use of org.apache.beam.runners.dataflow.worker.util.common.worker.Operation in project beam by apache.
the class BeamFnMapTaskExecutorFactory method createOperationTransformForGrpcPortNodes.
private Function<Node, Node> createOperationTransformForGrpcPortNodes(final Network<Node, Edge> network, final FnDataService beamFnDataService, final OperationContext context) {
return new TypeSafeNodeFunction<RemoteGrpcPortNode>(RemoteGrpcPortNode.class) {
@Override
public Node typedApply(RemoteGrpcPortNode input) {
RegisterAndProcessBundleOperation registerFnOperation = (RegisterAndProcessBundleOperation) Iterables.getOnlyElement(Iterables.filter(network.adjacentNodes(input), OperationNode.class)).getOperation();
// The coder comes from the one and only adjacent output node
Coder<?> coder = Iterables.getOnlyElement(Iterables.filter(network.adjacentNodes(input), OutputReceiverNode.class)).getCoder();
// We figure out whether we are outputting some where if the output node is a
// successor.
Iterable<OutputReceiverNode> outputReceiverNodes = Iterables.filter(network.successors(input), OutputReceiverNode.class);
Operation operation;
if (outputReceiverNodes.iterator().hasNext()) {
OutputReceiver[] outputReceivers = new OutputReceiver[] { Iterables.getOnlyElement(outputReceiverNodes).getOutputReceiver() };
operation = new RemoteGrpcPortReadOperation<>(beamFnDataService, input.getPrimitiveTransformId(), registerFnOperation::getProcessBundleInstructionId, (Coder) coder, outputReceivers, context);
} else {
operation = new RemoteGrpcPortWriteOperation<>(beamFnDataService, input.getPrimitiveTransformId(), registerFnOperation::getProcessBundleInstructionId, (Coder) coder, context);
}
return OperationNode.create(operation);
}
};
}
use of org.apache.beam.runners.dataflow.worker.util.common.worker.Operation in project beam by apache.
the class BeamFnMapTaskExecutorFactory method create.
/**
* Creates a new {@link DataflowMapTaskExecutor} from the given {@link MapTask} definition using
* the provided {@link ReaderFactory}.
*/
@Override
public DataflowMapTaskExecutor create(InstructionRequestHandler instructionRequestHandler, GrpcFnServer<GrpcDataService> grpcDataFnServer, Endpoints.ApiServiceDescriptor dataApiServiceDescriptor, GrpcFnServer<GrpcStateService> grpcStateFnServer, MutableNetwork<Node, Edge> network, PipelineOptions options, String stageName, ReaderFactory readerFactory, SinkFactory sinkFactory, DataflowExecutionContext<?> executionContext, CounterSet counterSet, IdGenerator idGenerator) {
// TODO: remove this once we trust the code paths
checkArgument(DataflowRunner.hasExperiment(options.as(DataflowPipelineDebugOptions.class), "beam_fn_api"), "%s should only be used when beam_fn_api is enabled", getClass().getSimpleName());
// Swap out all the InstructionOutput nodes with OutputReceiver nodes
Networks.replaceDirectedNetworkNodes(network, createOutputReceiversTransform(stageName, counterSet));
if (DataflowRunner.hasExperiment(options.as(DataflowPipelineDebugOptions.class), "use_executable_stage_bundle_execution")) {
LOG.debug("Using SingleEnvironmentInstanceJobBundleFactory");
JobBundleFactory jobBundleFactory = SingleEnvironmentInstanceJobBundleFactory.create(StaticRemoteEnvironmentFactory.forService(instructionRequestHandler), grpcDataFnServer, grpcStateFnServer, idGenerator);
// If the use_executable_stage_bundle_execution is enabled, use ExecutableStage instead.
Networks.replaceDirectedNetworkNodes(network, createOperationTransformForExecutableStageNode(network, stageName, executionContext, jobBundleFactory));
} else {
// Swap out all the RegisterFnRequest nodes with Operation nodes
Networks.replaceDirectedNetworkNodes(network, createOperationTransformForRegisterFnNodes(idGenerator, instructionRequestHandler, grpcStateFnServer.getService(), stageName, executionContext));
// Swap out all the RemoteGrpcPort nodes with Operation nodes, note that it is expected
// that the RegisterFnRequest nodes have already been replaced.
Networks.replaceDirectedNetworkNodes(network, createOperationTransformForGrpcPortNodes(network, grpcDataFnServer.getService(), // TODO: Set NameContext properly for these operations.
executionContext.createOperationContext(NameContext.create(stageName, stageName, stageName, stageName))));
}
// Swap out all the FetchAndFilterStreamingSideInput nodes with operation nodes
Networks.replaceDirectedNetworkNodes(network, createOperationTransformForFetchAndFilterStreamingSideInputNodes(network, idGenerator, instructionRequestHandler, grpcDataFnServer.getService(), dataApiServiceDescriptor, executionContext, stageName));
// Swap out all the ParallelInstruction nodes with Operation nodes
Networks.replaceDirectedNetworkNodes(network, createOperationTransformForParallelInstructionNodes(stageName, network, options, readerFactory, sinkFactory, executionContext));
// Collect all the operations within the network and attach all the operations as receivers
// to preceding output receivers.
List<Operation> topoSortedOperations = new ArrayList<>();
for (OperationNode node : Iterables.filter(Networks.topologicalOrder(network), OperationNode.class)) {
topoSortedOperations.add(node.getOperation());
for (Node predecessor : Iterables.filter(network.predecessors(node), OutputReceiverNode.class)) {
((OutputReceiverNode) predecessor).getOutputReceiver().addOutput((Receiver) node.getOperation());
}
}
if (LOG.isDebugEnabled()) {
LOG.info("Map task network: {}", Networks.toDot(network));
}
return BeamFnMapTaskExecutor.withSharedCounterSet(topoSortedOperations, counterSet, executionContext.getExecutionStateTracker());
}
use of org.apache.beam.runners.dataflow.worker.util.common.worker.Operation in project beam by apache.
the class IntrinsicMapTaskExecutorTest method testExceptionInFinishAbortsAllOperations.
@Test
public void testExceptionInFinishAbortsAllOperations() throws Exception {
Operation o1 = Mockito.mock(Operation.class);
Operation o2 = Mockito.mock(Operation.class);
Operation o3 = Mockito.mock(Operation.class);
Mockito.doThrow(new Exception("in finish")).when(o2).finish();
ExecutionStateTracker stateTracker = ExecutionStateTracker.newForTest();
try (IntrinsicMapTaskExecutor executor = IntrinsicMapTaskExecutor.withSharedCounterSet(Arrays.<Operation>asList(o1, o2, o3), counterSet, stateTracker)) {
executor.execute();
fail("Should have thrown");
} catch (Exception e) {
InOrder inOrder = Mockito.inOrder(o1, o2, o3);
inOrder.verify(o3).start();
inOrder.verify(o2).start();
inOrder.verify(o1).start();
inOrder.verify(o1).finish();
inOrder.verify(o2).finish();
// Order of abort doesn't matter
Mockito.verify(o1).abort();
Mockito.verify(o2).abort();
Mockito.verify(o3).abort();
Mockito.verifyNoMoreInteractions(o1, o2, o3);
}
}
use of org.apache.beam.runners.dataflow.worker.util.common.worker.Operation in project beam by apache.
the class IntrinsicMapTaskExecutorTest method testValidOperations.
@Test
public void testValidOperations() throws Exception {
TestOutputReceiver receiver = new TestOutputReceiver(counterSet, NameContextsForTests.nameContextForTest());
List<Operation> operations = Arrays.<Operation>asList(new TestReadOperation(receiver, createContext("ReadOperation")));
ExecutionStateTracker stateTracker = ExecutionStateTracker.newForTest();
try (IntrinsicMapTaskExecutor executor = IntrinsicMapTaskExecutor.withSharedCounterSet(operations, counterSet, stateTracker)) {
Assert.assertEquals(operations.get(0), executor.getReadOperation());
}
}
use of org.apache.beam.runners.dataflow.worker.util.common.worker.Operation in project beam by apache.
the class IntrinsicMapTaskExecutorTest method testGetMetricContainers.
@Test
@SuppressWarnings("unchecked")
public /**
* This test makes sure that any metrics reported within an operation are part of the metric
* containers returned by {@link getMetricContainers}.
*/
void testGetMetricContainers() throws Exception {
ExecutionStateTracker stateTracker = new DataflowExecutionStateTracker(ExecutionStateSampler.newForTest(), new TestDataflowExecutionState(NameContext.forStage("testStage"), "other", null, /* requestingStepName */
null, /* sideInputIndex */
null, /* metricsContainer */
NoopProfileScope.NOOP), new CounterSet(), PipelineOptionsFactory.create(), "test-work-item-id");
final String o1 = "o1";
TestOperationContext context1 = createContext(o1, stateTracker);
final String o2 = "o2";
TestOperationContext context2 = createContext(o2, stateTracker);
final String o3 = "o3";
TestOperationContext context3 = createContext(o3, stateTracker);
List<Operation> operations = Arrays.asList(new Operation(new OutputReceiver[] {}, context1) {
@Override
public void start() throws Exception {
super.start();
try (Closeable scope = context.enterStart()) {
Metrics.counter("TestMetric", "MetricCounter").inc(1L);
}
}
}, new Operation(new OutputReceiver[] {}, context2) {
@Override
public void start() throws Exception {
super.start();
try (Closeable scope = context.enterStart()) {
Metrics.counter("TestMetric", "MetricCounter").inc(2L);
}
}
}, new Operation(new OutputReceiver[] {}, context3) {
@Override
public void start() throws Exception {
super.start();
try (Closeable scope = context.enterStart()) {
Metrics.counter("TestMetric", "MetricCounter").inc(3L);
}
}
});
try (IntrinsicMapTaskExecutor executor = IntrinsicMapTaskExecutor.withSharedCounterSet(operations, counterSet, stateTracker)) {
// Call execute so that we run all the counters
executor.execute();
assertThat(context1.metricsContainer().getUpdates().counterUpdates(), contains(metricUpdate("TestMetric", "MetricCounter", o1, 1L)));
assertThat(context2.metricsContainer().getUpdates().counterUpdates(), contains(metricUpdate("TestMetric", "MetricCounter", o2, 2L)));
assertThat(context3.metricsContainer().getUpdates().counterUpdates(), contains(metricUpdate("TestMetric", "MetricCounter", o3, 3L)));
}
}
Aggregations