类org.apache.hadoop.mapred.Task源码实例Demo

下面列出了怎么用org.apache.hadoop.mapred.Task的API类实例代码及写法,或者点击链接到github查看源代码。

源代码1 项目: hadoop   文件: ContainerRemoteLaunchEvent.java
public ContainerRemoteLaunchEvent(TaskAttemptId taskAttemptID,
    ContainerLaunchContext containerLaunchContext,
    Container allocatedContainer, Task remoteTask) {
  super(taskAttemptID, allocatedContainer.getId(), StringInterner
    .weakIntern(allocatedContainer.getNodeId().toString()),
    allocatedContainer.getContainerToken(),
    ContainerLauncher.EventType.CONTAINER_REMOTE_LAUNCH);
  this.allocatedContainer = allocatedContainer;
  this.containerLaunchContext = containerLaunchContext;
  this.task = remoteTask;
}
 
源代码2 项目: hadoop   文件: TaskAttemptContextImpl.java
/**
 * Set the current status of the task to the given string.
 */
@Override
public void setStatus(String status) {
  String normalizedStatus = Task.normalizeStatus(status, conf);
  setStatusString(normalizedStatus);
  reporter.setStatus(normalizedStatus);
}
 
源代码3 项目: big-c   文件: ContainerRemoteLaunchEvent.java
public ContainerRemoteLaunchEvent(TaskAttemptId taskAttemptID,
    ContainerLaunchContext containerLaunchContext,
    Container allocatedContainer, Task remoteTask) {
  super(taskAttemptID, allocatedContainer.getId(), StringInterner
    .weakIntern(allocatedContainer.getNodeId().toString()),
    allocatedContainer.getContainerToken(),
    ContainerLauncher.EventType.CONTAINER_REMOTE_LAUNCH);
  this.allocatedContainer = allocatedContainer;
  this.containerLaunchContext = containerLaunchContext;
  this.task = remoteTask;
}
 
源代码4 项目: big-c   文件: TaskAttemptContextImpl.java
/**
 * Set the current status of the task to the given string.
 */
@Override
public void setStatus(String status) {
  String normalizedStatus = Task.normalizeStatus(status, conf);
  setStatusString(normalizedStatus);
  reporter.setStatus(normalizedStatus);
}
 
@Override
public List<HadoopCounterGroup> getCounterGroups() {
	List<HadoopCounterGroup> result = new ArrayList<HadoopCounterGroup>();
	result.add(new HadoopCounterGroup(JobInProgress.Counter.values(), "JobInProgressCounters", null));
	result.add(new HadoopCounterGroup(Task.Counter.values(), "TaskCounters", null));
	//TODO this is temporary solution for adding group for FileSystemCounters enum as the enum class itself cannot be found
	String fsCountersGroupName = "FileSystemCounters";
	List<HadoopCounterKey> fsCounters = new ArrayList<HadoopCounterKey>();
	fsCounters.add(new HadoopCounterKey("FILE_BYTES_READ", fsCountersGroupName));
	fsCounters.add(new HadoopCounterKey("HDFS_BYTES_READ", fsCountersGroupName));
	fsCounters.add(new HadoopCounterKey("FILE_BYTES_WRITTEN", fsCountersGroupName));
	fsCounters.add(new HadoopCounterKey("HDFS_BYTES_WRITTEN", fsCountersGroupName));
	result.add(new HadoopCounterGroup(fsCounters, fsCountersGroupName, null));
	return result;
}
 
源代码6 项目: hadoop   文件: TaskAttemptImpl.java
static ContainerLaunchContext createContainerLaunchContext(
    Map<ApplicationAccessType, String> applicationACLs,
    Configuration conf, Token<JobTokenIdentifier> jobToken, Task remoteTask,
    final org.apache.hadoop.mapred.JobID oldJobId,
    WrappedJvmID jvmID,
    TaskAttemptListener taskAttemptListener,
    Credentials credentials) {

  synchronized (commonContainerSpecLock) {
    if (commonContainerSpec == null) {
      commonContainerSpec = createCommonContainerLaunchContext(
          applicationACLs, conf, jobToken, oldJobId, credentials);
    }
  }

  // Fill in the fields needed per-container that are missing in the common
  // spec.

  boolean userClassesTakesPrecedence =
    conf.getBoolean(MRJobConfig.MAPREDUCE_JOB_USER_CLASSPATH_FIRST, false);

  // Setup environment by cloning from common env.
  Map<String, String> env = commonContainerSpec.getEnvironment();
  Map<String, String> myEnv = new HashMap<String, String>(env.size());
  myEnv.putAll(env);
  if (userClassesTakesPrecedence) {
    myEnv.put(Environment.CLASSPATH_PREPEND_DISTCACHE.name(), "true");
  }
  MapReduceChildJVM.setVMEnv(myEnv, remoteTask);

  // Set up the launch command
  List<String> commands = MapReduceChildJVM.getVMCommand(
      taskAttemptListener.getAddress(), remoteTask, jvmID);

  // Duplicate the ByteBuffers for access by multiple containers.
  Map<String, ByteBuffer> myServiceData = new HashMap<String, ByteBuffer>();
  for (Entry<String, ByteBuffer> entry : commonContainerSpec
              .getServiceData().entrySet()) {
    myServiceData.put(entry.getKey(), entry.getValue().duplicate());
  }

  // Construct the actual Container
  ContainerLaunchContext container = ContainerLaunchContext.newInstance(
      commonContainerSpec.getLocalResources(), myEnv, commands,
      myServiceData, commonContainerSpec.getTokens().duplicate(),
      applicationACLs);

  return container;
}
 
源代码7 项目: hadoop   文件: ContainerRemoteLaunchEvent.java
public Task getRemoteTask() {
  return this.task;
}
 
源代码8 项目: hadoop   文件: TestTaskImpl.java
@Override
protected Task createRemoteTask() {
  return new MockTask(taskType);
}
 
源代码9 项目: hadoop   文件: MergeManagerImpl.java
@Override
public void merge(List<InMemoryMapOutput<K,V>> inputs) throws IOException {
  if (inputs == null || inputs.size() == 0) {
    return;
  }
  
  //name this output file same as the name of the first file that is 
  //there in the current list of inmem files (this is guaranteed to
  //be absent on the disk currently. So we don't overwrite a prev. 
  //created spill). Also we need to create the output file now since
  //it is not guaranteed that this file will be present after merge
  //is called (we delete empty files as soon as we see them
  //in the merge method)

  //figure out the mapId 
  TaskAttemptID mapId = inputs.get(0).getMapId();
  TaskID mapTaskId = mapId.getTaskID();

  List<Segment<K, V>> inMemorySegments = new ArrayList<Segment<K, V>>();
  long mergeOutputSize = 
    createInMemorySegments(inputs, inMemorySegments,0);
  int noInMemorySegments = inMemorySegments.size();

  Path outputPath = 
    mapOutputFile.getInputFileForWrite(mapTaskId,
                                       mergeOutputSize).suffix(
                                           Task.MERGED_OUTPUT_PREFIX);

  FSDataOutputStream out = CryptoUtils.wrapIfNecessary(jobConf, rfs.create(outputPath));
  Writer<K, V> writer = new Writer<K, V>(jobConf, out,
      (Class<K>) jobConf.getMapOutputKeyClass(),
      (Class<V>) jobConf.getMapOutputValueClass(), codec, null, true);

  RawKeyValueIterator rIter = null;
  CompressAwarePath compressAwarePath;
  try {
    LOG.info("Initiating in-memory merge with " + noInMemorySegments + 
             " segments...");
    
    rIter = Merger.merge(jobConf, rfs,
                         (Class<K>)jobConf.getMapOutputKeyClass(),
                         (Class<V>)jobConf.getMapOutputValueClass(),
                         inMemorySegments, inMemorySegments.size(),
                         new Path(reduceId.toString()),
                         (RawComparator<K>)jobConf.getOutputKeyComparator(),
                         reporter, spilledRecordsCounter, null, null);
    
    if (null == combinerClass) {
      Merger.writeFile(rIter, writer, reporter, jobConf);
    } else {
      combineCollector.setWriter(writer);
      combineAndSpill(rIter, reduceCombineInputCounter);
    }
    writer.close();
    compressAwarePath = new CompressAwarePath(outputPath,
        writer.getRawLength(), writer.getCompressedLength());

    LOG.info(reduceId +  
        " Merge of the " + noInMemorySegments +
        " files in-memory complete." +
        " Local file is " + outputPath + " of size " + 
        localFS.getFileStatus(outputPath).getLen());
  } catch (IOException e) { 
    //make sure that we delete the ondisk file that we created 
    //earlier when we invoked cloneFileAttributes
    localFS.delete(outputPath, true);
    throw e;
  }

  // Note the output of the merge
  closeOnDiskFile(compressAwarePath);
}
 
源代码10 项目: hadoop   文件: MergeManagerImpl.java
@Override
public void merge(List<CompressAwarePath> inputs) throws IOException {
  // sanity check
  if (inputs == null || inputs.isEmpty()) {
    LOG.info("No ondisk files to merge...");
    return;
  }
  
  long approxOutputSize = 0;
  int bytesPerSum = 
    jobConf.getInt("io.bytes.per.checksum", 512);
  
  LOG.info("OnDiskMerger: We have  " + inputs.size() + 
           " map outputs on disk. Triggering merge...");
  
  // 1. Prepare the list of files to be merged. 
  for (CompressAwarePath file : inputs) {
    approxOutputSize += localFS.getFileStatus(file).getLen();
  }

  // add the checksum length
  approxOutputSize += 
    ChecksumFileSystem.getChecksumLength(approxOutputSize, bytesPerSum);

  // 2. Start the on-disk merge process
  Path outputPath = 
    localDirAllocator.getLocalPathForWrite(inputs.get(0).toString(), 
        approxOutputSize, jobConf).suffix(Task.MERGED_OUTPUT_PREFIX);

  FSDataOutputStream out = CryptoUtils.wrapIfNecessary(jobConf, rfs.create(outputPath));
  Writer<K, V> writer = new Writer<K, V>(jobConf, out,
      (Class<K>) jobConf.getMapOutputKeyClass(),
      (Class<V>) jobConf.getMapOutputValueClass(), codec, null, true);

  RawKeyValueIterator iter  = null;
  CompressAwarePath compressAwarePath;
  Path tmpDir = new Path(reduceId.toString());
  try {
    iter = Merger.merge(jobConf, rfs,
                        (Class<K>) jobConf.getMapOutputKeyClass(),
                        (Class<V>) jobConf.getMapOutputValueClass(),
                        codec, inputs.toArray(new Path[inputs.size()]), 
                        true, ioSortFactor, tmpDir, 
                        (RawComparator<K>) jobConf.getOutputKeyComparator(), 
                        reporter, spilledRecordsCounter, null, 
                        mergedMapOutputsCounter, null);

    Merger.writeFile(iter, writer, reporter, jobConf);
    writer.close();
    compressAwarePath = new CompressAwarePath(outputPath,
        writer.getRawLength(), writer.getCompressedLength());
  } catch (IOException e) {
    localFS.delete(outputPath, true);
    throw e;
  }

  closeOnDiskFile(compressAwarePath);

  LOG.info(reduceId +
      " Finished merging " + inputs.size() + 
      " map output files on disk of total-size " + 
      approxOutputSize + "." + 
      " Local output file is " + outputPath + " of size " +
      localFS.getFileStatus(outputPath).getLen());
}
 
源代码11 项目: hadoop   文件: TestShufflePlugin.java
@Test
/**
 * A testing method verifying availability and accessibility of API that is needed
 * for sub-classes of ShuffleConsumerPlugin
 */
public void testConsumerApi() {

  JobConf jobConf = new JobConf();
  ShuffleConsumerPlugin<K, V> shuffleConsumerPlugin = new TestShuffleConsumerPlugin<K, V>();

  //mock creation
  ReduceTask mockReduceTask = mock(ReduceTask.class);
  TaskUmbilicalProtocol mockUmbilical = mock(TaskUmbilicalProtocol.class);
  Reporter mockReporter = mock(Reporter.class);
  FileSystem mockFileSystem = mock(FileSystem.class);
  Class<? extends org.apache.hadoop.mapred.Reducer>  combinerClass = jobConf.getCombinerClass();
  @SuppressWarnings("unchecked")  // needed for mock with generic
  CombineOutputCollector<K, V>  mockCombineOutputCollector =
    (CombineOutputCollector<K, V>) mock(CombineOutputCollector.class);
  org.apache.hadoop.mapreduce.TaskAttemptID mockTaskAttemptID =
    mock(org.apache.hadoop.mapreduce.TaskAttemptID.class);
  LocalDirAllocator mockLocalDirAllocator = mock(LocalDirAllocator.class);
  CompressionCodec mockCompressionCodec = mock(CompressionCodec.class);
  Counter mockCounter = mock(Counter.class);
  TaskStatus mockTaskStatus = mock(TaskStatus.class);
  Progress mockProgress = mock(Progress.class);
  MapOutputFile mockMapOutputFile = mock(MapOutputFile.class);
  Task mockTask = mock(Task.class);

  try {
    String [] dirs = jobConf.getLocalDirs();
    // verify that these APIs are available through super class handler
    ShuffleConsumerPlugin.Context<K, V> context =
   new ShuffleConsumerPlugin.Context<K, V>(mockTaskAttemptID, jobConf, mockFileSystem,
                                              mockUmbilical, mockLocalDirAllocator,
                                              mockReporter, mockCompressionCodec,
                                              combinerClass, mockCombineOutputCollector,
                                              mockCounter, mockCounter, mockCounter,
                                              mockCounter, mockCounter, mockCounter,
                                              mockTaskStatus, mockProgress, mockProgress,
                                              mockTask, mockMapOutputFile, null);
    shuffleConsumerPlugin.init(context);
    shuffleConsumerPlugin.run();
    shuffleConsumerPlugin.close();
  }
  catch (Exception e) {
    assertTrue("Threw exception:" + e, false);
  }

  // verify that these APIs are available for 3rd party plugins
  mockReduceTask.getTaskID();
  mockReduceTask.getJobID();
  mockReduceTask.getNumMaps();
  mockReduceTask.getPartition();
  mockReporter.progress();
}
 
源代码12 项目: hadoop   文件: TestShuffleScheduler.java
@SuppressWarnings("rawtypes")
@Test
public <K, V> void TestSucceedAndFailedCopyMap() throws Exception {
  JobConf job = new JobConf();
  job.setNumMapTasks(2);
  //mock creation
  TaskUmbilicalProtocol mockUmbilical = mock(TaskUmbilicalProtocol.class);
  Reporter mockReporter = mock(Reporter.class);
  FileSystem mockFileSystem = mock(FileSystem.class);
  Class<? extends org.apache.hadoop.mapred.Reducer>  combinerClass = job.getCombinerClass();
  @SuppressWarnings("unchecked")  // needed for mock with generic
  CombineOutputCollector<K, V>  mockCombineOutputCollector =
      (CombineOutputCollector<K, V>) mock(CombineOutputCollector.class);
  org.apache.hadoop.mapreduce.TaskAttemptID mockTaskAttemptID =
      mock(org.apache.hadoop.mapreduce.TaskAttemptID.class);
  LocalDirAllocator mockLocalDirAllocator = mock(LocalDirAllocator.class);
  CompressionCodec mockCompressionCodec = mock(CompressionCodec.class);
  Counter mockCounter = mock(Counter.class);
  TaskStatus mockTaskStatus = mock(TaskStatus.class);
  Progress mockProgress = mock(Progress.class);
  MapOutputFile mockMapOutputFile = mock(MapOutputFile.class);
  Task mockTask = mock(Task.class);
  @SuppressWarnings("unchecked")
  MapOutput<K, V> output = mock(MapOutput.class);

  ShuffleConsumerPlugin.Context<K, V> context =
      new ShuffleConsumerPlugin.Context<K, V>(
          mockTaskAttemptID, job, mockFileSystem,
          mockUmbilical, mockLocalDirAllocator,
          mockReporter, mockCompressionCodec,
          combinerClass, mockCombineOutputCollector,
          mockCounter, mockCounter, mockCounter,
          mockCounter, mockCounter, mockCounter,
          mockTaskStatus, mockProgress, mockProgress,
          mockTask, mockMapOutputFile, null);
  TaskStatus status = new TaskStatus() {
    @Override
    public boolean getIsMap() {
      return false;
    }
    @Override
    public void addFetchFailedMap(TaskAttemptID mapTaskId) {
    }
  };
  Progress progress = new Progress();
  ShuffleSchedulerImpl<K, V> scheduler = new ShuffleSchedulerImpl<K, V>(job,
      status, null, null, progress, context.getShuffledMapsCounter(),
      context.getReduceShuffleBytes(), context.getFailedShuffleCounter());

  MapHost host1 = new MapHost("host1", null);
  TaskAttemptID failedAttemptID = new TaskAttemptID(
      new org.apache.hadoop.mapred.TaskID(
      new JobID("test",0), TaskType.MAP, 0), 0);

  TaskAttemptID succeedAttemptID = new TaskAttemptID(
      new org.apache.hadoop.mapred.TaskID(
      new JobID("test",0), TaskType.MAP, 1), 1);

  // handle output fetch failure for failedAttemptID, part I
  scheduler.hostFailed(host1.getHostName());

  // handle output fetch succeed for succeedAttemptID
  long bytes = (long)500 * 1024 * 1024;
  scheduler.copySucceeded(succeedAttemptID, host1, bytes, 0, 500000, output);

  // handle output fetch failure for failedAttemptID, part II
  // for MAPREDUCE-6361: verify no NPE exception get thrown out
  scheduler.copyFailed(failedAttemptID, host1, true, false);
}
 
源代码13 项目: big-c   文件: TaskAttemptImpl.java
static ContainerLaunchContext createContainerLaunchContext(
    Map<ApplicationAccessType, String> applicationACLs,
    Configuration conf, Token<JobTokenIdentifier> jobToken, Task remoteTask,
    final org.apache.hadoop.mapred.JobID oldJobId,
    WrappedJvmID jvmID,
    TaskAttemptListener taskAttemptListener,
    Credentials credentials) {

  synchronized (commonContainerSpecLock) {
    if (commonContainerSpec == null) {
      commonContainerSpec = createCommonContainerLaunchContext(
          applicationACLs, conf, jobToken, oldJobId, credentials);
    }
  }

  // Fill in the fields needed per-container that are missing in the common
  // spec.

  boolean userClassesTakesPrecedence =
    conf.getBoolean(MRJobConfig.MAPREDUCE_JOB_USER_CLASSPATH_FIRST, false);

  // Setup environment by cloning from common env.
  Map<String, String> env = commonContainerSpec.getEnvironment();
  Map<String, String> myEnv = new HashMap<String, String>(env.size());
  myEnv.putAll(env);
  if (userClassesTakesPrecedence) {
    myEnv.put(Environment.CLASSPATH_PREPEND_DISTCACHE.name(), "true");
  }
  MapReduceChildJVM.setVMEnv(myEnv, remoteTask);

  // Set up the launch command
  List<String> commands = MapReduceChildJVM.getVMCommand(
      taskAttemptListener.getAddress(), remoteTask, jvmID);

  // Duplicate the ByteBuffers for access by multiple containers.
  Map<String, ByteBuffer> myServiceData = new HashMap<String, ByteBuffer>();
  for (Entry<String, ByteBuffer> entry : commonContainerSpec
              .getServiceData().entrySet()) {
    myServiceData.put(entry.getKey(), entry.getValue().duplicate());
  }

  // Construct the actual Container
  ContainerLaunchContext container = ContainerLaunchContext.newInstance(
      commonContainerSpec.getLocalResources(), myEnv, commands,
      myServiceData, commonContainerSpec.getTokens().duplicate(),
      applicationACLs);

  return container;
}
 
源代码14 项目: big-c   文件: ContainerRemoteLaunchEvent.java
public Task getRemoteTask() {
  return this.task;
}
 
源代码15 项目: big-c   文件: TestTaskImpl.java
@Override
protected Task createRemoteTask() {
  return new MockTask(taskType);
}
 
源代码16 项目: big-c   文件: MergeManagerImpl.java
@Override
public void merge(List<InMemoryMapOutput<K,V>> inputs) throws IOException {
  if (inputs == null || inputs.size() == 0) {
    return;
  }
  
  //name this output file same as the name of the first file that is 
  //there in the current list of inmem files (this is guaranteed to
  //be absent on the disk currently. So we don't overwrite a prev. 
  //created spill). Also we need to create the output file now since
  //it is not guaranteed that this file will be present after merge
  //is called (we delete empty files as soon as we see them
  //in the merge method)

  //figure out the mapId 
  TaskAttemptID mapId = inputs.get(0).getMapId();
  TaskID mapTaskId = mapId.getTaskID();

  List<Segment<K, V>> inMemorySegments = new ArrayList<Segment<K, V>>();
  long mergeOutputSize = 
    createInMemorySegments(inputs, inMemorySegments,0);
  int noInMemorySegments = inMemorySegments.size();

  Path outputPath = 
    mapOutputFile.getInputFileForWrite(mapTaskId,
                                       mergeOutputSize).suffix(
                                           Task.MERGED_OUTPUT_PREFIX);

  FSDataOutputStream out = CryptoUtils.wrapIfNecessary(jobConf, rfs.create(outputPath));
  Writer<K, V> writer = new Writer<K, V>(jobConf, out,
      (Class<K>) jobConf.getMapOutputKeyClass(),
      (Class<V>) jobConf.getMapOutputValueClass(), codec, null, true);

  RawKeyValueIterator rIter = null;
  CompressAwarePath compressAwarePath;
  try {
    LOG.info("Initiating in-memory merge with " + noInMemorySegments + 
             " segments...");
    
    rIter = Merger.merge(jobConf, rfs,
                         (Class<K>)jobConf.getMapOutputKeyClass(),
                         (Class<V>)jobConf.getMapOutputValueClass(),
                         inMemorySegments, inMemorySegments.size(),
                         new Path(reduceId.toString()),
                         (RawComparator<K>)jobConf.getOutputKeyComparator(),
                         reporter, spilledRecordsCounter, null, null);
    
    if (null == combinerClass) {
      Merger.writeFile(rIter, writer, reporter, jobConf);
    } else {
      combineCollector.setWriter(writer);
      combineAndSpill(rIter, reduceCombineInputCounter);
    }
    writer.close();
    compressAwarePath = new CompressAwarePath(outputPath,
        writer.getRawLength(), writer.getCompressedLength());

    LOG.info(reduceId +  
        " Merge of the " + noInMemorySegments +
        " files in-memory complete." +
        " Local file is " + outputPath + " of size " + 
        localFS.getFileStatus(outputPath).getLen());
  } catch (IOException e) { 
    //make sure that we delete the ondisk file that we created 
    //earlier when we invoked cloneFileAttributes
    localFS.delete(outputPath, true);
    throw e;
  }

  // Note the output of the merge
  closeOnDiskFile(compressAwarePath);
}
 
源代码17 项目: big-c   文件: MergeManagerImpl.java
@Override
public void merge(List<CompressAwarePath> inputs) throws IOException {
  // sanity check
  if (inputs == null || inputs.isEmpty()) {
    LOG.info("No ondisk files to merge...");
    return;
  }
  
  long approxOutputSize = 0;
  int bytesPerSum = 
    jobConf.getInt("io.bytes.per.checksum", 512);
  
  LOG.info("OnDiskMerger: We have  " + inputs.size() + 
           " map outputs on disk. Triggering merge...");
  
  // 1. Prepare the list of files to be merged. 
  for (CompressAwarePath file : inputs) {
    approxOutputSize += localFS.getFileStatus(file).getLen();
  }

  // add the checksum length
  approxOutputSize += 
    ChecksumFileSystem.getChecksumLength(approxOutputSize, bytesPerSum);

  // 2. Start the on-disk merge process
  Path outputPath = 
    localDirAllocator.getLocalPathForWrite(inputs.get(0).toString(), 
        approxOutputSize, jobConf).suffix(Task.MERGED_OUTPUT_PREFIX);

  FSDataOutputStream out = CryptoUtils.wrapIfNecessary(jobConf, rfs.create(outputPath));
  Writer<K, V> writer = new Writer<K, V>(jobConf, out,
      (Class<K>) jobConf.getMapOutputKeyClass(),
      (Class<V>) jobConf.getMapOutputValueClass(), codec, null, true);

  RawKeyValueIterator iter  = null;
  CompressAwarePath compressAwarePath;
  Path tmpDir = new Path(reduceId.toString());
  try {
    iter = Merger.merge(jobConf, rfs,
                        (Class<K>) jobConf.getMapOutputKeyClass(),
                        (Class<V>) jobConf.getMapOutputValueClass(),
                        codec, inputs.toArray(new Path[inputs.size()]), 
                        true, ioSortFactor, tmpDir, 
                        (RawComparator<K>) jobConf.getOutputKeyComparator(), 
                        reporter, spilledRecordsCounter, null, 
                        mergedMapOutputsCounter, null);

    Merger.writeFile(iter, writer, reporter, jobConf);
    writer.close();
    compressAwarePath = new CompressAwarePath(outputPath,
        writer.getRawLength(), writer.getCompressedLength());
  } catch (IOException e) {
    localFS.delete(outputPath, true);
    throw e;
  }

  closeOnDiskFile(compressAwarePath);

  LOG.info(reduceId +
      " Finished merging " + inputs.size() + 
      " map output files on disk of total-size " + 
      approxOutputSize + "." + 
      " Local output file is " + outputPath + " of size " +
      localFS.getFileStatus(outputPath).getLen());
}
 
源代码18 项目: big-c   文件: TestShufflePlugin.java
@Test
/**
 * A testing method verifying availability and accessibility of API that is needed
 * for sub-classes of ShuffleConsumerPlugin
 */
public void testConsumerApi() {

  JobConf jobConf = new JobConf();
  ShuffleConsumerPlugin<K, V> shuffleConsumerPlugin = new TestShuffleConsumerPlugin<K, V>();

  //mock creation
  ReduceTask mockReduceTask = mock(ReduceTask.class);
  TaskUmbilicalProtocol mockUmbilical = mock(TaskUmbilicalProtocol.class);
  Reporter mockReporter = mock(Reporter.class);
  FileSystem mockFileSystem = mock(FileSystem.class);
  Class<? extends org.apache.hadoop.mapred.Reducer>  combinerClass = jobConf.getCombinerClass();
  @SuppressWarnings("unchecked")  // needed for mock with generic
  CombineOutputCollector<K, V>  mockCombineOutputCollector =
    (CombineOutputCollector<K, V>) mock(CombineOutputCollector.class);
  org.apache.hadoop.mapreduce.TaskAttemptID mockTaskAttemptID =
    mock(org.apache.hadoop.mapreduce.TaskAttemptID.class);
  LocalDirAllocator mockLocalDirAllocator = mock(LocalDirAllocator.class);
  CompressionCodec mockCompressionCodec = mock(CompressionCodec.class);
  Counter mockCounter = mock(Counter.class);
  TaskStatus mockTaskStatus = mock(TaskStatus.class);
  Progress mockProgress = mock(Progress.class);
  MapOutputFile mockMapOutputFile = mock(MapOutputFile.class);
  Task mockTask = mock(Task.class);

  try {
    String [] dirs = jobConf.getLocalDirs();
    // verify that these APIs are available through super class handler
    ShuffleConsumerPlugin.Context<K, V> context =
   new ShuffleConsumerPlugin.Context<K, V>(mockTaskAttemptID, jobConf, mockFileSystem,
                                              mockUmbilical, mockLocalDirAllocator,
                                              mockReporter, mockCompressionCodec,
                                              combinerClass, mockCombineOutputCollector,
                                              mockCounter, mockCounter, mockCounter,
                                              mockCounter, mockCounter, mockCounter,
                                              mockTaskStatus, mockProgress, mockProgress,
                                              mockTask, mockMapOutputFile, null);
    shuffleConsumerPlugin.init(context);
    shuffleConsumerPlugin.run();
    shuffleConsumerPlugin.close();
  }
  catch (Exception e) {
    assertTrue("Threw exception:" + e, false);
  }

  // verify that these APIs are available for 3rd party plugins
  mockReduceTask.getTaskID();
  mockReduceTask.getJobID();
  mockReduceTask.getNumMaps();
  mockReduceTask.getPartition();
  mockReporter.progress();
}
 
源代码19 项目: big-c   文件: TestShuffleScheduler.java
@SuppressWarnings("rawtypes")
@Test
public <K, V> void TestSucceedAndFailedCopyMap() throws Exception {
  JobConf job = new JobConf();
  job.setNumMapTasks(2);
  //mock creation
  TaskUmbilicalProtocol mockUmbilical = mock(TaskUmbilicalProtocol.class);
  Reporter mockReporter = mock(Reporter.class);
  FileSystem mockFileSystem = mock(FileSystem.class);
  Class<? extends org.apache.hadoop.mapred.Reducer>  combinerClass = job.getCombinerClass();
  @SuppressWarnings("unchecked")  // needed for mock with generic
  CombineOutputCollector<K, V>  mockCombineOutputCollector =
      (CombineOutputCollector<K, V>) mock(CombineOutputCollector.class);
  org.apache.hadoop.mapreduce.TaskAttemptID mockTaskAttemptID =
      mock(org.apache.hadoop.mapreduce.TaskAttemptID.class);
  LocalDirAllocator mockLocalDirAllocator = mock(LocalDirAllocator.class);
  CompressionCodec mockCompressionCodec = mock(CompressionCodec.class);
  Counter mockCounter = mock(Counter.class);
  TaskStatus mockTaskStatus = mock(TaskStatus.class);
  Progress mockProgress = mock(Progress.class);
  MapOutputFile mockMapOutputFile = mock(MapOutputFile.class);
  Task mockTask = mock(Task.class);
  @SuppressWarnings("unchecked")
  MapOutput<K, V> output = mock(MapOutput.class);

  ShuffleConsumerPlugin.Context<K, V> context =
      new ShuffleConsumerPlugin.Context<K, V>(
          mockTaskAttemptID, job, mockFileSystem,
          mockUmbilical, mockLocalDirAllocator,
          mockReporter, mockCompressionCodec,
          combinerClass, mockCombineOutputCollector,
          mockCounter, mockCounter, mockCounter,
          mockCounter, mockCounter, mockCounter,
          mockTaskStatus, mockProgress, mockProgress,
          mockTask, mockMapOutputFile, null);
  TaskStatus status = new TaskStatus() {
    @Override
    public boolean getIsMap() {
      return false;
    }
    @Override
    public void addFetchFailedMap(TaskAttemptID mapTaskId) {
    }
  };
  Progress progress = new Progress();
  ShuffleSchedulerImpl<K, V> scheduler = new ShuffleSchedulerImpl<K, V>(job,
      status, null, null, progress, context.getShuffledMapsCounter(),
      context.getReduceShuffleBytes(), context.getFailedShuffleCounter());

  MapHost host1 = new MapHost("host1", null);
  TaskAttemptID failedAttemptID = new TaskAttemptID(
      new org.apache.hadoop.mapred.TaskID(
      new JobID("test",0), TaskType.MAP, 0), 0);

  TaskAttemptID succeedAttemptID = new TaskAttemptID(
      new org.apache.hadoop.mapred.TaskID(
      new JobID("test",0), TaskType.MAP, 1), 1);

  // handle output fetch failure for failedAttemptID, part I
  scheduler.hostFailed(host1.getHostName());

  // handle output fetch succeed for succeedAttemptID
  long bytes = (long)500 * 1024 * 1024;
  scheduler.copySucceeded(succeedAttemptID, host1, bytes, 0, 500000, output);

  // handle output fetch failure for failedAttemptID, part II
  // for MAPREDUCE-6361: verify no NPE exception get thrown out
  scheduler.copyFailed(failedAttemptID, host1, true, false);
}
 
源代码20 项目: hadoop   文件: TaskAttemptListener.java
/**
 * Register a JVM with the listener.  This should be called as soon as a 
 * JVM ID is assigned to a task attempt, before it has been launched.
 * @param task the task itself for this JVM.
 * @param jvmID The ID of the JVM .
 */
void registerPendingTask(Task task, WrappedJvmID jvmID);
 
源代码21 项目: big-c   文件: TaskAttemptListener.java
/**
 * Register a JVM with the listener.  This should be called as soon as a 
 * JVM ID is assigned to a task attempt, before it has been launched.
 * @param task the task itself for this JVM.
 * @param jvmID The ID of the JVM .
 */
void registerPendingTask(Task task, WrappedJvmID jvmID);
 
源代码22 项目: hadoop   文件: TaskAttemptImpl.java
protected abstract org.apache.hadoop.mapred.Task createRemoteTask(); 
源代码23 项目: big-c   文件: TaskAttemptImpl.java
protected abstract org.apache.hadoop.mapred.Task createRemoteTask(); 
 类所在包
 类方法
 同包方法