mock(TaskManagerConfiguration.class);
        when(taskManagerServicesConfiguration.getNumberSlots()).thenReturn(1);
        TaskManagerLocation taskManagerLocation = mock(TaskManagerLocation.class);
        when(taskManagerLocation.getResourceID()).thenReturn(resourceID);
        final TestingFatalErrorHandler testingFatalErrorHandler = new TestingFatalErrorHandler();
        TaskExecutor taskManager = new TaskExecutor(taskManagerServicesConfiguration, taskManagerLocation, rpc, mock(MemoryManager.class), mock(IOManager.class), mock(NetworkEnvironment.class), haServices, mock(HeartbeatServices.class, RETURNS_MOCKS), mock(MetricRegistry.class), mock(TaskManagerMetricGroup.class), mock(BroadcastVariableManager.class), mock(FileCache.class), mock(TaskSlotTable.class), mock(JobManagerTable.class), mock(JobLeaderService.class), testingFatalErrorHandler);
        taskManager.start();
        String taskManagerAddress = taskManager.getAddress();
        // no connection initially, since there is no leader
        assertNull(taskManager.getResourceManagerConnection());
        // define a leader and see that a registration happens
        testLeaderService.notifyListener(address1, leaderId);
        verify(rmGateway1).registerTaskExecutor(eq(leaderId), eq(taskManagerAddress), eq(resourceID), any(SlotReport.class), any(Time.class));
        assertNotNull(taskManager.getResourceManagerConnection());
        // test that allocating a slot works
        final SlotID slotID = new SlotID(resourceID, 0);
        TMSlotRequestReply tmSlotRequestReply = taskManager.requestSlot(slotID, jobId, new AllocationID(), jobManagerAddress, leaderId);
        assertTrue(tmSlotRequestReply instanceof TMSlotRequestRegistered);
        // TODO: Figure out the concrete allocation behaviour between RM and TM. Maybe we don't need the SlotID...
        // test that we can't allocate slots which are blacklisted due to pending confirmation of the RM
        final SlotID unconfirmedFreeSlotID = new SlotID(resourceID, 1);
        TMSlotRequestReply tmSlotRequestReply2 = taskManager.requestSlot(unconfirmedFreeSlotID, jobId, new AllocationID(), jobManagerAddress, leaderId);
        assertTrue(tmSlotRequestReply2 instanceof TMSlotRequestRejected);
        // re-register
        verify(rmGateway1).registerTaskExecutor(eq(leaderId), eq(taskManagerAddress), eq(resourceID), any(SlotReport.class), any(Time.class));
        testLeaderService.notifyListener(address1, leaderId);
        // now we should be successful because the slots status has been synced
        // test that we can't allocate slots which are blacklisted due to pending confirmation of the RM
        TMSlotRequestReply tmSlotRequestReply3 = taskManager.requestSlot(unconfirmedFreeSlotID, jobId, new AllocationID(), jobManagerAddress, leaderId);
        assertTrue(tmSlotRequestReply3 instanceof TMSlotRequestRegistered);
        // check if a concurrent error occurred
        testingFatalErrorHandler.rethrowError();
    } finally {
        rpc.stopService();
    }
}