Search in sources :

Example 6 with MetricsFieldSetter

use of co.cask.cdap.internal.app.runtime.MetricsFieldSetter in project cdap by caskdata.

the class MapReduceProgramRunner method run.

@Override
public ProgramController run(final Program program, ProgramOptions options) {
    // Extract and verify parameters
    ApplicationSpecification appSpec = program.getApplicationSpecification();
    Preconditions.checkNotNull(appSpec, "Missing application specification.");
    ProgramType processorType = program.getType();
    Preconditions.checkNotNull(processorType, "Missing processor type.");
    Preconditions.checkArgument(processorType == ProgramType.MAPREDUCE, "Only MAPREDUCE process type is supported.");
    MapReduceSpecification spec = appSpec.getMapReduce().get(program.getName());
    Preconditions.checkNotNull(spec, "Missing MapReduceSpecification for %s", program.getName());
    Arguments arguments = options.getArguments();
    RunId runId = ProgramRunners.getRunId(options);
    WorkflowProgramInfo workflowInfo = WorkflowProgramInfo.create(arguments);
    DatasetFramework programDatasetFramework = workflowInfo == null ? datasetFramework : NameMappedDatasetFramework.createFromWorkflowProgramInfo(datasetFramework, workflowInfo, appSpec);
    // Setup dataset framework context, if required
    if (programDatasetFramework instanceof ProgramContextAware) {
        ProgramId programId = program.getId();
        ((ProgramContextAware) programDatasetFramework).setContext(new BasicProgramContext(programId.run(runId)));
    }
    MapReduce mapReduce;
    try {
        mapReduce = new InstantiatorFactory(false).get(TypeToken.of(program.<MapReduce>getMainClass())).create();
    } catch (Exception e) {
        LOG.error("Failed to instantiate MapReduce class for {}", spec.getClassName(), e);
        throw Throwables.propagate(e);
    }
    // List of all Closeable resources that needs to be cleanup
    List<Closeable> closeables = new ArrayList<>();
    try {
        PluginInstantiator pluginInstantiator = createPluginInstantiator(options, program.getClassLoader());
        if (pluginInstantiator != null) {
            closeables.add(pluginInstantiator);
        }
        final BasicMapReduceContext context = new BasicMapReduceContext(program, options, cConf, spec, workflowInfo, discoveryServiceClient, metricsCollectionService, txSystemClient, programDatasetFramework, streamAdmin, getPluginArchive(options), pluginInstantiator, secureStore, secureStoreManager, messagingService);
        Reflections.visit(mapReduce, mapReduce.getClass(), new PropertyFieldSetter(context.getSpecification().getProperties()), new MetricsFieldSetter(context.getMetrics()), new DataSetFieldSetter(context));
        // note: this sets logging context on the thread level
        LoggingContextAccessor.setLoggingContext(context.getLoggingContext());
        // Set the job queue to hConf if it is provided
        Configuration hConf = new Configuration(this.hConf);
        String schedulerQueue = options.getArguments().getOption(Constants.AppFabric.APP_SCHEDULER_QUEUE);
        if (schedulerQueue != null && !schedulerQueue.isEmpty()) {
            hConf.set(JobContext.QUEUE_NAME, schedulerQueue);
        }
        Service mapReduceRuntimeService = new MapReduceRuntimeService(injector, cConf, hConf, mapReduce, spec, context, program.getJarLocation(), locationFactory, streamAdmin, txSystemClient, authorizationEnforcer, authenticationContext);
        mapReduceRuntimeService.addListener(createRuntimeServiceListener(program.getId(), runId, closeables, arguments, options.getUserArguments()), Threads.SAME_THREAD_EXECUTOR);
        final ProgramController controller = new MapReduceProgramController(mapReduceRuntimeService, context);
        LOG.debug("Starting MapReduce Job: {}", context);
        // be running the job, but the data directory will be owned by cdap.
        if (MapReduceTaskContextProvider.isLocal(hConf) || UserGroupInformation.isSecurityEnabled()) {
            mapReduceRuntimeService.start();
        } else {
            ProgramRunners.startAsUser(cConf.get(Constants.CFG_HDFS_USER), mapReduceRuntimeService);
        }
        return controller;
    } catch (Exception e) {
        closeAllQuietly(closeables);
        throw Throwables.propagate(e);
    }
}
Also used : ApplicationSpecification(co.cask.cdap.api.app.ApplicationSpecification) CConfiguration(co.cask.cdap.common.conf.CConfiguration) Configuration(org.apache.hadoop.conf.Configuration) Closeable(java.io.Closeable) ArrayList(java.util.ArrayList) MapReduce(co.cask.cdap.api.mapreduce.MapReduce) NameMappedDatasetFramework(co.cask.cdap.internal.app.runtime.workflow.NameMappedDatasetFramework) DatasetFramework(co.cask.cdap.data2.dataset2.DatasetFramework) InstantiatorFactory(co.cask.cdap.common.lang.InstantiatorFactory) MetricsFieldSetter(co.cask.cdap.internal.app.runtime.MetricsFieldSetter) ProgramType(co.cask.cdap.proto.ProgramType) RunId(org.apache.twill.api.RunId) ProgramController(co.cask.cdap.app.runtime.ProgramController) MapReduceSpecification(co.cask.cdap.api.mapreduce.MapReduceSpecification) Arguments(co.cask.cdap.app.runtime.Arguments) MessagingService(co.cask.cdap.messaging.MessagingService) MetricsCollectionService(co.cask.cdap.api.metrics.MetricsCollectionService) Service(com.google.common.util.concurrent.Service) ProgramId(co.cask.cdap.proto.id.ProgramId) BasicProgramContext(co.cask.cdap.internal.app.runtime.BasicProgramContext) DataSetFieldSetter(co.cask.cdap.internal.app.runtime.DataSetFieldSetter) PropertyFieldSetter(co.cask.cdap.common.lang.PropertyFieldSetter) WorkflowProgramInfo(co.cask.cdap.internal.app.runtime.workflow.WorkflowProgramInfo) PluginInstantiator(co.cask.cdap.internal.app.runtime.plugin.PluginInstantiator) ProgramContextAware(co.cask.cdap.data.ProgramContextAware)

Example 7 with MetricsFieldSetter

use of co.cask.cdap.internal.app.runtime.MetricsFieldSetter in project cdap by caskdata.

the class MapperWrapper method run.

@SuppressWarnings("unchecked")
@Override
public void run(Context context) throws IOException, InterruptedException {
    MapReduceClassLoader classLoader = MapReduceClassLoader.getFromConfiguration(context.getConfiguration());
    ClassLoader weakReferenceClassLoader = new WeakReferenceDelegatorClassLoader(classLoader);
    BasicMapReduceTaskContext basicMapReduceContext = classLoader.getTaskContextProvider().get(context);
    String program = basicMapReduceContext.getProgramName();
    // this is a hook for periodic flushing of changes buffered by datasets (to avoid OOME)
    WrappedMapper.Context flushingContext = createAutoFlushingContext(context, basicMapReduceContext);
    basicMapReduceContext.setHadoopContext(flushingContext);
    InputSplit inputSplit = context.getInputSplit();
    if (inputSplit instanceof MultiInputTaggedSplit) {
        basicMapReduceContext.setInputContext(InputContexts.create((MultiInputTaggedSplit) inputSplit));
    }
    ClassLoader programClassLoader = classLoader.getProgramClassLoader();
    Mapper delegate = createMapperInstance(programClassLoader, getWrappedMapper(context.getConfiguration()), context, program);
    // injecting runtime components, like datasets, etc.
    try {
        Reflections.visit(delegate, delegate.getClass(), new PropertyFieldSetter(basicMapReduceContext.getSpecification().getProperties()), new MetricsFieldSetter(basicMapReduceContext.getMetrics()), new DataSetFieldSetter(basicMapReduceContext));
    } catch (Throwable t) {
        Throwable rootCause = Throwables.getRootCause(t);
        USERLOG.error("Failed to initialize program '{}' with error: {}. Please check the system logs for more details.", program, rootCause.getMessage(), rootCause);
        throw new IOException(String.format("Failed to inject fields to %s", delegate.getClass()), t);
    }
    ClassLoader oldClassLoader;
    if (delegate instanceof ProgramLifecycle) {
        oldClassLoader = ClassLoaders.setContextClassLoader(weakReferenceClassLoader);
        try {
            ((ProgramLifecycle) delegate).initialize(new MapReduceLifecycleContext(basicMapReduceContext));
        } catch (Exception e) {
            Throwable rootCause = Throwables.getRootCause(e);
            USERLOG.error("Failed to initialize program '{}' with error: {}. Please check the system logs for more " + "details.", program, rootCause.getMessage(), rootCause);
            throw new IOException(String.format("Failed to initialize mapper with %s", basicMapReduceContext), e);
        } finally {
            ClassLoaders.setContextClassLoader(oldClassLoader);
        }
    }
    oldClassLoader = ClassLoaders.setContextClassLoader(weakReferenceClassLoader);
    try {
        delegate.run(flushingContext);
    } finally {
        ClassLoaders.setContextClassLoader(oldClassLoader);
    }
    // memory by tx agent)
    try {
        basicMapReduceContext.flushOperations();
    } catch (Exception e) {
        throw new IOException("Failed to flush operations at the end of mapper of " + basicMapReduceContext, e);
    }
    // Close all writers created by MultipleOutputs
    basicMapReduceContext.closeMultiOutputs();
    if (delegate instanceof ProgramLifecycle) {
        oldClassLoader = ClassLoaders.setContextClassLoader(weakReferenceClassLoader);
        try {
            ((ProgramLifecycle<? extends RuntimeContext>) delegate).destroy();
        } catch (Exception e) {
            LOG.error("Error during destroy of mapper {}", basicMapReduceContext, e);
        // Do nothing, try to finish
        } finally {
            ClassLoaders.setContextClassLoader(oldClassLoader);
        }
    }
}
Also used : ProgramLifecycle(co.cask.cdap.api.ProgramLifecycle) IOException(java.io.IOException) DataSetFieldSetter(co.cask.cdap.internal.app.runtime.DataSetFieldSetter) IOException(java.io.IOException) Mapper(org.apache.hadoop.mapreduce.Mapper) WrappedMapper(org.apache.hadoop.mapreduce.lib.map.WrappedMapper) WeakReferenceDelegatorClassLoader(co.cask.cdap.common.lang.WeakReferenceDelegatorClassLoader) PropertyFieldSetter(co.cask.cdap.common.lang.PropertyFieldSetter) MetricsFieldSetter(co.cask.cdap.internal.app.runtime.MetricsFieldSetter) WrappedMapper(org.apache.hadoop.mapreduce.lib.map.WrappedMapper) MultiInputTaggedSplit(co.cask.cdap.internal.app.runtime.batch.dataset.input.MultiInputTaggedSplit) WeakReferenceDelegatorClassLoader(co.cask.cdap.common.lang.WeakReferenceDelegatorClassLoader) RuntimeContext(co.cask.cdap.api.RuntimeContext) TaggedInputSplit(co.cask.cdap.internal.app.runtime.batch.dataset.input.TaggedInputSplit) InputSplit(org.apache.hadoop.mapreduce.InputSplit)

Example 8 with MetricsFieldSetter

use of co.cask.cdap.internal.app.runtime.MetricsFieldSetter in project cdap by caskdata.

the class ReducerWrapper method run.

@SuppressWarnings("unchecked")
@Override
public void run(Context context) throws IOException, InterruptedException {
    MapReduceClassLoader classLoader = MapReduceClassLoader.getFromConfiguration(context.getConfiguration());
    ClassLoader weakReferenceClassLoader = new WeakReferenceDelegatorClassLoader(classLoader);
    BasicMapReduceTaskContext basicMapReduceContext = classLoader.getTaskContextProvider().get(context);
    // this is a hook for periodic flushing of changes buffered by datasets (to avoid OOME)
    WrappedReducer.Context flushingContext = createAutoFlushingContext(context, basicMapReduceContext);
    basicMapReduceContext.setHadoopContext(flushingContext);
    String userReducer = context.getConfiguration().get(ATTR_REDUCER_CLASS);
    ClassLoader programClassLoader = classLoader.getProgramClassLoader();
    Reducer delegate = createReducerInstance(programClassLoader, userReducer);
    // injecting runtime components, like datasets, etc.
    try {
        Reflections.visit(delegate, delegate.getClass(), new PropertyFieldSetter(basicMapReduceContext.getSpecification().getProperties()), new MetricsFieldSetter(basicMapReduceContext.getMetrics()), new DataSetFieldSetter(basicMapReduceContext));
    } catch (Throwable t) {
        LOG.error("Failed to inject fields to {}.", delegate.getClass(), t);
        throw Throwables.propagate(t);
    }
    ClassLoader oldClassLoader;
    if (delegate instanceof ProgramLifecycle) {
        oldClassLoader = ClassLoaders.setContextClassLoader(weakReferenceClassLoader);
        try {
            ((ProgramLifecycle) delegate).initialize(new MapReduceLifecycleContext(basicMapReduceContext));
        } catch (Exception e) {
            LOG.error("Failed to initialize reducer with {}", basicMapReduceContext, e);
            throw Throwables.propagate(e);
        } finally {
            ClassLoaders.setContextClassLoader(oldClassLoader);
        }
    }
    oldClassLoader = ClassLoaders.setContextClassLoader(weakReferenceClassLoader);
    try {
        delegate.run(flushingContext);
    } finally {
        ClassLoaders.setContextClassLoader(oldClassLoader);
    }
    // memory by tx agent)
    try {
        basicMapReduceContext.flushOperations();
    } catch (Exception e) {
        LOG.error("Failed to flush operations at the end of reducer of " + basicMapReduceContext, e);
        throw Throwables.propagate(e);
    }
    // Close all writers created by MultipleOutputs
    basicMapReduceContext.closeMultiOutputs();
    if (delegate instanceof ProgramLifecycle) {
        oldClassLoader = ClassLoaders.setContextClassLoader(weakReferenceClassLoader);
        try {
            ((ProgramLifecycle<? extends RuntimeContext>) delegate).destroy();
        } catch (Exception e) {
            LOG.error("Error during destroy of reducer {}", basicMapReduceContext, e);
        // Do nothing, try to finish
        } finally {
            ClassLoaders.setContextClassLoader(oldClassLoader);
        }
    }
}
Also used : ProgramLifecycle(co.cask.cdap.api.ProgramLifecycle) DataSetFieldSetter(co.cask.cdap.internal.app.runtime.DataSetFieldSetter) IOException(java.io.IOException) WeakReferenceDelegatorClassLoader(co.cask.cdap.common.lang.WeakReferenceDelegatorClassLoader) PropertyFieldSetter(co.cask.cdap.common.lang.PropertyFieldSetter) MetricsFieldSetter(co.cask.cdap.internal.app.runtime.MetricsFieldSetter) WeakReferenceDelegatorClassLoader(co.cask.cdap.common.lang.WeakReferenceDelegatorClassLoader) WrappedReducer(org.apache.hadoop.mapreduce.lib.reduce.WrappedReducer) WrappedReducer(org.apache.hadoop.mapreduce.lib.reduce.WrappedReducer) Reducer(org.apache.hadoop.mapreduce.Reducer) RuntimeContext(co.cask.cdap.api.RuntimeContext)

Example 9 with MetricsFieldSetter

use of co.cask.cdap.internal.app.runtime.MetricsFieldSetter in project cdap by caskdata.

the class FlowletProgramRunner method run.

@SuppressWarnings("unchecked")
@Override
public ProgramController run(Program program, ProgramOptions options) {
    BasicFlowletContext flowletContext = null;
    try {
        // Extract and verify parameters
        String flowletName = options.getName();
        int instanceId = Integer.parseInt(options.getArguments().getOption(ProgramOptionConstants.INSTANCE_ID, "-1"));
        Preconditions.checkArgument(instanceId >= 0, "Missing instance Id");
        int instanceCount = Integer.parseInt(options.getArguments().getOption(ProgramOptionConstants.INSTANCES, "0"));
        Preconditions.checkArgument(instanceCount > 0, "Invalid or missing instance count");
        RunId runId = ProgramRunners.getRunId(options);
        ApplicationSpecification appSpec = program.getApplicationSpecification();
        Preconditions.checkNotNull(appSpec, "Missing application specification.");
        ProgramType processorType = program.getType();
        Preconditions.checkNotNull(processorType, "Missing processor type.");
        Preconditions.checkArgument(processorType == ProgramType.FLOW, "Only FLOW process type is supported.");
        String processorName = program.getName();
        Preconditions.checkNotNull(processorName, "Missing processor name.");
        FlowSpecification flowSpec = appSpec.getFlows().get(processorName);
        FlowletDefinition flowletDef = flowSpec.getFlowlets().get(flowletName);
        Preconditions.checkNotNull(flowletDef, "Definition missing for flowlet \"%s\"", flowletName);
        Class<?> clz = Class.forName(flowletDef.getFlowletSpec().getClassName(), true, program.getClassLoader());
        Preconditions.checkArgument(Flowlet.class.isAssignableFrom(clz), "%s is not a Flowlet.", clz);
        // Setup dataset framework context, if required
        ProgramId programId = program.getId();
        FlowletId flowletId = programId.flowlet(flowletName);
        ProgramRunId run = programId.run(runId);
        ProgramContext programContext = new BasicProgramContext(run, flowletId);
        if (dsFramework instanceof ProgramContextAware) {
            ((ProgramContextAware) dsFramework).setContext(programContext);
        }
        Class<? extends Flowlet> flowletClass = (Class<? extends Flowlet>) clz;
        // Creates flowlet context
        flowletContext = new BasicFlowletContext(program, options, flowletId, instanceId, instanceCount, flowletDef.getDatasets(), flowletDef.getFlowletSpec(), metricsCollectionService, discoveryServiceClient, txClient, dsFramework, secureStore, secureStoreManager, messageService, cConf);
        // Creates tx related objects
        DataFabricFacade dataFabricFacade = dataFabricFacadeFactory.create(program, flowletContext.getDatasetCache());
        if (dataFabricFacade instanceof ProgramContextAware) {
            ((ProgramContextAware) dataFabricFacade).setContext(programContext);
        }
        // Creates QueueSpecification
        Table<Node, String, Set<QueueSpecification>> queueSpecs = new SimpleQueueSpecificationGenerator(new ApplicationId(program.getNamespaceId(), program.getApplicationId())).create(flowSpec);
        Flowlet flowlet = new InstantiatorFactory(false).get(TypeToken.of(flowletClass)).create();
        TypeToken<? extends Flowlet> flowletType = TypeToken.of(flowletClass);
        // Set the context classloader to the cdap classloader. It is needed for the DatumWriterFactory be able
        // to load cdap classes
        Thread.currentThread().setContextClassLoader(FlowletProgramRunner.class.getClassLoader());
        // Inject DataSet, OutputEmitter, Metric fields
        ImmutableList.Builder<ProducerSupplier> queueProducerSupplierBuilder = ImmutableList.builder();
        Reflections.visit(flowlet, flowlet.getClass(), new PropertyFieldSetter(flowletDef.getFlowletSpec().getProperties()), new DataSetFieldSetter(flowletContext), new MetricsFieldSetter(flowletContext.getMetrics()), new OutputEmitterFieldSetter(outputEmitterFactory(flowletContext, flowletName, dataFabricFacade, queueProducerSupplierBuilder, queueSpecs)));
        ImmutableList.Builder<ConsumerSupplier<?>> queueConsumerSupplierBuilder = ImmutableList.builder();
        Collection<ProcessSpecification<?>> processSpecs = createProcessSpecification(flowletContext, flowletType, processMethodFactory(flowlet), processSpecificationFactory(flowletContext, dataFabricFacade, queueReaderFactory, flowletName, queueSpecs, queueConsumerSupplierBuilder, createSchemaCache(program)), Lists.<ProcessSpecification<?>>newLinkedList());
        List<ConsumerSupplier<?>> consumerSuppliers = queueConsumerSupplierBuilder.build();
        // Create the flowlet driver
        AtomicReference<FlowletProgramController> controllerRef = new AtomicReference<>();
        Service serviceHook = createServiceHook(flowletName, consumerSuppliers, controllerRef);
        FlowletRuntimeService driver = new FlowletRuntimeService(flowlet, flowletContext, processSpecs, createCallback(flowlet, flowletDef.getFlowletSpec()), dataFabricFacade, serviceHook);
        FlowletProgramController controller = new FlowletProgramController(program.getId(), flowletName, flowletContext, driver, queueProducerSupplierBuilder.build(), consumerSuppliers);
        controllerRef.set(controller);
        LOG.info("Starting flowlet: {}", flowletContext);
        driver.start();
        LOG.info("Flowlet started: {}", flowletContext);
        return controller;
    } catch (Exception e) {
        // of the flowlet context.
        if (flowletContext != null) {
            flowletContext.close();
        }
        throw Throwables.propagate(e);
    }
}
Also used : ApplicationSpecification(co.cask.cdap.api.app.ApplicationSpecification) Set(java.util.Set) ImmutableSet(com.google.common.collect.ImmutableSet) ImmutableList(com.google.common.collect.ImmutableList) Node(co.cask.cdap.app.queue.QueueSpecificationGenerator.Node) BasicProgramContext(co.cask.cdap.internal.app.runtime.BasicProgramContext) ProgramContext(co.cask.cdap.data.ProgramContext) InstantiatorFactory(co.cask.cdap.common.lang.InstantiatorFactory) FlowletDefinition(co.cask.cdap.api.flow.FlowletDefinition) FlowSpecification(co.cask.cdap.api.flow.FlowSpecification) MetricsFieldSetter(co.cask.cdap.internal.app.runtime.MetricsFieldSetter) ProgramType(co.cask.cdap.proto.ProgramType) RunId(org.apache.twill.api.RunId) ProgramRunId(co.cask.cdap.proto.id.ProgramRunId) Flowlet(co.cask.cdap.api.flow.flowlet.Flowlet) MetricsCollectionService(co.cask.cdap.api.metrics.MetricsCollectionService) Service(com.google.common.util.concurrent.Service) MessagingService(co.cask.cdap.messaging.MessagingService) AbstractService(com.google.common.util.concurrent.AbstractService) AtomicReference(java.util.concurrent.atomic.AtomicReference) ProgramId(co.cask.cdap.proto.id.ProgramId) BasicProgramContext(co.cask.cdap.internal.app.runtime.BasicProgramContext) DataSetFieldSetter(co.cask.cdap.internal.app.runtime.DataSetFieldSetter) UnsupportedTypeException(co.cask.cdap.api.data.schema.UnsupportedTypeException) IOException(java.io.IOException) SimpleQueueSpecificationGenerator(co.cask.cdap.internal.app.queue.SimpleQueueSpecificationGenerator) PropertyFieldSetter(co.cask.cdap.common.lang.PropertyFieldSetter) FlowletId(co.cask.cdap.proto.id.FlowletId) ProgramRunId(co.cask.cdap.proto.id.ProgramRunId) DataFabricFacade(co.cask.cdap.internal.app.runtime.DataFabricFacade) ApplicationId(co.cask.cdap.proto.id.ApplicationId) ProgramContextAware(co.cask.cdap.data.ProgramContextAware)

Aggregations

MetricsFieldSetter (co.cask.cdap.internal.app.runtime.MetricsFieldSetter)9 PropertyFieldSetter (co.cask.cdap.common.lang.PropertyFieldSetter)8 DataSetFieldSetter (co.cask.cdap.internal.app.runtime.DataSetFieldSetter)7 InstantiatorFactory (co.cask.cdap.common.lang.InstantiatorFactory)4 ProgramLifecycle (co.cask.cdap.api.ProgramLifecycle)3 IOException (java.io.IOException)3 RuntimeContext (co.cask.cdap.api.RuntimeContext)2 TransactionControl (co.cask.cdap.api.annotation.TransactionControl)2 ApplicationSpecification (co.cask.cdap.api.app.ApplicationSpecification)2 MetricsCollectionService (co.cask.cdap.api.metrics.MetricsCollectionService)2 CConfiguration (co.cask.cdap.common.conf.CConfiguration)2 WeakReferenceDelegatorClassLoader (co.cask.cdap.common.lang.WeakReferenceDelegatorClassLoader)2 ProgramContextAware (co.cask.cdap.data.ProgramContextAware)2 BasicProgramContext (co.cask.cdap.internal.app.runtime.BasicProgramContext)2 MessagingService (co.cask.cdap.messaging.MessagingService)2 ProgramType (co.cask.cdap.proto.ProgramType)2 ProgramId (co.cask.cdap.proto.id.ProgramId)2 Service (com.google.common.util.concurrent.Service)2 ArrayList (java.util.ArrayList)2 Configuration (org.apache.hadoop.conf.Configuration)2