Package org.apache.hadoop.mapred

Source Code of org.apache.hadoop.mapred.JobInProgress$FallowSlotInfo

/**
* Licensed to the Apache Software Foundation (ASF) under one
* or more contributor license agreements.  See the NOTICE file
* distributed with this work for additional information
* regarding copyright ownership.  The ASF licenses this file
* to you under the Apache License, Version 2.0 (the
* "License"); you may not use this file except in compliance
* with the License.  You may obtain a copy of the License at
*
*     http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.hadoop.mapred;

import java.io.File;
import java.io.FileOutputStream;
import java.io.IOException;
import java.security.PrivilegedExceptionAction;
import java.util.ArrayList;
import java.util.Collection;
import java.util.Collections;
import java.util.Comparator;
import java.util.HashMap;
import java.util.HashSet;
import java.util.IdentityHashMap;
import java.util.Iterator;
import java.util.LinkedHashSet;
import java.util.LinkedList;
import java.util.List;
import java.util.Map;
import java.util.Set;
import java.util.TreeMap;
import java.util.Vector;
import java.util.concurrent.atomic.AtomicBoolean;

import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
import org.apache.hadoop.classification.InterfaceAudience;
import org.apache.hadoop.classification.InterfaceStability;
import org.apache.hadoop.fs.FSDataOutputStream;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.LocalFileSystem;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapred.CleanupQueue.PathDeletionContext;
import org.apache.hadoop.mapreduce.JobContext;
import org.apache.hadoop.mapreduce.JobCounter;
import org.apache.hadoop.mapreduce.JobACL;
import org.apache.hadoop.mapreduce.JobSubmissionFiles;
import org.apache.hadoop.mapreduce.MRJobConfig;
import org.apache.hadoop.mapreduce.TaskType;
import org.apache.hadoop.mapreduce.jobhistory.JobFinishedEvent;
import org.apache.hadoop.mapreduce.jobhistory.JobHistory;
import org.apache.hadoop.mapreduce.jobhistory.JobInfoChangeEvent;
import org.apache.hadoop.mapreduce.jobhistory.JobInitedEvent;
import org.apache.hadoop.mapreduce.jobhistory.JobPriorityChangeEvent;
import org.apache.hadoop.mapreduce.jobhistory.JobStatusChangedEvent;
import org.apache.hadoop.mapreduce.jobhistory.JobSubmittedEvent;
import org.apache.hadoop.mapreduce.jobhistory.JobUnsuccessfulCompletionEvent;
import org.apache.hadoop.mapreduce.jobhistory.MapAttemptFinishedEvent;
import org.apache.hadoop.mapreduce.jobhistory.ReduceAttemptFinishedEvent;
import org.apache.hadoop.mapreduce.jobhistory.TaskAttemptStartedEvent;
import org.apache.hadoop.mapreduce.jobhistory.TaskAttemptUnsuccessfulCompletionEvent;
import org.apache.hadoop.mapreduce.jobhistory.TaskFailedEvent;
import org.apache.hadoop.mapreduce.jobhistory.TaskFinishedEvent;
import org.apache.hadoop.mapreduce.jobhistory.TaskStartedEvent;
import org.apache.hadoop.mapreduce.security.TokenCache;
import org.apache.hadoop.security.TokenStorage;
import org.apache.hadoop.mapreduce.security.token.DelegationTokenRenewal;
import org.apache.hadoop.mapreduce.security.token.JobTokenIdentifier;
import org.apache.hadoop.mapreduce.server.jobtracker.TaskTracker;
import org.apache.hadoop.mapreduce.split.JobSplit;
import org.apache.hadoop.mapreduce.split.SplitMetaInfoReader;
import org.apache.hadoop.mapreduce.split.JobSplit.TaskSplitMetaInfo;
import org.apache.hadoop.mapreduce.task.JobContextImpl;
import org.apache.hadoop.metrics.MetricsContext;
import org.apache.hadoop.metrics.MetricsRecord;
import org.apache.hadoop.metrics.MetricsUtil;
import org.apache.hadoop.net.NetUtils;
import org.apache.hadoop.net.NetworkTopology;
import org.apache.hadoop.net.Node;
import org.apache.hadoop.security.AccessControlException;
import org.apache.hadoop.security.UserGroupInformation;
import org.apache.hadoop.security.token.Token;
import org.apache.hadoop.security.token.TokenIdentifier;
import org.apache.hadoop.util.StringUtils;

/**
* JobInProgress maintains all the info for keeping a Job on the straight and
* narrow. It keeps its JobProfile and its latest JobStatus, plus a set of
* tables for doing bookkeeping of its Tasks.
*/
@InterfaceAudience.LimitedPrivate({"MapReduce"})
@InterfaceStability.Unstable
public class JobInProgress {
  /**
   * Used when the a kill is issued to a job which is initializing.
   */
  static class KillInterruptedException extends InterruptedException {
   private static final long serialVersionUID = 1L;
    public KillInterruptedException(String msg) {
      super(msg);
    }
  }
 
  static final Log LOG = LogFactory.getLog(JobInProgress.class);
   
  JobProfile profile;
  JobStatus status;
  Path jobFile = null;
  Path localJobFile = null;

  TaskInProgress maps[] = new TaskInProgress[0];
  TaskInProgress reduces[] = new TaskInProgress[0];
  TaskInProgress cleanup[] = new TaskInProgress[0];
  TaskInProgress setup[] = new TaskInProgress[0];
  int numMapTasks = 0;
  int numReduceTasks = 0;
  int numSlotsPerMap = 1;
  int numSlotsPerReduce = 1;
 
  // Counters to track currently running/finished/failed Map/Reduce task-attempts
  int runningMapTasks = 0;
  int runningReduceTasks = 0;
  int finishedMapTasks = 0;
  int finishedReduceTasks = 0;
  int failedMapTasks = 0;
  int failedReduceTasks = 0;
 
  static final float DEFAULT_COMPLETED_MAPS_PERCENT_FOR_REDUCE_SLOWSTART = 0.05f;
  int completedMapsForReduceSlowstart = 0;
 
  // runningMapTasks include speculative tasks, so we need to capture
  // speculative tasks separately
  int speculativeMapTasks = 0;
  int speculativeReduceTasks = 0;
 
  int mapFailuresPercent = 0;
  int reduceFailuresPercent = 0;
  int failedMapTIPs = 0;
  int failedReduceTIPs = 0;
  private volatile boolean launchedCleanup = false;
  private volatile boolean launchedSetup = false;
  private volatile boolean jobKilled = false;
  private volatile boolean jobFailed = false;
  private final boolean jobSetupCleanupNeeded;

  JobPriority priority = JobPriority.NORMAL;
  protected JobTracker jobtracker;
 
  protected TokenStorage tokenStorage;
 
  JobHistory jobHistory;

  // NetworkTopology Node to the set of TIPs
  Map<Node, List<TaskInProgress>> nonRunningMapCache;
 
  // Map of NetworkTopology Node to set of running TIPs
  Map<Node, Set<TaskInProgress>> runningMapCache;

  // A list of non-local non-running maps
  List<TaskInProgress> nonLocalMaps;

  // A set of non-local running maps
  Set<TaskInProgress> nonLocalRunningMaps;

  // A list of non-running reduce TIPs
  List<TaskInProgress> nonRunningReduces;

  // A set of running reduce TIPs
  Set<TaskInProgress> runningReduces;
 
  // A list of cleanup tasks for the map task attempts, to be launched
  List<TaskAttemptID> mapCleanupTasks = new LinkedList<TaskAttemptID>();
 
  // A list of cleanup tasks for the reduce task attempts, to be launched
  List<TaskAttemptID> reduceCleanupTasks = new LinkedList<TaskAttemptID>();

  int maxLevel;

  /**
   * A special value indicating that
   * {@link #findNewMapTask(TaskTrackerStatus, int, int, int, double)} should
   * schedule any available map tasks for this job, including speculative tasks.
   */
  int anyCacheLevel;
 
  /**
   * A special value indicating that
   * {@link #findNewMapTask(TaskTrackerStatus, int, int, int, double)} should
   * schedule any only off-switch and speculative map tasks for this job.
   */
  private static final int NON_LOCAL_CACHE_LEVEL = -1;

  private int taskCompletionEventTracker = 0;
  List<TaskCompletionEvent> taskCompletionEvents;
   
  // The maximum percentage of trackers in cluster added to the 'blacklist'.
  private static final double CLUSTER_BLACKLIST_PERCENT = 0.25;
 
  // The maximum percentage of fetch failures allowed for a map
  private static final double MAX_ALLOWED_FETCH_FAILURES_PERCENT = 0.5;
 
  // No. of tasktrackers in the cluster
  private volatile int clusterSize = 0;
 
  // The no. of tasktrackers where >= conf.getMaxTaskFailuresPerTracker()
  // tasks have failed
  private volatile int flakyTaskTrackers = 0;
  // Map of trackerHostName -> no. of task failures
  private Map<String, Integer> trackerToFailuresMap =
    new TreeMap<String, Integer>();
   
  //Confine estimation algorithms to an "oracle" class that JIP queries.
  ResourceEstimator resourceEstimator;
 
  long startTime;
  long launchTime;
  long finishTime;
 
  // Indicates how many times the job got restarted
  private final int restartCount;

  JobConf conf;
  protected AtomicBoolean tasksInited = new AtomicBoolean(false);
  private JobInitKillStatus jobInitKillStatus = new JobInitKillStatus();

  LocalFileSystem localFs;
  FileSystem fs;
  String user;
  JobID jobId;
  private boolean hasSpeculativeMaps;
  private boolean hasSpeculativeReduces;
  long inputLength = 0;
 
  Counters jobCounters = new Counters();
 
  // Maximum no. of fetch-failure notifications after which map task is killed
  private static final int MAX_FETCH_FAILURES_NOTIFICATIONS = 3;

  // Don't lower speculativeCap below one TT's worth (for small clusters)
  private static final int MIN_SPEC_CAP = 10;
 
  private static final float MIN_SLOTS_CAP = 0.01f;
 
  // Map of mapTaskId -> no. of fetch failures
  private Map<TaskAttemptID, Integer> mapTaskIdToFetchFailuresMap =
    new TreeMap<TaskAttemptID, Integer>();

  private Object schedulingInfo;

  //thresholds for speculative execution
  float slowTaskThreshold;
  float speculativeCap;
  float slowNodeThreshold; //standard deviations

  //Statistics are maintained for a couple of things
  //mapTaskStats is used for maintaining statistics about
  //the completion time of map tasks on the trackers. On a per
  //tracker basis, the mean time for task completion is maintained
  private DataStatistics mapTaskStats = new DataStatistics();
  //reduceTaskStats is used for maintaining statistics about
  //the completion time of reduce tasks on the trackers. On a per
  //tracker basis, the mean time for task completion is maintained
  private DataStatistics reduceTaskStats = new DataStatistics();
  //trackerMapStats used to maintain a mapping from the tracker to the
  //the statistics about completion time of map tasks
  private Map<String,DataStatistics> trackerMapStats =
    new HashMap<String,DataStatistics>();
  //trackerReduceStats used to maintain a mapping from the tracker to the
  //the statistics about completion time of reduce tasks
  private Map<String,DataStatistics> trackerReduceStats =
    new HashMap<String,DataStatistics>();
  //runningMapStats used to maintain the RUNNING map tasks' statistics
  private DataStatistics runningMapTaskStats = new DataStatistics();
  //runningReduceStats used to maintain the RUNNING reduce tasks' statistics
  private DataStatistics runningReduceTaskStats = new DataStatistics();
  private static class FallowSlotInfo {
    long timestamp;
    int numSlots;
   
    public FallowSlotInfo(long timestamp, int numSlots) {
      this.timestamp = timestamp;
      this.numSlots = numSlots;
    }

    public long getTimestamp() {
      return timestamp;
    }

    public void setTimestamp(long timestamp) {
      this.timestamp = timestamp;
    }

    public int getNumSlots() {
      return numSlots;
    }

    public void setNumSlots(int numSlots) {
      this.numSlots = numSlots;
    }
  }
 
  private Map<TaskTracker, FallowSlotInfo> trackersReservedForMaps =
    new HashMap<TaskTracker, FallowSlotInfo>();
  private Map<TaskTracker, FallowSlotInfo> trackersReservedForReduces =
    new HashMap<TaskTracker, FallowSlotInfo>();
  private Path jobSubmitDir = null;
 
  /**
   * Create an almost empty JobInProgress, which can be used only for tests
   */
  protected JobInProgress(JobID jobid, JobConf conf, JobTracker tracker) {
    this.conf = conf;
    this.jobId = jobid;
    this.numMapTasks = conf.getNumMapTasks();
    this.numReduceTasks = conf.getNumReduceTasks();
    this.maxLevel = NetworkTopology.DEFAULT_HOST_LEVEL;
    this.anyCacheLevel = this.maxLevel+1;
    this.jobtracker = tracker;
    this.restartCount = 0;
    this.profile = new JobProfile(conf.getUser(), jobid, "", "",
                                  conf.getJobName(),conf.getQueueName());

   
    hasSpeculativeMaps = conf.getMapSpeculativeExecution();
    hasSpeculativeReduces = conf.getReduceSpeculativeExecution();
    this.nonLocalMaps = new LinkedList<TaskInProgress>();
    this.nonLocalRunningMaps = new LinkedHashSet<TaskInProgress>();
    this.runningMapCache = new IdentityHashMap<Node, Set<TaskInProgress>>();
    this.nonRunningReduces = new LinkedList<TaskInProgress>();   
    this.runningReduces = new LinkedHashSet<TaskInProgress>();
    this.resourceEstimator = new ResourceEstimator(this);
    this.status = new JobStatus(jobid, 0.0f, 0.0f, JobStatus.PREP,
        this.profile.getUser(), this.profile.getJobName(),
        this.profile.getJobFile(), "");
    this.jobtracker.getInstrumentation().addPrepJob(conf, jobid);
    this.taskCompletionEvents = new ArrayList<TaskCompletionEvent>
    (numMapTasks + numReduceTasks + 10);
   
    this.slowTaskThreshold = Math.max(0.0f,
        conf.getFloat(MRJobConfig.SPECULATIVE_SLOWTASK_THRESHOLD,1.0f));
    this.speculativeCap = conf.getFloat(
        MRJobConfig.SPECULATIVECAP,0.1f);
    this.slowNodeThreshold = conf.getFloat(
        MRJobConfig.SPECULATIVE_SLOWNODE_THRESHOLD,1.0f);
    this.jobSetupCleanupNeeded = conf.getBoolean(
        MRJobConfig.SETUP_CLEANUP_NEEDED, true);
    if (tracker != null) { // Some mock tests have null tracker
      this.jobHistory = tracker.getJobHistory();
    }
    this.tokenStorage = null;
  }
 
  JobInProgress() {
    restartCount = 0;
    jobSetupCleanupNeeded = false;
  }
 
  /**
   * Create a JobInProgress with the given job file, plus a handle
   * to the tracker.
   */
  public JobInProgress(JobTracker jobtracker,
                       final JobConf default_conf, int rCount,
                       JobInfo jobInfo,
                       TokenStorage ts
                      ) throws IOException, InterruptedException {
    this.restartCount = rCount;
    this.jobId = JobID.downgrade(jobInfo.getJobID());
    String url = "http://" + jobtracker.getJobTrackerMachine() + ":"
        + jobtracker.getInfoPort() + "/jobdetails.jsp?jobid=" + this.jobId;
    this.jobtracker = jobtracker;
    this.jobHistory = jobtracker.getJobHistory();
    this.startTime = System.currentTimeMillis();
   
    this.localFs = jobtracker.getLocalFileSystem();
    this.tokenStorage = ts;
    // use the user supplied token to add user credentials to the conf
    jobSubmitDir = jobInfo.getJobSubmitDir();
    user = jobInfo.getUser().toString();

    UserGroupInformation ugi = UserGroupInformation.createRemoteUser(user);
    if (ts != null) {
      for (Token<? extends TokenIdentifier> token : ts.getAllTokens()) {
        ugi.addToken(token);
      }
    }

    fs = ugi.doAs(new PrivilegedExceptionAction<FileSystem>() {
      public FileSystem run() throws IOException {
        return jobSubmitDir.getFileSystem(default_conf);
      }});
    this.localJobFile =
      default_conf.getLocalPath(JobTracker.SUBDIR + "/" + this.jobId + ".xml");
   
    jobFile = JobSubmissionFiles.getJobConfPath(jobSubmitDir);
    fs.copyToLocalFile(jobFile, localJobFile);
    conf = new JobConf(localJobFile);
    if (conf.getUser() == null) {
      this.conf.setUser(user);
    }
    if (!conf.getUser().equals(user)) {
      throw new IOException("The username obtained from the conf doesn't " +
          "match the username the user authenticated as");
    }
    this.priority = conf.getJobPriority();
    this.profile = new JobProfile(conf.getUser(), this.jobId,
                                  jobFile.toString(), url,
                                  conf.getJobName(), conf.getQueueName());
    this.status = new JobStatus(this.jobId, 0.0f, 0.0f, JobStatus.PREP,
        profile.getUser(), profile.getJobName(), profile.getJobFile(),
        profile.getURL().toString());
    this.jobtracker.getInstrumentation().addPrepJob(conf, this.jobId);
    status.setStartTime(startTime);
    this.status.setJobPriority(this.priority);

    this.numMapTasks = conf.getNumMapTasks();
    this.numReduceTasks = conf.getNumReduceTasks();
    this.taskCompletionEvents = new ArrayList<TaskCompletionEvent>
       (numMapTasks + numReduceTasks + 10);
    JobContext jobContext = new JobContextImpl(conf, jobId);
    this.jobSetupCleanupNeeded = jobContext.getJobSetupCleanupNeeded();

    // Construct the jobACLs
    status.setJobACLs(jobtracker.getJobACLsManager().constructJobACLs(conf));

    this.mapFailuresPercent = conf.getMaxMapTaskFailuresPercent();
    this.reduceFailuresPercent = conf.getMaxReduceTaskFailuresPercent();
       
    hasSpeculativeMaps = conf.getMapSpeculativeExecution();
    hasSpeculativeReduces = conf.getReduceSpeculativeExecution();
    this.maxLevel = jobtracker.getNumTaskCacheLevels();
    this.anyCacheLevel = this.maxLevel+1;
    this.nonLocalMaps = new LinkedList<TaskInProgress>();
    this.nonLocalRunningMaps = new LinkedHashSet<TaskInProgress>();
    this.runningMapCache = new IdentityHashMap<Node, Set<TaskInProgress>>();
    this.nonRunningReduces = new LinkedList<TaskInProgress>();   
    this.runningReduces = new LinkedHashSet<TaskInProgress>();
    this.resourceEstimator = new ResourceEstimator(this);
   
    this.nonLocalMaps = new LinkedList<TaskInProgress>();
    this.nonLocalRunningMaps = new LinkedHashSet<TaskInProgress>();
    this.runningMapCache = new IdentityHashMap<Node, Set<TaskInProgress>>();
    this.nonRunningReduces = new LinkedList<TaskInProgress>();   
    this.runningReduces = new LinkedHashSet<TaskInProgress>();
    this.slowTaskThreshold = Math.max(0.0f,
        conf.getFloat(MRJobConfig.SPECULATIVE_SLOWTASK_THRESHOLD,1.0f));
    this.speculativeCap = conf.getFloat(
        MRJobConfig.SPECULATIVECAP,0.1f);
    this.slowNodeThreshold = conf.getFloat(
        MRJobConfig.SPECULATIVE_SLOWNODE_THRESHOLD,1.0f);
    // register job's tokens for renewal
    DelegationTokenRenewal.registerDelegationTokensForRenewal(
        jobInfo.getJobID(), ts, this.conf);
  }
   
  private void printCache (Map<Node, List<TaskInProgress>> cache) {
    LOG.info("The taskcache info:");
    for (Map.Entry<Node, List<TaskInProgress>> n : cache.entrySet()) {
      List <TaskInProgress> tips = n.getValue();
      LOG.info("Cached TIPs on node: " + n.getKey());
      for (TaskInProgress tip : tips) {
        LOG.info("tip : " + tip.getTIPId());
      }
    }
  }
 
  Map<Node, List<TaskInProgress>> createCache(
                         TaskSplitMetaInfo[] splits, int maxLevel) {
    Map<Node, List<TaskInProgress>> cache =
      new IdentityHashMap<Node, List<TaskInProgress>>(maxLevel);
   
    for (int i = 0; i < splits.length; i++) {
      String[] splitLocations = splits[i].getLocations();
      if (splitLocations.length == 0) {
        nonLocalMaps.add(maps[i]);
        continue;
      }

      for(String host: splitLocations) {
        Node node = jobtracker.resolveAndAddToTopology(host);
        LOG.info("tip:" + maps[i].getTIPId() + " has split on node:" + node);
        for (int j = 0; j < maxLevel; j++) {
          List<TaskInProgress> hostMaps = cache.get(node);
          if (hostMaps == null) {
            hostMaps = new ArrayList<TaskInProgress>();
            cache.put(node, hostMaps);
            hostMaps.add(maps[i]);
          }
          //check whether the hostMaps already contains an entry for a TIP
          //This will be true for nodes that are racks and multiple nodes in
          //the rack contain the input for a tip. Note that if it already
          //exists in the hostMaps, it must be the last element there since
          //we process one TIP at a time sequentially in the split-size order
          if (hostMaps.get(hostMaps.size() - 1) != maps[i]) {
            hostMaps.add(maps[i]);
          }
          node = node.getParent();
        }
      }
    }
    return cache;
  }
 
  /**
   * Check if the job has been initialized.
   * @return <code>true</code> if the job has been initialized,
   *         <code>false</code> otherwise
   */
  public boolean inited() {
    return tasksInited.get();
  }
 
  /**
   * Get the user for the job
   */
  public String getUser() {
    return user;
  }

  /**
   * Get the number of slots required to run a single map task-attempt.
   * @return the number of slots required to run a single map task-attempt
   */
  synchronized int getNumSlotsPerMap() {
    return numSlotsPerMap;
  }

  /**
   * Set the number of slots required to run a single map task-attempt.
   * This is typically set by schedulers which support high-ram jobs.
   * @param slots the number of slots required to run a single map task-attempt
   */
  synchronized void setNumSlotsPerMap(int numSlotsPerMap) {
    this.numSlotsPerMap = numSlotsPerMap;
  }

  /**
   * Get the number of slots required to run a single reduce task-attempt.
   * @return the number of slots required to run a single reduce task-attempt
   */
  synchronized int getNumSlotsPerReduce() {
    return numSlotsPerReduce;
  }

  /**
   * Set the number of slots required to run a single reduce task-attempt.
   * This is typically set by schedulers which support high-ram jobs.
   * @param slots the number of slots required to run a single reduce
   *              task-attempt
   */
  synchronized void setNumSlotsPerReduce(int numSlotsPerReduce) {
    this.numSlotsPerReduce = numSlotsPerReduce;
  }

  /**
   * Construct the splits, etc.  This is invoked from an async
   * thread so that split-computation doesn't block anyone. Only the
   * {@link JobTracker} should invoke this api. Look
   * at {@link JobTracker#initJob(JobInProgress)} for more details.
   */
  public synchronized void initTasks()
  throws IOException, KillInterruptedException {
    if (tasksInited.get() || isComplete()) {
      return;
    }
    synchronized(jobInitKillStatus){
      if(jobInitKillStatus.killed || jobInitKillStatus.initStarted) {
        return;
      }
      jobInitKillStatus.initStarted = true;
    }

    LOG.info("Initializing " + jobId);

    logSubmissionToJobHistory();
   
    // log the job priority
    setPriority(this.priority);
   
    //
    // generate security keys needed by Tasks
    //
    generateAndStoreTokens();
   
    //
    // read input splits and create a map per a split
    //
    TaskSplitMetaInfo[] taskSplitMetaInfo = createSplits(jobId);
    numMapTasks = taskSplitMetaInfo.length;

    checkTaskLimits();

    jobtracker.getInstrumentation().addWaitingMaps(getJobID(), numMapTasks);
    jobtracker.getInstrumentation().addWaitingReduces(getJobID(), numReduceTasks);

    createMapTasks(jobFile.toString(), taskSplitMetaInfo);
   
    if (numMapTasks > 0) {
      nonRunningMapCache = createCache(taskSplitMetaInfo,
          maxLevel);
    }
       
    // set the launch time
    this.launchTime = JobTracker.getClock().getTime();

    createReduceTasks(jobFile.toString());
   
    // Calculate the minimum number of maps to be complete before
    // we should start scheduling reduces
    completedMapsForReduceSlowstart =
      (int)Math.ceil(
          (conf.getFloat(MRJobConfig.COMPLETED_MAPS_FOR_REDUCE_SLOWSTART,
                         DEFAULT_COMPLETED_MAPS_PERCENT_FOR_REDUCE_SLOWSTART) *
           numMapTasks));
   
    initSetupCleanupTasks(jobFile.toString());
   
    synchronized(jobInitKillStatus){
      jobInitKillStatus.initDone = true;
      if(jobInitKillStatus.killed) {
        //setup not launched so directly terminate
        throw new KillInterruptedException("Job " + jobId + " killed in init");
      }
    }
   
    tasksInited.set(true);
    JobInitedEvent jie = new JobInitedEvent(
        profile.getJobID()this.launchTime,
        numMapTasks, numReduceTasks,
        JobStatus.getJobRunState(JobStatus.PREP));
   
    jobHistory.logEvent(jie, jobId);
  
    // Log the number of map and reduce tasks
    LOG.info("Job " + jobId + " initialized successfully with " + numMapTasks
             + " map tasks and " + numReduceTasks + " reduce tasks.");
  }

  // Returns true if the job is empty (0 maps, 0 reduces and no setup-cleanup)
  // else return false.
  synchronized boolean isJobEmpty() {
    return maps.length == 0 && reduces.length == 0 && !jobSetupCleanupNeeded;
  }
 
  synchronized boolean isSetupCleanupRequired() {
   return jobSetupCleanupNeeded;
  }

  // Should be called once the init is done. This will complete the job
  // because the job is empty (0 maps, 0 reduces and no setup-cleanup).
  synchronized void completeEmptyJob() {
    jobComplete();
  }

  synchronized void completeSetup() {
    setupComplete();
  }

  void logSubmissionToJobHistory() throws IOException {
    // log job info
    String username = conf.getUser();
    if (username == null) { username = ""; }
    String jobname = conf.getJobName();
    if (jobname == null) { jobname = ""; }
    setUpLocalizedJobConf(conf, jobId);
    jobHistory.setupEventWriter(jobId, conf);
    JobSubmittedEvent jse =
        new JobSubmittedEvent(jobId, jobname, username, this.startTime,
            jobFile.toString(), status.getJobACLs());
    jobHistory.logEvent(jse, jobId);
   
  }

  TaskSplitMetaInfo[] createSplits(org.apache.hadoop.mapreduce.JobID jobId)
  throws IOException {
    TaskSplitMetaInfo[] allTaskSplitMetaInfo =
      SplitMetaInfoReader.readSplitMetaInfo(jobId, fs, conf, jobSubmitDir);
    return allTaskSplitMetaInfo;
  }

  /**
   * If authorization is enabled on the JobTracker, checks whether the user (in
   * the callerUGI) is authorized to perform the operation specify by
   * 'jobOperation' on the job.
   * <ul>
   * <li>The owner of the job can do any operation on the job</li>
   * <li>The superuser/supergroup of the JobTracker is always permitted to do
   * operations on any job.</li>
   * <li>For all other users/groups job-acls are checked</li>
   * </ul>
   *
   * @param callerUGI
   * @param jobOperation
   */
  void checkAccess(UserGroupInformation callerUGI, JobACL jobOperation)
      throws AccessControlException {
    jobtracker.getJobACLsManager().checkAccess(status, callerUGI, jobOperation);
  }

  /**
   * If the number of taks is greater than the configured value
   * throw an exception that will fail job initialization
   */
  void checkTaskLimits() throws IOException {
    int maxTasks = jobtracker.getMaxTasksPerJob();
    if (maxTasks > 0 && numMapTasks + numReduceTasks > maxTasks) {
      throw new IOException(
                "The number of tasks for this job " +
                (numMapTasks + numReduceTasks) +
                " exceeds the configured limit " + maxTasks);
    }
  }

  synchronized void createMapTasks(String jobFile,
      TaskSplitMetaInfo[] splits) {
    maps = new TaskInProgress[numMapTasks];
    for(int i=0; i < numMapTasks; ++i) {
      inputLength += splits[i].getInputDataLength();
      maps[i] = new TaskInProgress(jobId, jobFile,
                                   splits[i],
                                   jobtracker, conf, this,
                                   i, numSlotsPerMap);
    }
    LOG.info("Input size for job " + jobId + " = " + inputLength
        + ". Number of splits = " + splits.length);

  }

  synchronized void createReduceTasks(String jobFile) {
    this.reduces = new TaskInProgress[numReduceTasks];
    for (int i = 0; i < numReduceTasks; i++) {
      reduces[i] = new TaskInProgress(jobId, jobFile,
                                      numMapTasks, i,
                                      jobtracker, conf,
                                      this, numSlotsPerReduce);
      nonRunningReduces.add(reduces[i]);
    }
  }

 
  synchronized void initSetupCleanupTasks(String jobFile) {
    if (!jobSetupCleanupNeeded) {
      LOG.info("Setup/Cleanup not needed for job " + jobId);
      // nothing to initialize
      return;
    }
    // create cleanup two cleanup tips, one map and one reduce.
    cleanup = new TaskInProgress[2];

    // cleanup map tip. This map doesn't use any splits. Just assign an empty
    // split.
    TaskSplitMetaInfo emptySplit = JobSplit.EMPTY_TASK_SPLIT;
    cleanup[0] = new TaskInProgress(jobId, jobFile, emptySplit,
            jobtracker, conf, this, numMapTasks, 1);
    cleanup[0].setJobCleanupTask();

    // cleanup reduce tip.
    cleanup[1] = new TaskInProgress(jobId, jobFile, numMapTasks,
                       numReduceTasks, jobtracker, conf, this, 1);
    cleanup[1].setJobCleanupTask();

    // create two setup tips, one map and one reduce.
    setup = new TaskInProgress[2];

    // setup map tip. This map doesn't use any split. Just assign an empty
    // split.
    setup[0] = new TaskInProgress(jobId, jobFile, emptySplit,
            jobtracker, conf, this, numMapTasks + 1, 1);
    setup[0].setJobSetupTask();

    // setup reduce tip.
    setup[1] = new TaskInProgress(jobId, jobFile, numMapTasks,
                       numReduceTasks + 1, jobtracker, conf, this, 1);
    setup[1].setJobSetupTask();
  }
 
  void setupComplete() {
    status.setSetupProgress(1.0f);
    if (this.status.getRunState() == JobStatus.PREP) {
      changeStateTo(JobStatus.RUNNING);
      JobStatusChangedEvent jse =
        new JobStatusChangedEvent(profile.getJobID(),
         JobStatus.getJobRunState(JobStatus.RUNNING));
      jobHistory.logEvent(jse, profile.getJobID());
    }
  }

  /////////////////////////////////////////////////////
  // Accessors for the JobInProgress
  /////////////////////////////////////////////////////
  public JobProfile getProfile() {
    return profile;
  }
  public JobStatus getStatus() {
    return status;
  }
  public synchronized long getLaunchTime() {
    return launchTime;
  }
  public long getStartTime() {
    return startTime;
  }
  public long getFinishTime() {
    return finishTime;
  }
  public int desiredMaps() {
    return numMapTasks;
  }
  public synchronized int finishedMaps() {
    return finishedMapTasks;
  }
  public int desiredReduces() {
    return numReduceTasks;
  }
  public synchronized int runningMaps() {
    return runningMapTasks;
  }
  public synchronized int runningReduces() {
    return runningReduceTasks;
  }
  public synchronized int finishedReduces() {
    return finishedReduceTasks;
  }
  public synchronized int pendingMaps() {
    return numMapTasks - runningMapTasks - failedMapTIPs -
    finishedMapTasks + speculativeMapTasks;
  }
  public synchronized int pendingReduces() {
    return numReduceTasks - runningReduceTasks - failedReduceTIPs -
    finishedReduceTasks + speculativeReduceTasks;
  }
  public synchronized int getNumSlotsPerTask(TaskType taskType) {
    if (taskType == TaskType.MAP) {
      return numSlotsPerMap;
    } else if (taskType == TaskType.REDUCE) {
      return numSlotsPerReduce;
    } else {
      return 1;
    }
  }
  public JobPriority getPriority() {
    return this.priority;
  }
  public void setPriority(JobPriority priority) {
    if(priority == null) {
      priority = JobPriority.NORMAL;
    }
    synchronized (this) {
      this.priority = priority;
      status.setJobPriority(priority);
      // log and change to the job's priority
      JobPriorityChangeEvent prEvent =
        new JobPriorityChangeEvent(jobId, priority);
      
      jobHistory.logEvent(prEvent, jobId);
     
    }
  }

  // Update the job start/launch time (upon restart) and log to history
  synchronized void updateJobInfo(long startTime, long launchTime) {
    // log and change to the job's start/launch time
    this.startTime = startTime;
    this.launchTime = launchTime;
    JobInfoChangeEvent event =
      new JobInfoChangeEvent(jobId, startTime, launchTime);
    
    jobHistory.logEvent(event, jobId);
   
  }

  /**
   * Get the number of times the job has restarted
   */
  int getNumRestarts() {
    return restartCount;
  }
 
  long getInputLength() {
    return inputLength;
  }
  boolean isCleanupLaunched() {
    return launchedCleanup;
  }

  boolean isSetupLaunched() {
    return launchedSetup;
  }

  /**
   * Get all the tasks of the desired type in this job.
   * @param type {@link TaskType} of the tasks required
   * @return An array of {@link TaskInProgress} matching the given type.
   *         Returns an empty array if no tasks are found for the given type. 
   */
  TaskInProgress[] getTasks(TaskType type) {
    TaskInProgress[] tasks = null;
    switch (type) {
      case MAP:
      {
        tasks = maps;
      }
      break;
      case REDUCE:
      {
        tasks = reduces;
      }
      break;
      case JOB_SETUP:
      {
        tasks = setup;
      }
      break;
      case JOB_CLEANUP:
      {
        tasks = cleanup;
      }
      break;
      default:
      {
          tasks = new TaskInProgress[0];
      }
      break;
    }
    return tasks;
  }

  /**
   * Return the nonLocalRunningMaps
   * @return
   */
  Set<TaskInProgress> getNonLocalRunningMaps()
  {
    return nonLocalRunningMaps;
  }
 
  /**
   * Return the runningMapCache
   * @return
   */
  Map<Node, Set<TaskInProgress>> getRunningMapCache()
  {
    return runningMapCache;
  }
 
  /**
   * Return runningReduces
   * @return
   */
  Set<TaskInProgress> getRunningReduces()
  {
    return runningReduces;
  }
 
  /**
   * Get the job configuration
   * @return the job's configuration
   */
  JobConf getJobConf() {
    return conf;
  }
   
  /**
   * Return a vector of completed TaskInProgress objects
   */
  public synchronized Vector<TaskInProgress> reportTasksInProgress(boolean shouldBeMap,
                                                      boolean shouldBeComplete) {
   
    Vector<TaskInProgress> results = new Vector<TaskInProgress>();
    TaskInProgress tips[] = null;
    if (shouldBeMap) {
      tips = maps;
    } else {
      tips = reduces;
    }
    for (int i = 0; i < tips.length; i++) {
      if (tips[i].isComplete() == shouldBeComplete) {
        results.add(tips[i]);
      }
    }
    return results;
  }
 
  /**
   * Return a vector of cleanup TaskInProgress objects
   */
  public synchronized Vector<TaskInProgress> reportCleanupTIPs(
                                               boolean shouldBeComplete) {
   
    Vector<TaskInProgress> results = new Vector<TaskInProgress>();
    for (int i = 0; i < cleanup.length; i++) {
      if (cleanup[i].isComplete() == shouldBeComplete) {
        results.add(cleanup[i]);
      }
    }
    return results;
  }

  /**
   * Return a vector of setup TaskInProgress objects
   */
  public synchronized Vector<TaskInProgress> reportSetupTIPs(
                                               boolean shouldBeComplete) {
   
    Vector<TaskInProgress> results = new Vector<TaskInProgress>();
    for (int i = 0; i < setup.length; i++) {
      if (setup[i].isComplete() == shouldBeComplete) {
        results.add(setup[i]);
      }
    }
    return results;
  }

  ////////////////////////////////////////////////////
  // Status update methods
  ////////////////////////////////////////////////////

  /**
   * Assuming {@link JobTracker} is locked on entry.
   */
  public synchronized void updateTaskStatus(TaskInProgress tip,
                                            TaskStatus status) {

    double oldProgress = tip.getProgress();   // save old progress
    boolean wasRunning = tip.isRunning();
    boolean wasComplete = tip.isComplete();
    boolean wasPending = tip.isOnlyCommitPending();
    TaskAttemptID taskid = status.getTaskID();
    boolean wasAttemptRunning = tip.isAttemptRunning(taskid);

   
    // If the TIP is already completed and the task reports as SUCCEEDED then
    // mark the task as KILLED.
    // In case of task with no promotion the task tracker will mark the task
    // as SUCCEEDED.
    // User has requested to kill the task, but TT reported SUCCEEDED,
    // mark the task KILLED.
    if ((wasComplete || tip.wasKilled(taskid)) &&
        (status.getRunState() == TaskStatus.State.SUCCEEDED)) {
      status.setRunState(TaskStatus.State.KILLED);
    }
   
    // If the job is complete and a task has just reported its
    // state as FAILED_UNCLEAN/KILLED_UNCLEAN,
    // make the task's state FAILED/KILLED without launching cleanup attempt.
    // Note that if task is already a cleanup attempt,
    // we don't change the state to make sure the task gets a killTaskAction
    if ((this.isComplete() || jobFailed || jobKilled) &&
        !tip.isCleanupAttempt(taskid)) {
      if (status.getRunState() == TaskStatus.State.FAILED_UNCLEAN) {
        status.setRunState(TaskStatus.State.FAILED);
      } else if (status.getRunState() == TaskStatus.State.KILLED_UNCLEAN) {
        status.setRunState(TaskStatus.State.KILLED);
      }
    }
   
    boolean change = tip.updateStatus(status);
    if (change) {
      TaskStatus.State state = status.getRunState();
      // get the TaskTrackerStatus where the task ran
      TaskTracker taskTracker =
        this.jobtracker.getTaskTracker(tip.machineWhereTaskRan(taskid));
      TaskTrackerStatus ttStatus =
        (taskTracker == null) ? null : taskTracker.getStatus();
      String taskTrackerHttpLocation = null;

      if (null != ttStatus){
        String host;
        if (NetUtils.getStaticResolution(ttStatus.getHost()) != null) {
          host = NetUtils.getStaticResolution(ttStatus.getHost());
        } else {
          host = ttStatus.getHost();
        }
        taskTrackerHttpLocation = "http://" + host + ":"
            + ttStatus.getHttpPort();
      }

      TaskCompletionEvent taskEvent = null;
      if (state == TaskStatus.State.SUCCEEDED) {
        taskEvent = new TaskCompletionEvent(
                                            taskCompletionEventTracker,
                                            taskid,
                                            tip.idWithinJob(),
                                            status.getIsMap() &&
                                            !tip.isJobCleanupTask() &&
                                            !tip.isJobSetupTask(),
                                            TaskCompletionEvent.Status.SUCCEEDED,
                                            taskTrackerHttpLocation
                                           );
        taskEvent.setTaskRunTime((int)(status.getFinishTime()
                                       - status.getStartTime()));
        tip.setSuccessEventNumber(taskCompletionEventTracker);
      } else if (state == TaskStatus.State.COMMIT_PENDING) {
        // If it is the first attempt reporting COMMIT_PENDING
        // ask the task to commit.
        if (!wasComplete && !wasPending) {
          tip.doCommit(taskid);
        }
        return;
      } else if (state == TaskStatus.State.FAILED_UNCLEAN ||
                 state == TaskStatus.State.KILLED_UNCLEAN) {
        tip.incompleteSubTask(taskid, this.status);
        // add this task, to be rescheduled as cleanup attempt
        if (tip.isMapTask()) {
          mapCleanupTasks.add(taskid);
        } else {
          reduceCleanupTasks.add(taskid);
        }
        // Remove the task entry from jobtracker
        jobtracker.removeTaskEntry(taskid);
      }
      //For a failed task update the JT datastructures.
      else if (state == TaskStatus.State.FAILED ||
               state == TaskStatus.State.KILLED) {
        // Get the event number for the (possibly) previously successful
        // task. If there exists one, then set that status to OBSOLETE
        int eventNumber;
        if ((eventNumber = tip.getSuccessEventNumber()) != -1) {
          TaskCompletionEvent t =
            this.taskCompletionEvents.get(eventNumber);
          if (t.getTaskAttemptId().equals(taskid))
            t.setTaskStatus(TaskCompletionEvent.Status.OBSOLETE);
        }
       
        // Tell the job to fail the relevant task
        failedTask(tip, taskid, status, taskTracker,
                   wasRunning, wasComplete, wasAttemptRunning);

        // Did the task failure lead to tip failure?
        TaskCompletionEvent.Status taskCompletionStatus =
          (state == TaskStatus.State.FAILED ) ?
              TaskCompletionEvent.Status.FAILED :
              TaskCompletionEvent.Status.KILLED;
        if (tip.isFailed()) {
          taskCompletionStatus = TaskCompletionEvent.Status.TIPFAILED;
        }
        taskEvent = new TaskCompletionEvent(taskCompletionEventTracker,
                                            taskid,
                                            tip.idWithinJob(),
                                            status.getIsMap() &&
                                            !tip.isJobCleanupTask() &&
                                            !tip.isJobSetupTask(),
                                            taskCompletionStatus,
                                            taskTrackerHttpLocation
                                           );
      }         

      // Add the 'complete' task i.e. successful/failed
      // It _is_ safe to add the TaskCompletionEvent.Status.SUCCEEDED
      // *before* calling TIP.completedTask since:
      // a. One and only one task of a TIP is declared as a SUCCESS, the
      //    other (speculative tasks) are marked KILLED
      // b. TIP.completedTask *does not* throw _any_ exception at all.
      if (taskEvent != null) {
        this.taskCompletionEvents.add(taskEvent);
        taskCompletionEventTracker++;
        JobTrackerStatistics.TaskTrackerStat ttStat = jobtracker.
           getStatistics().getTaskTrackerStat(tip.machineWhereTaskRan(taskid));
        if(ttStat != null) { // ttStat can be null in case of lost tracker
          ttStat.incrTotalTasks();
        }
        if (state == TaskStatus.State.SUCCEEDED) {
          completedTask(tip, status);
          if(ttStat != null) {
            ttStat.incrSucceededTasks();
          }
        }
      }
    }
       
    //
    // Update JobInProgress status
    //
    if(LOG.isDebugEnabled()) {
      LOG.debug("Taking progress for " + tip.getTIPId() + " from " +
                 oldProgress + " to " + tip.getProgress());
    }
   
    if (!tip.isJobCleanupTask() && !tip.isJobSetupTask()) {
      double progressDelta = tip.getProgress() - oldProgress;
      if (tip.isMapTask()) {
          this.status.setMapProgress((float) (this.status.mapProgress() +
                                              progressDelta / maps.length));
      } else {
        this.status.setReduceProgress((float) (this.status.reduceProgress() +
                                           (progressDelta / reduces.length)));
      }
    }
  }

  /**
   * Returns the job-level counters.
   *
   * @return the job-level counters.
   */
  public synchronized Counters getJobCounters() {
    return jobCounters;
  }
 
  /**
   *  Returns map phase counters by summing over all map tasks in progress.
   */
  public synchronized Counters getMapCounters() {
    return incrementTaskCounters(new Counters(), maps);
  }
   
  /**
   *  Returns map phase counters by summing over all map tasks in progress.
   */
  public synchronized Counters getReduceCounters() {
    return incrementTaskCounters(new Counters(), reduces);
  }
   
  /**
   *  Returns the total job counters, by adding together the job,
   *  the map and the reduce counters.
   */
  public synchronized Counters getCounters() {
    Counters result = new Counters();
    result.incrAllCounters(getJobCounters());
    incrementTaskCounters(result, maps);
    return incrementTaskCounters(result, reduces);
  }
   
  /**
   * Increments the counters with the counters from each task.
   * @param counters the counters to increment
   * @param tips the tasks to add in to counters
   * @return counters the same object passed in as counters
   */
  private Counters incrementTaskCounters(Counters counters,
                                         TaskInProgress[] tips) {
    for (TaskInProgress tip : tips) {
      counters.incrAllCounters(tip.getCounters());
    }
    return counters;
  }

  /////////////////////////////////////////////////////
  // Create/manage tasks
  /////////////////////////////////////////////////////
  /**
   * Return a MapTask, if appropriate, to run on the given tasktracker
   */
  public synchronized Task obtainNewMapTask(TaskTrackerStatus tts,
                                            int clusterSize,
                                            int numUniqueHosts,
                                            int maxCacheLevel
                                           ) throws IOException {
    if (status.getRunState() != JobStatus.RUNNING) {
      LOG.info("Cannot create task split for " + profile.getJobID());
      return null;
    }
      
    int target = findNewMapTask(tts, clusterSize, numUniqueHosts,
        maxCacheLevel);
    if (target == -1) {
      return null;
    }
   
    Task result = maps[target].getTaskToRun(tts.getTrackerName());
    if (result != null) {
      addRunningTaskToTIP(maps[target], result.getTaskID(), tts, true);
    }

    return result;
  }
 
  /**
   * Return a MapTask, if appropriate, to run on the given tasktracker
   */
  public synchronized Task obtainNewMapTask(TaskTrackerStatus tts,
                                            int clusterSize,
                                            int numUniqueHosts
                                           ) throws IOException {
    return obtainNewMapTask(tts, clusterSize, numUniqueHosts, anyCacheLevel);
  }   

  /*
   * Return task cleanup attempt if any, to run on a given tracker
   */
  public Task obtainTaskCleanupTask(TaskTrackerStatus tts,
                                                 boolean isMapSlot)
  throws IOException {
    if (!tasksInited.get()) {
      return null;
    }
    synchronized (this) {
      if (this.status.getRunState() != JobStatus.RUNNING ||
          jobFailed || jobKilled) {
        return null;
      }
      String taskTracker = tts.getTrackerName();
      if (!shouldRunOnTaskTracker(taskTracker)) {
        return null;
      }
      TaskAttemptID taskid = null;
      TaskInProgress tip = null;
      if (isMapSlot) {
        if (!mapCleanupTasks.isEmpty()) {
          taskid = mapCleanupTasks.remove(0);
          tip = maps[taskid.getTaskID().getId()];
        }
      } else {
        if (!reduceCleanupTasks.isEmpty()) {
          taskid = reduceCleanupTasks.remove(0);
          tip = reduces[taskid.getTaskID().getId()];
        }
      }
      if (tip != null) {
        return tip.addRunningTask(taskid, taskTracker, true);
      }
      return null;
    }
  }
 
  public synchronized Task obtainNewLocalMapTask(TaskTrackerStatus tts,
                                                     int clusterSize,
                                                     int numUniqueHosts)
  throws IOException {
    if (!tasksInited.get()) {
      LOG.info("Cannot create task split for " + profile.getJobID());
      return null;
    }
 
    return obtainNewMapTask(tts, clusterSize, numUniqueHosts, maxLevel);
  }
 
  public synchronized Task obtainNewNonLocalMapTask(TaskTrackerStatus tts,
                                                    int clusterSize,
                                                    int numUniqueHosts)
  throws IOException {
    if (!tasksInited.get()) {
      LOG.info("Cannot create task split for " + profile.getJobID());
      return null;
    }
 
    return obtainNewMapTask(tts, clusterSize, numUniqueHosts,
        NON_LOCAL_CACHE_LEVEL);
  }
 
  /**
   * Return a CleanupTask, if appropriate, to run on the given tasktracker
   *
   */
  public Task obtainJobCleanupTask(TaskTrackerStatus tts,
                                             int clusterSize,
                                             int numUniqueHosts,
                                             boolean isMapSlot
                                            ) throws IOException {
    if(!tasksInited.get() || !jobSetupCleanupNeeded) {
      return null;
    }
   
    synchronized(this) {
      if (!canLaunchJobCleanupTask()) {
        return null;
      }
     
      String taskTracker = tts.getTrackerName();
      // Update the last-known clusterSize
      this.clusterSize = clusterSize;
      if (!shouldRunOnTaskTracker(taskTracker)) {
        return null;
      }
     
      List<TaskInProgress> cleanupTaskList = new ArrayList<TaskInProgress>();
      if (isMapSlot) {
        cleanupTaskList.add(cleanup[0]);
      } else {
        cleanupTaskList.add(cleanup[1]);
      }
      TaskInProgress tip = findTaskFromList(cleanupTaskList,
                             tts, numUniqueHosts, false);
      if (tip == null) {
        return null;
      }
     
      // Now launch the cleanupTask
      Task result = tip.getTaskToRun(tts.getTrackerName());
      if (result != null) {
        addRunningTaskToTIP(tip, result.getTaskID(), tts, true);
        if (jobFailed) {
          result.setJobCleanupTaskState(org.apache.hadoop.mapreduce.JobStatus
                .State.FAILED);
        } else if (jobKilled) {
          result.setJobCleanupTaskState(org.apache.hadoop.mapreduce.JobStatus
                .State.KILLED);
        } else {
          result.setJobCleanupTaskState(org.apache.hadoop.mapreduce
                .JobStatus.State.SUCCEEDED);
        }
      }
      return result;
    }
   
  }
 
  /**
   * Check whether cleanup task can be launched for the job.
   *
   * Cleanup task can be launched if it is not already launched
   * or job is Killed
   * or all maps and reduces are complete
   * @return true/false
   */
  private synchronized boolean canLaunchJobCleanupTask() {
    // check if the job is running
    if (status.getRunState() != JobStatus.RUNNING &&
        status.getRunState() != JobStatus.PREP) {
      return false;
    }
    // check if cleanup task has been launched already or if setup isn't
    // launched already. The later check is useful when number of maps is
    // zero.
    if (launchedCleanup || !isSetupFinished()) {
      return false;
    }
    // check if job has failed or killed
    if (jobKilled || jobFailed) {
      return true;
    }
    // Check if all maps and reducers have finished.
    boolean launchCleanupTask =
        ((finishedMapTasks + failedMapTIPs) == (numMapTasks));
    if (launchCleanupTask) {
      launchCleanupTask =
        ((finishedReduceTasks + failedReduceTIPs) == numReduceTasks);
    }
    return launchCleanupTask;
  }

  /**
   * Return a SetupTask, if appropriate, to run on the given tasktracker
   *
   */
  public Task obtainJobSetupTask(TaskTrackerStatus tts,
                                             int clusterSize,
                                             int numUniqueHosts,
                                             boolean isMapSlot
                                            ) throws IOException {
    if(!tasksInited.get() || !jobSetupCleanupNeeded) {
      return null;
    }
   
    synchronized(this) {
      if (!canLaunchSetupTask()) {
        return null;
      }
      String taskTracker = tts.getTrackerName();
      // Update the last-known clusterSize
      this.clusterSize = clusterSize;
      if (!shouldRunOnTaskTracker(taskTracker)) {
        return null;
      }
     
      List<TaskInProgress> setupTaskList = new ArrayList<TaskInProgress>();
      if (isMapSlot) {
        setupTaskList.add(setup[0]);
      } else {
        setupTaskList.add(setup[1]);
      }
      TaskInProgress tip = findTaskFromList(setupTaskList,
                             tts, numUniqueHosts, false);
      if (tip == null) {
        return null;
      }
     
      // Now launch the setupTask
      Task result = tip.getTaskToRun(tts.getTrackerName());
      if (result != null) {
        addRunningTaskToTIP(tip, result.getTaskID(), tts, true);
      }
      return result;
    }
  }
 
  public synchronized boolean scheduleReduces() {
    return finishedMapTasks >= completedMapsForReduceSlowstart;
  }
 
  /**
   * Check whether setup task can be launched for the job.
   *
   * Setup task can be launched after the tasks are inited
   * and Job is in PREP state
   * and if it is not already launched
   * or job is not Killed/Failed
   * @return true/false
   */
  private synchronized boolean canLaunchSetupTask() {
    return (tasksInited.get() && status.getRunState() == JobStatus.PREP &&
           !launchedSetup && !jobKilled && !jobFailed);
  }
 

  /**
   * Return a ReduceTask, if appropriate, to run on the given tasktracker.
   * We don't have cache-sensitivity for reduce tasks, as they
   *  work on temporary MapRed files. 
   */
  public synchronized Task obtainNewReduceTask(TaskTrackerStatus tts,
                                               int clusterSize,
                                               int numUniqueHosts
                                              ) throws IOException {
    if (status.getRunState() != JobStatus.RUNNING) {
      LOG.info("Cannot create task split for " + profile.getJobID());
      return null;
    }
   
    // Ensure we have sufficient map outputs ready to shuffle before
    // scheduling reduces
    if (!scheduleReduces()) {
      return null;
    }

    int  target = findNewReduceTask(tts, clusterSize, numUniqueHosts);
    if (target == -1) {
      return null;
    }
   
    Task result = reduces[target].getTaskToRun(tts.getTrackerName());
    if (result != null) {
      addRunningTaskToTIP(reduces[target], result.getTaskID(), tts, true);
    }

    return result;
  }
 
  // returns the (cache)level at which the nodes matches
  private int getMatchingLevelForNodes(Node n1, Node n2) {
    int count = 0;
    do {
      if (n1.equals(n2)) {
        return count;
      }
      ++count;
      n1 = n1.getParent();
      n2 = n2.getParent();
    } while (n1 != null);
    return this.maxLevel;
  }

  /**
   * Populate the data structures as a task is scheduled.
   *
   * Assuming {@link JobTracker} is locked on entry.
   *
   * @param tip The tip for which the task is added
   * @param id The attempt-id for the task
   * @param tts task-tracker status
   * @param isScheduled Whether this task is scheduled from the JT or has
   *        joined back upon restart
   */
  synchronized void addRunningTaskToTIP(TaskInProgress tip, TaskAttemptID id,
                                        TaskTrackerStatus tts,
                                        boolean isScheduled) {
    // Make an entry in the tip if the attempt is not scheduled i.e externally
    // added
    if (!isScheduled) {
      tip.addRunningTask(id, tts.getTrackerName());
    }
    final JobTrackerInstrumentation metrics = jobtracker.getInstrumentation();

    // keeping the earlier ordering intact
    TaskType name;
    String splits = "";
    Enum counter = null;
    if (tip.isJobSetupTask()) {
      launchedSetup = true;
      name = TaskType.JOB_SETUP;
    } else if (tip.isJobCleanupTask()) {
      launchedCleanup = true;
      name = TaskType.JOB_CLEANUP;
    } else if (tip.isMapTask()) {
      ++runningMapTasks;
      name = TaskType.MAP;
      counter = JobCounter.TOTAL_LAUNCHED_MAPS;
      splits = tip.getSplitNodes();
      if (tip.isSpeculating()) {
        speculativeMapTasks++;
        LOG.debug("Chosen speculative task, current speculativeMap task count: "
            + speculativeMapTasks);
      }
      metrics.launchMap(id);
    } else {
      ++runningReduceTasks;
      name = TaskType.REDUCE;
      counter = JobCounter.TOTAL_LAUNCHED_REDUCES;
      if (tip.isSpeculating()) {
        speculativeReduceTasks++;
        LOG.debug("Chosen speculative task, current speculativeReduce task count: "
          + speculativeReduceTasks);
      }
      metrics.launchReduce(id);
    }
    // Note that the logs are for the scheduled tasks only. Tasks that join on
    // restart has already their logs in place.
    if (tip.isFirstAttempt(id)) {
      TaskStartedEvent tse = new TaskStartedEvent(tip.getTIPId(),
          tip.getExecStartTime(),
          name, splits);
     
      jobHistory.logEvent(tse, tip.getJob().jobId);
     
    }
    if (!tip.isJobSetupTask() && !tip.isJobCleanupTask()) {
      jobCounters.incrCounter(counter, 1);
    }
   
    //TODO The only problem with these counters would be on restart.
    // The jobtracker updates the counter only when the task that is scheduled
    // if from a non-running tip and is local (data, rack ...). But upon restart
    // as the reports come from the task tracker, there is no good way to infer
    // when exactly to increment the locality counters. The only solution is to
    // increment the counters for all the tasks irrespective of
    //    - whether the tip is running or not
    //    - whether its a speculative task or not
    //
    // So to simplify, increment the data locality counter whenever there is
    // data locality.
    if (tip.isMapTask() && !tip.isJobSetupTask() && !tip.isJobCleanupTask()) {
      // increment the data locality counter for maps
      int level = getLocalityLevel(tip, tts);
      switch (level) {
      case 0 :
        LOG.info("Choosing data-local task " + tip.getTIPId());
        jobCounters.incrCounter(JobCounter.DATA_LOCAL_MAPS, 1);
        break;
      case 1:
        LOG.info("Choosing rack-local task " + tip.getTIPId());
        jobCounters.incrCounter(JobCounter.RACK_LOCAL_MAPS, 1);
        break;
      default :
        // check if there is any locality
        if (level != this.maxLevel) {
          LOG.info("Choosing cached task at level " + level + tip.getTIPId());
          jobCounters.incrCounter(JobCounter.OTHER_LOCAL_MAPS, 1);
        }
        break;
      }
    }
  }
   
  public static String convertTrackerNameToHostName(String trackerName) {
    // Ugly!
    // Convert the trackerName to it's host name
    int indexOfColon = trackerName.indexOf(":");
    String trackerHostName = (indexOfColon == -1) ?
      trackerName :
      trackerName.substring(0, indexOfColon);
    return trackerHostName.substring("tracker_".length());
  }
   
  /**
   * Note that a task has failed on a given tracker and add the tracker 
   * to the blacklist iff too many trackers in the cluster i.e.
   * (clusterSize * CLUSTER_BLACKLIST_PERCENT) haven't turned 'flaky' already.
   *
   * @param taskTracker task-tracker on which a task failed
   */
  synchronized void addTrackerTaskFailure(String trackerName,
                                          TaskTracker taskTracker) {
    if (flakyTaskTrackers < (clusterSize * CLUSTER_BLACKLIST_PERCENT)) {
      String trackerHostName = convertTrackerNameToHostName(trackerName);

      Integer trackerFailures = trackerToFailuresMap.get(trackerHostName);
      if (trackerFailures == null) {
        trackerFailures = 0;
      }
      trackerToFailuresMap.put(trackerHostName, ++trackerFailures);

      // Check if this tasktracker has turned 'flaky'
      if (trackerFailures.intValue() == conf.getMaxTaskFailuresPerTracker()) {
        ++flakyTaskTrackers;
       
        // Cancel reservations if appropriate
        if (taskTracker != null) {
          if (trackersReservedForMaps.containsKey(taskTracker)) {
            taskTracker.unreserveSlots(TaskType.MAP, this);
          }
          if (trackersReservedForReduces.containsKey(taskTracker)) {
            taskTracker.unreserveSlots(TaskType.REDUCE, this);
          }
        }
        LOG.info("TaskTracker at '" + trackerHostName + "' turned 'flaky'");
      }
    }
  }
 
  public synchronized void reserveTaskTracker(TaskTracker taskTracker,
                                              TaskType type, int numSlots) {
    Map<TaskTracker, FallowSlotInfo> map =
      (type == TaskType.MAP) ? trackersReservedForMaps : trackersReservedForReduces;
   
    long now = System.currentTimeMillis();
   
    FallowSlotInfo info = map.get(taskTracker);
    int reservedSlots = 0;
    if (info == null) {
      info = new FallowSlotInfo(now, numSlots);
      reservedSlots = numSlots;
    } else {
      // Increment metering info if the reservation is changing
      if (info.getNumSlots() != numSlots) {
        Enum<JobCounter> counter =
          (type == TaskType.MAP) ?
              JobCounter.FALLOW_SLOTS_MILLIS_MAPS :
              JobCounter.FALLOW_SLOTS_MILLIS_REDUCES;
        long fallowSlotMillis = (now - info.getTimestamp()) * info.getNumSlots();
        jobCounters.incrCounter(counter, fallowSlotMillis);
       
        // Update
        reservedSlots = numSlots - info.getNumSlots();
        info.setTimestamp(now);
        info.setNumSlots(numSlots);
      }
    }
    map.put(taskTracker, info);
    if (type == TaskType.MAP) {
      jobtracker.getInstrumentation().addReservedMapSlots(reservedSlots);
    }
    else {
      jobtracker.getInstrumentation().addReservedReduceSlots(reservedSlots);
    }
    jobtracker.incrementReservations(type, reservedSlots);
  }
 
  public synchronized void unreserveTaskTracker(TaskTracker taskTracker,
                                                TaskType type) {
    Map<TaskTracker, FallowSlotInfo> map =
      (type == TaskType.MAP) ? trackersReservedForMaps :
                               trackersReservedForReduces;

    FallowSlotInfo info = map.get(taskTracker);
    if (info == null) {
      LOG.warn("Cannot find information about fallow slots for " +
               taskTracker.getTrackerName());
      return;
    }
   
    long now = System.currentTimeMillis();

    Enum<JobCounter> counter =
      (type == TaskType.MAP) ?
          JobCounter.FALLOW_SLOTS_MILLIS_MAPS :
          JobCounter.FALLOW_SLOTS_MILLIS_REDUCES;
    long fallowSlotMillis = (now - info.getTimestamp()) * info.getNumSlots();
    jobCounters.incrCounter(counter, fallowSlotMillis);

    map.remove(taskTracker);
    if (type == TaskType.MAP) {
      jobtracker.getInstrumentation().decReservedMapSlots(info.getNumSlots());
    }
    else {
      jobtracker.getInstrumentation().decReservedReduceSlots(
        info.getNumSlots());
    }
    jobtracker.decrementReservations(type, info.getNumSlots());
  }
 
  public int getNumReservedTaskTrackersForMaps() {
    return trackersReservedForMaps.size();
  }
 
  public int getNumReservedTaskTrackersForReduces() {
    return trackersReservedForReduces.size();
  }
 
  private int getTrackerTaskFailures(String trackerName) {
    String trackerHostName = convertTrackerNameToHostName(trackerName);
    Integer failedTasks = trackerToFailuresMap.get(trackerHostName);
    return (failedTasks != null) ? failedTasks.intValue() : 0;
  }
   
  /**
   * Get the black listed trackers for the job
   *
   * @return List of blacklisted tracker names
   */
  List<String> getBlackListedTrackers() {
    List<String> blackListedTrackers = new ArrayList<String>();
    for (Map.Entry<String,Integer> e : trackerToFailuresMap.entrySet()) {
       if (e.getValue().intValue() >= conf.getMaxTaskFailuresPerTracker()) {
         blackListedTrackers.add(e.getKey());
       }
    }
    return blackListedTrackers;
  }
 
  /**
   * Get the no. of 'flaky' tasktrackers for a given job.
   *
   * @return the no. of 'flaky' tasktrackers for a given job.
   */
  int getNoOfBlackListedTrackers() {
    return flakyTaskTrackers;
  }
   
  /**
   * Get the information on tasktrackers and no. of errors which occurred
   * on them for a given job.
   *
   * @return the map of tasktrackers and no. of errors which occurred
   *         on them for a given job.
   */
  synchronized Map<String, Integer> getTaskTrackerErrors() {
    // Clone the 'trackerToFailuresMap' and return the copy
    Map<String, Integer> trackerErrors =
      new TreeMap<String, Integer>(trackerToFailuresMap);
    return trackerErrors;
  }

  /**
   * Remove a map TIP from the lists for running maps.
   * Called when a map fails/completes (note if a map is killed,
   * it won't be present in the list since it was completed earlier)
   * @param tip the tip that needs to be retired
   */
  private synchronized void retireMap(TaskInProgress tip) {
    if (runningMapCache == null) {
      LOG.warn("Running cache for maps missing!! "
               + "Job details are missing.");
      return;
    }
   
    String[] splitLocations = tip.getSplitLocations();

    // Remove the TIP from the list for running non-local maps
    if (splitLocations == null || splitLocations.length == 0) {
      nonLocalRunningMaps.remove(tip);
      return;
    }

    // Remove from the running map caches
    for(String host: splitLocations) {
      Node node = jobtracker.getNode(host);

      for (int j = 0; j < maxLevel; ++j) {
        Set<TaskInProgress> hostMaps = runningMapCache.get(node);
        if (hostMaps != null) {
          hostMaps.remove(tip);
          if (hostMaps.size() == 0) {
            runningMapCache.remove(node);
          }
        }
        node = node.getParent();
      }
    }
  }

  /**
   * Remove a reduce TIP from the list for running-reduces
   * Called when a reduce fails/completes
   * @param tip the tip that needs to be retired
   */
  private synchronized void retireReduce(TaskInProgress tip) {
    if (runningReduces == null) {
      LOG.warn("Running list for reducers missing!! "
               + "Job details are missing.");
      return;
    }
    runningReduces.remove(tip);
  }

  /**
   * Adds a map tip to the list of running maps.
   * @param tip the tip that needs to be scheduled as running
   */
  protected synchronized void scheduleMap(TaskInProgress tip) {
   
    runningMapTaskStats.add(0.0f);
    if (runningMapCache == null) {
      LOG.warn("Running cache for maps is missing!! "
               + "Job details are missing.");
      return;
    }
    String[] splitLocations = tip.getSplitLocations();

    // Add the TIP to the list of non-local running TIPs
    if (splitLocations == null || splitLocations.length == 0) {
      nonLocalRunningMaps.add(tip);
      return;
    }

    for(String host: splitLocations) {
      Node node = jobtracker.getNode(host);

      for (int j = 0; j < maxLevel; ++j) {
        Set<TaskInProgress> hostMaps = runningMapCache.get(node);
        if (hostMaps == null) {
          // create a cache if needed
          hostMaps = new LinkedHashSet<TaskInProgress>();
          runningMapCache.put(node, hostMaps);
        }
        hostMaps.add(tip);
        node = node.getParent();
      }
    }
  }
 
  /**
   * Adds a reduce tip to the list of running reduces
   * @param tip the tip that needs to be scheduled as running
   */
  protected synchronized void scheduleReduce(TaskInProgress tip) {
    runningReduceTaskStats.add(0.0f);
    if (runningReduces == null) {
      LOG.warn("Running cache for reducers missing!! "
               + "Job details are missing.");
      return;
    }
    runningReduces.add(tip);
  }
 
  /**
   * Adds the failed TIP in the front of the list for non-running maps
   * @param tip the tip that needs to be failed
   */
  private synchronized void failMap(TaskInProgress tip) {
    if (nonRunningMapCache == null) {
      LOG.warn("Non-running cache for maps missing!! "
               + "Job details are missing.");
      return;
    }

    // 1. Its added everywhere since other nodes (having this split local)
    //    might have removed this tip from their local cache
    // 2. Give high priority to failed tip - fail early

    String[] splitLocations = tip.getSplitLocations();

    // Add the TIP in the front of the list for non-local non-running maps
    if (splitLocations.length == 0) {
      nonLocalMaps.add(0, tip);
      return;
    }

    for(String host: splitLocations) {
      Node node = jobtracker.getNode(host);
     
      for (int j = 0; j < maxLevel; ++j) {
        List<TaskInProgress> hostMaps = nonRunningMapCache.get(node);
        if (hostMaps == null) {
          hostMaps = new LinkedList<TaskInProgress>();
          nonRunningMapCache.put(node, hostMaps);
        }
        hostMaps.add(0, tip);
        node = node.getParent();
      }
    }
  }
 
  /**
   * Adds a failed TIP in the front of the list for non-running reduces
   * @param tip the tip that needs to be failed
   */
  private synchronized void failReduce(TaskInProgress tip) {
    if (nonRunningReduces == null) {
      LOG.warn("Failed cache for reducers missing!! "
               + "Job details are missing.");
      return;
    }
    nonRunningReduces.add(0, tip);
  }
 
  /**
   * Find a non-running task in the passed list of TIPs
   * @param tips a collection of TIPs
   * @param ttStatus the status of tracker that has requested a task to run
   * @param numUniqueHosts number of unique hosts that run trask trackers
   * @param removeFailedTip whether to remove the failed tips
   */
  private synchronized TaskInProgress findTaskFromList(
      Collection<TaskInProgress> tips, TaskTrackerStatus ttStatus,
      int numUniqueHosts,
      boolean removeFailedTip) {
    Iterator<TaskInProgress> iter = tips.iterator();
    while (iter.hasNext()) {
      TaskInProgress tip = iter.next();

      // Select a tip if
      //   1. runnable   : still needs to be run and is not completed
      //   2. ~running   : no other node is running it
      //   3. earlier attempt failed : has not failed on this host
      //                               and has failed on all the other hosts
      // A TIP is removed from the list if
      // (1) this tip is scheduled
      // (2) if the passed list is a level 0 (host) cache
      // (3) when the TIP is non-schedulable (running, killed, complete)
      if (tip.isRunnable() && !tip.isRunning()) {
        // check if the tip has failed on this host
        if (!tip.hasFailedOnMachine(ttStatus.getHost()) ||
             tip.getNumberOfFailedMachines() >= numUniqueHosts) {
          // check if the tip has failed on all the nodes
          iter.remove();
          return tip;
        } else if (removeFailedTip) {
          // the case where we want to remove a failed tip from the host cache
          // point#3 in the TIP removal logic above
          iter.remove();
        }
      } else {
        // see point#3 in the comment above for TIP removal logic
        iter.remove();
      }
    }
    return null;
  }
 
  public boolean hasSpeculativeMaps() {
    return hasSpeculativeMaps;
  }

  public boolean hasSpeculativeReduces() {
    return hasSpeculativeReduces;
  }

  /**
   * Retrieve a task for speculation.
   * If a task slot becomes available and there are less than SpeculativeCap
   * speculative tasks running:
   *  1)Ignore the request if the TT's progressRate is < SlowNodeThreshold
   *  2)Choose candidate tasks - those tasks whose progress rate is below
   *    slowTaskThreshold * mean(progress-rates)
   *  3)Speculate task that's expected to complete last
   * @param list pool of tasks to choose from
   * @param taskTrackerName the name of the TaskTracker asking for a task
   * @param taskTrackerHost the hostname of the TaskTracker asking for a task
   * @param taskType the type of task (MAP/REDUCE) that we are considering
   * @return the TIP to speculatively re-execute
   */
  protected synchronized TaskInProgress findSpeculativeTask(
      Collection<TaskInProgress> list, String taskTrackerName,
      String taskTrackerHost, TaskType taskType) {
    if (list.isEmpty()) {
      return null;
    }
    long now = JobTracker.getClock().getTime();
    if (isSlowTracker(taskTrackerName) || atSpeculativeCap(list, taskType)) {
      return null;
    }
    // List of speculatable candidates, start with all, and chop it down
    ArrayList<TaskInProgress> candidates = new ArrayList<TaskInProgress>(list);
   
    Iterator<TaskInProgress> iter = candidates.iterator();
    while (iter.hasNext()) {
      TaskInProgress tip = iter.next();
      if (tip.hasRunOnMachine(taskTrackerHost, taskTrackerName) ||
          !tip.canBeSpeculated(now)) {
          //remove it from candidates
          iter.remove();
      }
    }
    //resort according to expected time till completion
    Comparator<TaskInProgress> LateComparator =
      new EstimatedTimeLeftComparator(now);
    Collections.sort(candidates, LateComparator);
    if (candidates.size() > 0 ) {
      TaskInProgress tip = candidates.get(0);
      if (LOG.isDebugEnabled()) {
        LOG.debug("Chose task " + tip.getTIPId() + ". Statistics: Task's : " +
            tip.getCurrentProgressRate(now) + " Job's : " +
            (tip.isMapTask() ? runningMapTaskStats : runningReduceTaskStats));
      }
      return tip;
    } else {
      return null;
    }
  }

  /**
   * Find new map task
   * @param tts The task tracker that is asking for a task
   * @param clusterSize The number of task trackers in the cluster
   * @param numUniqueHosts The number of hosts that run task trackers
   * @param maxCacheLevel The maximum topology level until which to schedule
   *                      maps.
   *                      A value of {@link #anyCacheLevel} implies any
   *                      available task (node-local, rack-local, off-switch and
   *                      speculative tasks).
   *                      A value of {@link #NON_LOCAL_CACHE_LEVEL} implies only
   *                      off-switch/speculative tasks should be scheduled.
   * @return the index in tasks of the selected task (or -1 for no task)
   */
  private synchronized int findNewMapTask(final TaskTrackerStatus tts,
                                          final int clusterSize,
                                          final int numUniqueHosts,
                                          final int maxCacheLevel) {
    String taskTrackerName = tts.getTrackerName();
    String taskTrackerHost = tts.getHost();
    if (numMapTasks == 0) {
      if(LOG.isDebugEnabled()) {
        LOG.debug("No maps to schedule for " + profile.getJobID());
      }
      return -1;
    }

    TaskInProgress tip = null;
   
    //
    // Update the last-known clusterSize
    //
    this.clusterSize = clusterSize;

    if (!shouldRunOnTaskTracker(taskTrackerName)) {
      return -1;
    }

    // Check to ensure this TaskTracker has enough resources to
    // run tasks from this job
    long outSize = resourceEstimator.getEstimatedMapOutputSize();
    long availSpace = tts.getResourceStatus().getAvailableSpace();
    if(availSpace < outSize) {
      LOG.warn("No room for map task. Node " + tts.getHost() +
               " has " + availSpace +
               " bytes free; but we expect map to take " + outSize);

      return -1; //see if a different TIP might work better.
    }
   
   
    // For scheduling a map task, we have two caches and a list (optional)
    //  I)   one for non-running task
    //  II)  one for running task (this is for handling speculation)
    //  III) a list of TIPs that have empty locations (e.g., dummy splits),
    //       the list is empty if all TIPs have associated locations

    // First a look up is done on the non-running cache and on a miss, a look
    // up is done on the running cache. The order for lookup within the cache:
    //   1. from local node to root [bottom up]
    //   2. breadth wise for all the parent nodes at max level

    // We fall to linear scan of the list (III above) if we have misses in the
    // above caches

    Node node = jobtracker.getNode(tts.getHost());
   
    //
    // I) Non-running TIP :
    //

    // 1. check from local node to the root [bottom up cache lookup]
    //    i.e if the cache is available and the host has been resolved
    //    (node!=null)
    if (node != null) {
      Node key = node;
      int level = 0;
      // maxCacheLevel might be greater than this.maxLevel if findNewMapTask is
      // called to schedule any task (local, rack-local, off-switch or speculative)
      // tasks or it might be NON_LOCAL_CACHE_LEVEL (i.e. -1) if findNewMapTask is
      //  (i.e. -1) if findNewMapTask is to only schedule off-switch/speculative
      // tasks
      int maxLevelToSchedule = Math.min(maxCacheLevel, maxLevel);
      for (level = 0;level < maxLevelToSchedule; ++level) {
        List <TaskInProgress> cacheForLevel = nonRunningMapCache.get(key);
        if (cacheForLevel != null) {
          tip = findTaskFromList(cacheForLevel, tts,
              numUniqueHosts,level == 0);
          if (tip != null) {
            // Add to running cache
            scheduleMap(tip);

            // remove the cache if its empty
            if (cacheForLevel.size() == 0) {
              nonRunningMapCache.remove(key);
            }

            return tip.getIdWithinJob();
          }
        }
        key = key.getParent();
      }
     
      // Check if we need to only schedule a local task (node-local/rack-local)
      if (level == maxCacheLevel) {
        return -1;
      }
    }

    //2. Search breadth-wise across parents at max level for non-running
    //   TIP if
    //     - cache exists and there is a cache miss
    //     - node information for the tracker is missing (tracker's topology
    //       info not obtained yet)

    // collection of node at max level in the cache structure
    Collection<Node> nodesAtMaxLevel = jobtracker.getNodesAtMaxLevel();

    // get the node parent at max level
    Node nodeParentAtMaxLevel =
      (node == null) ? null : JobTracker.getParentNode(node, maxLevel - 1);
   
    for (Node parent : nodesAtMaxLevel) {

      // skip the parent that has already been scanned
      if (parent == nodeParentAtMaxLevel) {
        continue;
      }

      List<TaskInProgress> cache = nonRunningMapCache.get(parent);
      if (cache != null) {
        tip = findTaskFromList(cache, tts, numUniqueHosts, false);
        if (tip != null) {
          // Add to the running cache
          scheduleMap(tip);

          // remove the cache if empty
          if (cache.size() == 0) {
            nonRunningMapCache.remove(parent);
          }
          LOG.info("Choosing a non-local task " + tip.getTIPId());
          return tip.getIdWithinJob();
        }
      }
    }

    // 3. Search non-local tips for a new task
    tip = findTaskFromList(nonLocalMaps, tts, numUniqueHosts, false);
    if (tip != null) {
      // Add to the running list
      scheduleMap(tip);

      LOG.info("Choosing a non-local task " + tip.getTIPId());
      return tip.getIdWithinJob();
    }

    //
    // II) Running TIP :
    //
    if (hasSpeculativeMaps) {
      tip = getSpeculativeMap(taskTrackerName, taskTrackerHost);
      if (tip != null) {
        return tip.getIdWithinJob();
      }
    }
   return -1;
  }

  private synchronized TaskInProgress getSpeculativeMap(String taskTrackerName,
      String taskTrackerHost) {

    //////// Populate allTips with all TaskInProgress
    Set<TaskInProgress> allTips = new HashSet<TaskInProgress>();
   
    // collection of node at max level in the cache structure
    Collection<Node> nodesAtMaxLevel = jobtracker.getNodesAtMaxLevel();
    // Add all tasks from max-level nodes breadth-wise
    for (Node parent : nodesAtMaxLevel) {
      Set<TaskInProgress> cache = runningMapCache.get(parent);
      if (cache != null) {
        allTips.addAll(cache);
      }
    }
    // Add all non-local TIPs
    allTips.addAll(nonLocalRunningMaps);
   
    ///////// Select a TIP to run on
    TaskInProgress tip = findSpeculativeTask(allTips, taskTrackerName,
        taskTrackerHost, TaskType.MAP);
   
    if (tip != null) {
      LOG.info("Choosing map task " + tip.getTIPId() +
          " for speculative execution");
    } else {
      LOG.debug("No speculative map task found for tracker " + taskTrackerName);
    }
    return tip;
  }
 
  /**
   * Find new reduce task
   * @param tts The task tracker that is asking for a task
   * @param clusterSize The number of task trackers in the cluster
   * @param numUniqueHosts The number of hosts that run task trackers
   * @return the index in tasks of the selected task (or -1 for no task)
   */
  private synchronized int findNewReduceTask(TaskTrackerStatus tts,
                                             int clusterSize,
                                             int numUniqueHosts) {
    String taskTrackerName = tts.getTrackerName();
    String taskTrackerHost = tts.getHost();
    if (numReduceTasks == 0) {
      if(LOG.isDebugEnabled()) {
        LOG.debug("No reduces to schedule for " + profile.getJobID());
      }
      return -1;
    }
    TaskInProgress tip = null;
   
    // Update the last-known clusterSize
    this.clusterSize = clusterSize;

    if (!shouldRunOnTaskTracker(taskTrackerName)) {
      return -1;
    }

    long outSize = resourceEstimator.getEstimatedReduceInputSize();
    long availSpace = tts.getResourceStatus().getAvailableSpace();
    if(availSpace < outSize) {
      LOG.warn("No room for reduce task. Node " + taskTrackerName + " has " +
                availSpace +
               " bytes free; but we expect reduce input to take " + outSize);

      return -1; //see if a different TIP might work better.
    }
   
    // 1. check for a never-executed reduce tip
    // reducers don't have a cache and so pass -1 to explicitly call that out
    tip = findTaskFromList(nonRunningReduces, tts, numUniqueHosts, false);
    if (tip != null) {
      scheduleReduce(tip);
      return tip.getIdWithinJob();
    }

    // 2. check for a reduce tip to be speculated
    if (hasSpeculativeReduces) {
      tip = getSpeculativeReduce(taskTrackerName, taskTrackerHost);
      if (tip != null) {
        return tip.getIdWithinJob();
      }
    }

    return -1;
  }

  private synchronized TaskInProgress getSpeculativeReduce(
      String taskTrackerName, String taskTrackerHost) {
    TaskInProgress tip = findSpeculativeTask(
        runningReduces, taskTrackerName, taskTrackerHost, TaskType.REDUCE);
    if (tip != null) {
      LOG.info("Choosing reduce task " + tip.getTIPId() +
          " for speculative execution");
    }else {
      LOG.debug("No speculative map task found for tracker " + taskTrackerHost);
    }
    return tip;
  }

    /**
     * Check to see if the maximum number of speculative tasks are
     * already being executed currently.
     * @param tasks the set of tasks to test
     * @param type the type of task (MAP/REDUCE) that we are considering
     * @return has the cap been reached?
     */
   private boolean atSpeculativeCap(Collection<TaskInProgress> tasks,
       TaskType type) {
     float numTasks = tasks.size();
     if (numTasks == 0){
       return true; // avoid divide by zero
     }
     int speculativeTaskCount = type == TaskType.MAP ? speculativeMapTasks
         : speculativeReduceTasks;
     //return true if totalSpecTask < max(10, 0.01 * total-slots,
     //                                   0.1 * total-running-tasks)

     if (speculativeTaskCount < MIN_SPEC_CAP) {
       return false; // at least one slow tracker's worth of slots(default=10)
     }
     ClusterStatus c = jobtracker.getClusterStatus(false);
     int numSlots = (type == TaskType.MAP ? c.getMaxMapTasks() : c.getMaxReduceTasks());
     if ((float)speculativeTaskCount < numSlots * MIN_SLOTS_CAP) {
       return false;
     }
     boolean atCap = (((float)(speculativeTaskCount)/numTasks) >= speculativeCap);
     if (LOG.isDebugEnabled()) {
       LOG.debug("SpeculativeCap is "+speculativeCap+", specTasks/numTasks is " +
           ((float)(speculativeTaskCount)/numTasks)+
           ", so atSpecCap() is returning "+atCap);
     }
     return atCap;
   }
 
  /**
   * A class for comparing the estimated time to completion of two tasks
   */
  private static class EstimatedTimeLeftComparator
  implements Comparator<TaskInProgress> {
    private long time;
    public EstimatedTimeLeftComparator(long now) {
      this.time = now;
    }
    /**
     * Estimated time to completion is measured as:
     *   % of task left to complete (1 - progress) / progress rate of the task.
     *
     * This assumes that tasks are linear in their progress, which is
     * often wrong, especially since progress for reducers is currently
     * calculated by evenly weighting their three stages (shuffle, sort, map)
     * which rarely account for 1/3 each. This should be fixed in the future
     * by calculating progressRate more intelligently or splitting these
     * multi-phase tasks into individual tasks.
     *
     * The ordering this comparator defines is: task1 < task2 if task1 is
     * estimated to finish farther in the future => compare(t1,t2) returns -1
     */
    public int compare(TaskInProgress tip1, TaskInProgress tip2) {
      //we have to use the Math.max in the denominator to avoid divide by zero
      //error because prog and progRate can both be zero (if one is zero,
      //the other one will be 0 too).
      //We use inverse of time_reminaing=[(1- prog) / progRate]
      //so that (1-prog) is in denom. because tasks can have arbitrarily
      //low progRates in practice (e.g. a task that is half done after 1000
      //seconds will have progRate of 0.0000005) so we would rather
      //use Math.maxnon (1-prog) by putting it in the denominator
      //which will cause tasks with prog=1 look 99.99% done instead of 100%
      //which is okay
      double t1 = tip1.getCurrentProgressRate(time) / Math.max(0.0001,
          1.0 - tip1.getProgress());
      double t2 = tip2.getCurrentProgressRate(time) / Math.max(0.0001,
          1.0 - tip2.getProgress());
      if (t1 < t2) return -1;
      else if (t2 < t1) return 1;
      else return 0;
    }
  }
 
  /**
   * Compares the ave progressRate of tasks that have finished on this
   * taskTracker to the ave of all succesfull tasks thus far to see if this
   * TT one is too slow for speculating.
   * slowNodeThreshold is used to determine the number of standard deviations
   * @param taskTracker the name of the TaskTracker we are checking
   * @return is this TaskTracker slow
   */
  protected boolean isSlowTracker(String taskTracker) {
    if (trackerMapStats.get(taskTracker) != null &&
        trackerMapStats.get(taskTracker).mean() -
        mapTaskStats.mean() > mapTaskStats.std()*slowNodeThreshold) {
      if (LOG.isDebugEnabled()) {
        LOG.debug("Tracker " + taskTracker +
            " declared slow. trackerMapStats.get(taskTracker).mean() :" + trackerMapStats.get(taskTracker).mean() +
            " mapTaskStats :" + mapTaskStats);
      }
      return true;
    }
    if (trackerReduceStats.get(taskTracker) != null &&
        trackerReduceStats.get(taskTracker).mean() -
        reduceTaskStats.mean() > reduceTaskStats.std()*slowNodeThreshold) {
      if (LOG.isDebugEnabled()) {
        LOG.debug("Tracker " + taskTracker +
            " declared slow. trackerReduceStats.get(taskTracker).mean() :" + trackerReduceStats.get(taskTracker).mean() +
            " reduceTaskStats :" + reduceTaskStats);
      }
      return true;
    }
    return false;
  }
 
  static class DataStatistics{
    private int count = 0;
    private double sum = 0;
    private double sumSquares = 0;
   
    public DataStatistics() {
    }
   
    public DataStatistics(double initNum) {
      this.count = 1;
      this.sum = initNum;
      this.sumSquares = initNum * initNum;
    }
   
    public void add(double newNum) {
      this.count++;
      this.sum += newNum;
      this.sumSquares += newNum * newNum;
    }

    public void updateStatistics(double old, double update) {
      sub(old);
      add(update);
    }
    private void sub(double oldNum) {
      this.count--;
      this.sum = Math.max(this.sum -= oldNum, 0.0d);
      this.sumSquares = Math.max(this.sumSquares -= oldNum * oldNum, 0.0d);
    }
   
    public double mean() {
      return sum/count;     
    }
 
    public double var() {
      // E(X^2) - E(X)^2
      return Math.max((sumSquares/count) - mean() * mean(), 0.0d);
    }
   
    public double std() {
      return Math.sqrt(this.var());
    }
   
    public String toString() {
      return "DataStatistics: count is " + count + ", sum is " + sum +
      ", sumSquares is " + sumSquares + " mean is " + mean() + " std() is " + std();
    }
   
  }
 
  private boolean shouldRunOnTaskTracker(String taskTracker) {
    //
    // Check if too many tasks of this job have failed on this
    // tasktracker prior to assigning it a new one.
    //
    int taskTrackerFailedTasks = getTrackerTaskFailures(taskTracker);
    if ((flakyTaskTrackers < (clusterSize * CLUSTER_BLACKLIST_PERCENT)) &&
        taskTrackerFailedTasks >= conf.getMaxTaskFailuresPerTracker()) {
      if (LOG.isDebugEnabled()) {
        String flakyTracker = convertTrackerNameToHostName(taskTracker);
        LOG.debug("Ignoring the black-listed tasktracker: '" + flakyTracker
                  + "' for assigning a new task");
      }
      return false;
    }
    return true;
  }
 

  /**
   * Metering: Occupied Slots * (Finish - Start)
   * @param tip {@link TaskInProgress} to be metered which just completed,
   *            cannot be <code>null</code>
   * @param status {@link TaskStatus} of the completed task, cannot be
   *               <code>null</code>
   */
  private void meterTaskAttempt(TaskInProgress tip, TaskStatus status) {
    JobCounter slotCounter =
      (tip.isMapTask()) ? JobCounter.SLOTS_MILLIS_MAPS :
                          JobCounter.SLOTS_MILLIS_REDUCES;
    jobCounters.incrCounter(slotCounter,
                            tip.getNumSlotsRequired() *
                            (status.getFinishTime() - status.getStartTime()));
  }
 
  /**
   * A taskid assigned to this JobInProgress has reported in successfully.
   */
  public synchronized boolean completedTask(TaskInProgress tip,
                                            TaskStatus status)
  {
    TaskAttemptID taskid = status.getTaskID();
    final JobTrackerInstrumentation metrics = jobtracker.getInstrumentation();
       
    // Metering
    meterTaskAttempt(tip, status);
   
    // Sanity check: is the TIP already complete?
    // This would not happen,
    // because no two tasks are SUCCEEDED at the same time.
    if (tip.isComplete()) {
      // Mark this task as KILLED
      tip.alreadyCompletedTask(taskid);

      // Let the JobTracker cleanup this taskid if the job isn't running
      if (this.status.getRunState() != JobStatus.RUNNING) {
        jobtracker.markCompletedTaskAttempt(status.getTaskTracker(), taskid);
      }
      return false;
    }
    boolean wasSpeculating = tip.isSpeculating(); //store this fact
    LOG.info("Task '" + taskid + "' has completed " + tip.getTIPId() +
             " successfully.");         
    // Mark the TIP as complete
    tip.completed(taskid);
    resourceEstimator.updateWithCompletedTask(status, tip);

    // Update jobhistory
    TaskTrackerStatus ttStatus =
      this.jobtracker.getTaskTrackerStatus(status.getTaskTracker());
    String trackerHostname = jobtracker.getNode(ttStatus.getHost()).toString();
    TaskType taskType = getTaskType(tip);

    TaskAttemptStartedEvent tse = new TaskAttemptStartedEvent(
        status.getTaskID(), taskType, status.getStartTime(),
        status.getTaskTracker(),  ttStatus.getHttpPort());
   
    jobHistory.logEvent(tse, status.getTaskID().getJobID());
   

    if (status.getIsMap()){
      MapAttemptFinishedEvent mfe = new MapAttemptFinishedEvent(
          status.getTaskID(), taskType, TaskStatus.State.SUCCEEDED.toString(),
          status.getMapFinishTime(),
          status.getFinishTime(),  trackerHostname,
          status.getStateString(),
          new org.apache.hadoop.mapreduce.Counters(status.getCounters()));
     
      jobHistory.logEvent(mfe,  status.getTaskID().getJobID());
     
    }else{
      ReduceAttemptFinishedEvent rfe = new ReduceAttemptFinishedEvent(
          status.getTaskID(), taskType, TaskStatus.State.SUCCEEDED.toString(),
          status.getShuffleFinishTime(),
          status.getSortFinishTime(), status.getFinishTime(),
          trackerHostname, status.getStateString(),
          new org.apache.hadoop.mapreduce.Counters(status.getCounters()));
     
      jobHistory.logEvent(rfe,  status.getTaskID().getJobID());
     
    }

    TaskFinishedEvent tfe = new TaskFinishedEvent(tip.getTIPId(),
        tip.getExecFinishTime(), taskType,
        TaskStatus.State.SUCCEEDED.toString(),
        new org.apache.hadoop.mapreduce.Counters(status.getCounters()));
   
    jobHistory.logEvent(tfe, tip.getJob().getJobID());
   
  
    if (tip.isJobSetupTask()) {
      // setup task has finished. kill the extra setup tip
      killSetupTip(!tip.isMapTask());
      setupComplete();
    } else if (tip.isJobCleanupTask()) {
      // cleanup task has finished. Kill the extra cleanup tip
      if (tip.isMapTask()) {
        // kill the reduce tip
        cleanup[1].kill();
      } else {
        cleanup[0].kill();
      }
      //
      // The Job is done
      // if the job is failed, then mark the job failed.
      if (jobFailed) {
        terminateJob(JobStatus.FAILED);
      }
      // if the job is killed, then mark the job killed.
      if (jobKilled) {
        terminateJob(JobStatus.KILLED);
      }
      else {
        jobComplete();
      }
      // The job has been killed/failed/successful
      // JobTracker should cleanup this task
      jobtracker.markCompletedTaskAttempt(status.getTaskTracker(), taskid);
    } else if (tip.isMapTask()) {
      runningMapTasks -= 1;
      finishedMapTasks += 1;
      metrics.completeMap(taskid);
      if (!tip.isJobSetupTask() && hasSpeculativeMaps) {
        updateTaskTrackerStats(tip,ttStatus,trackerMapStats,mapTaskStats);
      }
      // remove the completed map from the resp running caches
      retireMap(tip);
      if ((finishedMapTasks + failedMapTIPs) == (numMapTasks)) {
        this.status.setMapProgress(1.0f);
      }
    } else {
      runningReduceTasks -= 1;
      finishedReduceTasks += 1;
      metrics.completeReduce(taskid);
      if (!tip.isJobSetupTask() && hasSpeculativeReduces) {
        updateTaskTrackerStats(tip,ttStatus,trackerReduceStats,reduceTaskStats);
      }
      // remove the completed reduces from the running reducers set
      retireReduce(tip);
      if ((finishedReduceTasks + failedReduceTIPs) == (numReduceTasks)) {
        this.status.setReduceProgress(1.0f);
      }
    }
    decrementSpeculativeCount(wasSpeculating, tip);
    // is job complete?
    if (!jobSetupCleanupNeeded && canLaunchJobCleanupTask()) {
      jobComplete();
    }
    return true;
  }
 
  private void updateTaskTrackerStats(TaskInProgress tip, TaskTrackerStatus ttStatus,
      Map<String,DataStatistics> trackerStats, DataStatistics overallStats) {
    float tipDuration = tip.getExecFinishTime()-tip.getDispatchTime(tip.getSuccessfulTaskid());
    DataStatistics ttStats =
      trackerStats.get(ttStatus.getTrackerName());
    double oldMean = 0.0d;
    //We maintain the mean of TaskTrackers' means. That way, we get a single
    //data-point for every tracker (used in the evaluation in isSlowTracker)
    if (ttStats != null) {
      oldMean = ttStats.mean();
      ttStats.add(tipDuration);
      overallStats.updateStatistics(oldMean, ttStats.mean());
    } else {
      trackerStats.put(ttStatus.getTrackerName(),
          (ttStats = new DataStatistics(tipDuration)));
      overallStats.add(tipDuration);
    }
    if (LOG.isDebugEnabled()) {
      LOG.debug("Added mean of " +ttStats.mean() + " to trackerStats of type "+
          (tip.isMapTask() ? "Map" : "Reduce") +
          " on "+ttStatus.getTrackerName()+". DataStatistics is now: " +
          trackerStats.get(ttStatus.getTrackerName()));
    }
  }
 
  public void updateStatistics(double oldProg, double newProg, boolean isMap) {
    if (isMap) {  
      runningMapTaskStats.updateStatistics(oldProg, newProg);
    } else {
      runningReduceTaskStats.updateStatistics(oldProg, newProg);
    }
  }
 
  public DataStatistics getRunningTaskStatistics(boolean isMap) {
    if (isMap) {
      return runningMapTaskStats;
    } else {
      return runningReduceTaskStats;
    }
  }
 
  public float getSlowTaskThreshold() {
    return slowTaskThreshold;
  }

  /**
   * Job state change must happen thru this call
   */
  private void changeStateTo(int newState) {
    int oldState = this.status.getRunState();
    if (oldState == newState) {
      return; //old and new states are same
    }
    this.status.setRunState(newState);
   
    //update the metrics
    if (oldState == JobStatus.PREP) {
      this.jobtracker.getInstrumentation().decPrepJob(conf, jobId);
    } else if (oldState == JobStatus.RUNNING) {
      this.jobtracker.getInstrumentation().decRunningJob(conf, jobId);
    }
   
    if (newState == JobStatus.PREP) {
      this.jobtracker.getInstrumentation().addPrepJob(conf, jobId);
    } else if (newState == JobStatus.RUNNING) {
      this.jobtracker.getInstrumentation().addRunningJob(conf, jobId);
    }
   
  }

  /**
   * The job is done since all it's component tasks are either
   * successful or have failed.
   */
  private void jobComplete() {
    final JobTrackerInstrumentation metrics = jobtracker.getInstrumentation();
    //
    // All tasks are complete, then the job is done!
    //
    if (this.status.getRunState() == JobStatus.RUNNING ||
        this.status.getRunState() == JobStatus.PREP) {
      changeStateTo(JobStatus.SUCCEEDED);
      this.status.setCleanupProgress(1.0f);
      if (maps.length == 0) {
        this.status.setMapProgress(1.0f);
      }
      if (reduces.length == 0) {
        this.status.setReduceProgress(1.0f);
      }
      this.finishTime = JobTracker.getClock().getTime();
      this.status.setFinishTime(this.finishTime);
      LOG.info("Job " + this.status.getJobID() +
               " has completed successfully.");
     
      // Log the job summary (this should be done prior to logging to
      // job-history to ensure job-counters are in-sync
      JobSummary.logJobSummary(this, jobtracker.getClusterStatus(false));

      // Log job-history
      JobFinishedEvent jfe =
        new JobFinishedEvent(this.status.getJobID(),
          this.finishTime,
          this.finishedMapTasks,this.finishedReduceTasks, failedMapTasks,
          failedReduceTasks,
          new org.apache.hadoop.mapreduce.Counters(getMapCounters()),
          new org.apache.hadoop.mapreduce.Counters(getReduceCounters()),
          new org.apache.hadoop.mapreduce.Counters(getCounters()));
     
      jobHistory.logEvent(jfe, this.status.getJobID());
      jobHistory.closeWriter(this.status.getJobID());

      // Note that finalize will close the job history handles which garbage collect
      // might try to finalize
      garbageCollect();
     
      metrics.completeJob(this.conf, this.status.getJobID());
    }
  }
 
  private synchronized void terminateJob(int jobTerminationState) {
    if ((status.getRunState() == JobStatus.RUNNING) ||
        (status.getRunState() == JobStatus.PREP)) {

      this.finishTime = JobTracker.getClock().getTime();
      this.status.setMapProgress(1.0f);
      this.status.setReduceProgress(1.0f);
      this.status.setCleanupProgress(1.0f);
      this.status.setFinishTime(this.finishTime);

      if (jobTerminationState == JobStatus.FAILED) {
        changeStateTo(JobStatus.FAILED);
      } else {
        changeStateTo(JobStatus.KILLED);
      }
      // Log the job summary
      JobSummary.logJobSummary(this, jobtracker.getClusterStatus(false));

      JobUnsuccessfulCompletionEvent failedEvent =
        new JobUnsuccessfulCompletionEvent(this.status.getJobID(),
            finishTime,
            this.finishedMapTasks,
            this.finishedReduceTasks,
            JobStatus.getJobRunState(jobTerminationState));
     
      jobHistory.logEvent(failedEvent, this.status.getJobID());
      jobHistory.closeWriter(this.status.getJobID());

      garbageCollect();

      jobtracker.getInstrumentation().terminateJob(
          this.conf, this.status.getJobID());
      if (jobTerminationState == JobStatus.FAILED) {
        jobtracker.getInstrumentation().failedJob(
            this.conf, this.status.getJobID());
      } else {
        jobtracker.getInstrumentation().killedJob(
            this.conf, this.status.getJobID());
      }
    }
  }

  /**
   * Terminate the job and all its component tasks.
   * Calling this will lead to marking the job as failed/killed. Cleanup
   * tip will be launched. If the job has not inited, it will directly call
   * terminateJob as there is no need to launch cleanup tip.
   * This method is reentrant.
   * @param jobTerminationState job termination state
   */
  private synchronized void terminate(int jobTerminationState) {
    if(!tasksInited.get()) {
      //init could not be done, we just terminate directly.
      terminateJob(jobTerminationState);
      return;
    }

    if ((status.getRunState() == JobStatus.RUNNING) ||
         (status.getRunState() == JobStatus.PREP)) {
      LOG.info("Killing job '" + this.status.getJobID() + "'");
      if (jobTerminationState == JobStatus.FAILED) {
        if(jobFailed) {//reentrant
          return;
        }
        jobFailed = true;
      } else if (jobTerminationState == JobStatus.KILLED) {
        if(jobKilled) {//reentrant
          return;
        }
        jobKilled = true;
      }
      // clear all unclean tasks
      clearUncleanTasks();
      //
      // kill all TIPs.
      //
      for (int i = 0; i < setup.length; i++) {
        setup[i].kill();
      }
      for (int i = 0; i < maps.length; i++) {
        maps[i].kill();
      }
      for (int i = 0; i < reduces.length; i++) {
        reduces[i].kill();
      }
     
      if (!jobSetupCleanupNeeded) {
        terminateJob(jobTerminationState);
      }
    }
  }

  /**
   * Cancel all reservations since the job is done
   */
  private void cancelReservedSlots() {
    // Make a copy of the set of TaskTrackers to prevent a
    // ConcurrentModificationException ...
    Set<TaskTracker> tm =
      new HashSet<TaskTracker>(trackersReservedForMaps.keySet());
    for (TaskTracker tt : tm) {
      tt.unreserveSlots(TaskType.MAP, this);
    }

    Set<TaskTracker> tr =
      new HashSet<TaskTracker>(trackersReservedForReduces.keySet());
    for (TaskTracker tt : tr) {
      tt.unreserveSlots(TaskType.REDUCE, this);
    }
  }
 
  private void clearUncleanTasks() {
    TaskAttemptID taskid = null;
    TaskInProgress tip = null;
    while (!mapCleanupTasks.isEmpty()) {
      taskid = mapCleanupTasks.remove(0);
      tip = maps[taskid.getTaskID().getId()];
      updateTaskStatus(tip, tip.getTaskStatus(taskid));
    }
    while (!reduceCleanupTasks.isEmpty()) {
      taskid = reduceCleanupTasks.remove(0);
      tip = reduces[taskid.getTaskID().getId()];
      updateTaskStatus(tip, tip.getTaskStatus(taskid));
    }
  }

  /**
   * Kill the job and all its component tasks. This method should be called from
   * jobtracker and should return fast as it locks the jobtracker.
   */
  public void kill() {
    boolean killNow = false;
    synchronized(jobInitKillStatus) {
      jobInitKillStatus.killed = true;
      //if not in middle of init, terminate it now
      if(!jobInitKillStatus.initStarted || jobInitKillStatus.initDone) {
        //avoiding nested locking by setting flag
        killNow = true;
      }
    }
    if(killNow) {
      terminate(JobStatus.KILLED);
    }
  }
 
  /**
   * Fails the job and all its component tasks. This should be called only from
   * {@link JobInProgress} or {@link JobTracker}. Look at
   * {@link JobTracker#failJob(JobInProgress)} for more details.
   * Note that the job doesnt expect itself to be failed before its inited.
   * Only when the init is done (successfully or otherwise), the job can be
   * failed.
   */
  synchronized void fail() {
    terminate(JobStatus.FAILED);
  }
 
  private void decrementSpeculativeCount(boolean wasSpeculating,
      TaskInProgress tip) {
    if (wasSpeculating) {
      if (tip.isMapTask()) {
        speculativeMapTasks--;
        LOG.debug("Decremented count for " +
            tip.getTIPId()+"/"+tip.getJob().getJobID() +
          ". Current speculativeMap task count: " + speculativeMapTasks);
      } else {
        speculativeReduceTasks--;
        LOG.debug("Decremented count for " +
            tip.getTIPId()+"/"+tip.getJob().getJobID() +
          ". Current speculativeReduce task count: " + speculativeReduceTasks);
      }
    }
  }
 
  /**
   * A task assigned to this JobInProgress has reported in as failed.
   * Most of the time, we'll just reschedule execution.  However, after
   * many repeated failures we may instead decide to allow the entire
   * job to fail or succeed if the user doesn't care about a few tasks failing.
   *
   * Even if a task has reported as completed in the past, it might later
   * be reported as failed.  That's because the TaskTracker that hosts a map
   * task might die before the entire job can complete.  If that happens,
   * we need to schedule reexecution so that downstream reduce tasks can
   * obtain the map task's output.
   */
  private void failedTask(TaskInProgress tip, TaskAttemptID taskid,
                          TaskStatus status,
                          TaskTracker taskTracker, boolean wasRunning,
                          boolean wasComplete, boolean wasAttemptRunning) {
    // check if the TIP is already failed
    boolean wasFailed = tip.isFailed();
    boolean wasSpeculating = tip.isSpeculating();

    // Mark the taskid as FAILED or KILLED
    tip.incompleteSubTask(taskid, this.status);
    decrementSpeculativeCount(wasSpeculating, tip);
  
    boolean isRunning = tip.isRunning();
    boolean isComplete = tip.isComplete();

    if(wasAttemptRunning) {
      // We are decrementing counters without looking for isRunning ,
      // because we increment the counters when we obtain
      // new map task attempt or reduce task attempt.We do not really check
      // for tip being running.
      // Whenever we obtain new task attempt runningMapTasks incremented.
      // hence we are decrementing the same.     
      if(!tip.isJobCleanupTask() && !tip.isJobSetupTask()) {
        if(tip.isMapTask()) {
          runningMapTasks -= 1;
        } else {
          runningReduceTasks -= 1;
        }
      }
     
      // Metering
      meterTaskAttempt(tip, status);
    }
       
    //update running  count on task failure.
    if (wasRunning && !isRunning) {
      if (tip.isJobCleanupTask()) {
        launchedCleanup = false;
      } else if (tip.isJobSetupTask()) {
        launchedSetup = false;
      } else if (tip.isMapTask()) {
        // remove from the running queue and put it in the non-running cache
        // if the tip is not complete i.e if the tip still needs to be run
        if (!isComplete) {
          retireMap(tip);
          failMap(tip);
        }
      } else {
        // remove from the running queue and put in the failed queue if the tip
        // is not complete
        if (!isComplete) {
          retireReduce(tip);
          failReduce(tip);
        }
      }
    }
       
    // The case when the map was complete but the task tracker went down.
    // However, we don't need to do any metering here...
    if (wasComplete && !isComplete) {
      if (tip.isMapTask()) {
        // Put the task back in the cache. This will help locality for cases
        // where we have a different TaskTracker from the same rack/switch
        // asking for a task.
        // We bother about only those TIPs that were successful
        // earlier (wasComplete and !isComplete)
        // (since they might have been removed from the cache of other
        // racks/switches, if the input split blocks were present there too)
        failMap(tip);
        finishedMapTasks -= 1;
      }
    }
       
    // update job history
    // get taskStatus from tip
    TaskStatus taskStatus = tip.getTaskStatus(taskid);
    String taskTrackerName = taskStatus.getTaskTracker();
    String taskTrackerHostName = convertTrackerNameToHostName(taskTrackerName);
    int taskTrackerPort = -1;
    TaskTrackerStatus taskTrackerStatus =
      (taskTracker == null) ? null : taskTracker.getStatus();
    if (taskTrackerStatus != null) {
      taskTrackerPort = taskTrackerStatus.getHttpPort();
    }
    long startTime = taskStatus.getStartTime();
    long finishTime = taskStatus.getFinishTime();
    List<String> taskDiagnosticInfo = tip.getDiagnosticInfo(taskid);
    String diagInfo = taskDiagnosticInfo == null ? "" :
      StringUtils.arrayToString(taskDiagnosticInfo.toArray(new String[0]));
    TaskType taskType = getTaskType(tip);
    TaskAttemptStartedEvent tse = new TaskAttemptStartedEvent(
        taskid, taskType, startTime, taskTrackerName, taskTrackerPort);
   
    jobHistory.logEvent(tse, taskid.getJobID());
  
    TaskAttemptUnsuccessfulCompletionEvent tue =
      new TaskAttemptUnsuccessfulCompletionEvent(taskid,
          taskType, taskStatus.getRunState().toString(),
          finishTime,
          taskTrackerHostName, diagInfo);
    jobHistory.logEvent(tue, taskid.getJobID());
       
    // After this, try to assign tasks with the one after this, so that
    // the failed task goes to the end of the list.
    if (!tip.isJobCleanupTask() && !tip.isJobSetupTask()) {
      if (tip.isMapTask()) {
        failedMapTasks++;
      } else {
        failedReduceTasks++;
      }
    }
           
    //
    // Note down that a task has failed on this tasktracker
    //
    if (status.getRunState() == TaskStatus.State.FAILED) {
      addTrackerTaskFailure(taskTrackerName, taskTracker);
    }
       
    //
    // Let the JobTracker know that this task has failed
    //
    jobtracker.markCompletedTaskAttempt(status.getTaskTracker(), taskid);

    //
    // Check if we need to kill the job because of too many failures or
    // if the job is complete since all component tasks have completed

    // We do it once per TIP and that too for the task that fails the TIP
    if (!wasFailed && tip.isFailed()) {
      //
      // Allow upto 'mapFailuresPercent' of map tasks to fail or
      // 'reduceFailuresPercent' of reduce tasks to fail
      //
      boolean killJob = tip.isJobCleanupTask() || tip.isJobSetupTask() ? true :
                        tip.isMapTask() ?
            ((++failedMapTIPs*100) > (mapFailuresPercent*numMapTasks)) :
            ((++failedReduceTIPs*100) > (reduceFailuresPercent*numReduceTasks));
     
      if (killJob) {
        LOG.info("Aborting job " + profile.getJobID());
        TaskFailedEvent tfe =
          new TaskFailedEvent(tip.getTIPId(), finishTime, taskType, diagInfo,
              TaskStatus.State.FAILED.toString(),
              null);
       
        jobHistory.logEvent(tfe, tip.getJob().getJobID());
       
        if (tip.isJobCleanupTask()) {
          // kill the other tip
          if (tip.isMapTask()) {
            cleanup[1].kill();
          } else {
            cleanup[0].kill();
          }
          terminateJob(JobStatus.FAILED);
        } else {
          if (tip.isJobSetupTask()) {
            // kill the other tip
            killSetupTip(!tip.isMapTask());
          }
          fail();
        }
      }
     
      //
      // Update the counters
      //
      if (!tip.isJobCleanupTask() && !tip.isJobSetupTask()) {
        if (tip.isMapTask()) {
          jobCounters.incrCounter(JobCounter.NUM_FAILED_MAPS, 1);
        } else {
          jobCounters.incrCounter(JobCounter.NUM_FAILED_REDUCES, 1);
        }
      }
    }
  }

  void killSetupTip(boolean isMap) {
    if (isMap) {
      setup[0].kill();
    } else {
      setup[1].kill();
    }
  }

  boolean isSetupFinished() {
    // if there is no setup to be launched, consider setup is finished. 
    if ((tasksInited.get() && setup.length == 0) ||
        setup[0].isComplete() || setup[0].isFailed() || setup[1].isComplete()
        || setup[1].isFailed()) {
      return true;
    }
    return false;
  }

  /**
   * Fail a task with a given reason, but without a status object.
   *
   * Assuming {@link JobTracker} is locked on entry.
   *
   * @param tip The task's tip
   * @param taskid The task id
   * @param reason The reason that the task failed
   * @param trackerName The task tracker the task failed on
   */
  public synchronized void failedTask(TaskInProgress tip, TaskAttemptID taskid,
      String reason, TaskStatus.Phase phase, TaskStatus.State state,
                         String trackerName) {
    TaskStatus status = TaskStatus.createTaskStatus(tip.isMapTask(),
                                                    taskid,
                                                    0.0f,
                                                    tip.isMapTask() ?
                                                        numSlotsPerMap :
                                                        numSlotsPerReduce,
                                                    state,
                                                    reason,
                                                    reason,
                                                    trackerName, phase,
                                                    new Counters());
    // update the actual start-time of the attempt
    TaskStatus oldStatus = tip.getTaskStatus(taskid);
    long startTime = oldStatus == null
                     ? JobTracker.getClock().getTime()
                     : oldStatus.getStartTime();
    status.setStartTime(startTime);
    status.setFinishTime(JobTracker.getClock().getTime());
    boolean wasComplete = tip.isComplete();
    updateTaskStatus(tip, status);
    boolean isComplete = tip.isComplete();
    if (wasComplete && !isComplete) { // mark a successful tip as failed
      TaskType taskType = getTaskType(tip);
      TaskFailedEvent tfe =
        new TaskFailedEvent(tip.getTIPId(), tip.getExecFinishTime(), taskType,
            reason, TaskStatus.State.FAILED.toString(),
            taskid);
     
        jobHistory.logEvent(tfe, tip.getJob().getJobID());
     
    }
  }
      
                          
  /**
   * The job is dead.  We're now GC'ing it, getting rid of the job
   * from all tables.  Be sure to remove all of this job's tasks
   * from the various tables.
   */
   void garbageCollect() {
     synchronized(this) {
       // Cancel task tracker reservation
       cancelReservedSlots();


       // Let the JobTracker know that a job is complete
       jobtracker.getInstrumentation().decWaitingMaps(getJobID(), pendingMaps());
       jobtracker.getInstrumentation().decWaitingReduces(getJobID(), pendingReduces());
       jobtracker.storeCompletedJob(this);
       jobtracker.finalizeJob(this);

       try {
         // Definitely remove the local-disk copy of the job file
         if (localJobFile != null) {
           localFs.delete(localJobFile, true);
           localJobFile = null;
         }

         Path tempDir = jobtracker.getSystemDirectoryForJob(getJobID());
         new CleanupQueue().addToQueue(new PathDeletionContext(
             jobtracker.getFileSystem(), tempDir.toUri().getPath()));
       } catch (IOException e) {
         LOG.warn("Error cleaning up "+profile.getJobID()+": "+e);
       }

       // free up the memory used by the data structures
       this.nonRunningMapCache = null;
       this.runningMapCache = null;
       this.nonRunningReduces = null;
       this.runningReduces = null;

     }
     // remove jobs delegation tokens
     if(conf.getBoolean(MRJobConfig.JOB_CANCEL_DELEGATION_TOKEN, true)) {
       DelegationTokenRenewal.removeDelegationTokenRenewalForJob(jobId);
     } // else don't remove it.May be used by spawned tasks
   }

  /**
   * Return the TaskInProgress that matches the tipid.
   */
  public synchronized TaskInProgress getTaskInProgress(TaskID tipid) {
    if (tipid.getTaskType() == TaskType.MAP) {
      // cleanup map tip
      if (cleanup.length > 0 && tipid.equals(cleanup[0].getTIPId())) {
        return cleanup[0];
      }
      // setup map tip
      if (setup.length > 0 && tipid.equals(setup[0].getTIPId())) {
        return setup[0];
      }
      for (int i = 0; i < maps.length; i++) {
        if (tipid.equals(maps[i].getTIPId())){
          return maps[i];
        }
      }
    } else {
      // cleanup reduce tip
      if (cleanup.length > 0 && tipid.equals(cleanup[1].getTIPId())) {
        return cleanup[1];
      }
      // setup reduce tip
      if (setup.length > 0 && tipid.equals(setup[1].getTIPId())) {
        return setup[1];
      }
      for (int i = 0; i < reduces.length; i++) {
        if (tipid.equals(reduces[i].getTIPId())){
          return reduces[i];
        }
      }
    }
    return null;
  }
   
  /**
   * Find the details of someplace where a map has finished
   * @param mapId the id of the map
   * @return the task status of the completed task
   */
  public synchronized TaskStatus findFinishedMap(int mapId) {
    TaskInProgress tip = maps[mapId];
    if (tip.isComplete()) {
      TaskStatus[] statuses = tip.getTaskStatuses();
      for(int i=0; i < statuses.length; i++) {
        if (statuses[i].getRunState() == TaskStatus.State.SUCCEEDED) {
          return statuses[i];
        }
      }
    }
    return null;
  }
 
  synchronized int getNumTaskCompletionEvents() {
    return taskCompletionEvents.size();
  }
   
  synchronized public TaskCompletionEvent[] getTaskCompletionEvents(
                                                                    int fromEventId, int maxEvents) {
    TaskCompletionEvent[] events = TaskCompletionEvent.EMPTY_ARRAY;
    if (taskCompletionEvents.size() > fromEventId) {
      int actualMax = Math.min(maxEvents,
                               (taskCompletionEvents.size() - fromEventId));
      events = taskCompletionEvents.subList(fromEventId, actualMax + fromEventId).toArray(events);       
    }
    return events;
  }
 
  synchronized void fetchFailureNotification(TaskInProgress tip,
                                             TaskAttemptID mapTaskId,
                                             String trackerName) {
    Integer fetchFailures = mapTaskIdToFetchFailuresMap.get(mapTaskId);
    fetchFailures = (fetchFailures == null) ? 1 : (fetchFailures+1);
    mapTaskIdToFetchFailuresMap.put(mapTaskId, fetchFailures);
    LOG.info("Failed fetch notification #" + fetchFailures + " for task " +
            mapTaskId);
   
    float failureRate = (float)fetchFailures / runningReduceTasks;
    // declare faulty if fetch-failures >= max-allowed-failures
    boolean isMapFaulty = (failureRate >= MAX_ALLOWED_FETCH_FAILURES_PERCENT)
                          ? true
                          : false;
    if (fetchFailures >= MAX_FETCH_FAILURES_NOTIFICATIONS
        && isMapFaulty) {
      LOG.info("Too many fetch-failures for output of task: " + mapTaskId
               + " ... killing it");
     
      failedTask(tip, mapTaskId, "Too many fetch-failures",                           
                 (tip.isMapTask() ? TaskStatus.Phase.MAP :
                                    TaskStatus.Phase.REDUCE),
                 TaskStatus.State.FAILED, trackerName);
     
      mapTaskIdToFetchFailuresMap.remove(mapTaskId);
    }
  }
 
  /**
   * @return The JobID of this JobInProgress.
   */
  public JobID getJobID() {
    return jobId;
  }
 
  public synchronized Object getSchedulingInfo() {
    return this.schedulingInfo;
  }
 
  public synchronized void setSchedulingInfo(Object schedulingInfo) {
    this.schedulingInfo = schedulingInfo;
    this.status.setSchedulingInfo(schedulingInfo.toString());
  }
 
  /**
   * To keep track of kill and initTasks status of this job. initTasks() take
   * a lock on JobInProgress object. kill should avoid waiting on
   * JobInProgress lock since it may take a while to do initTasks().
   */
  private static class JobInitKillStatus {
    //flag to be set if kill is called
    boolean killed;
   
    boolean initStarted;
    boolean initDone;
  }

  boolean isComplete() {
    return status.isJobComplete();
  }
 
  /**
   * Get the task type for logging it to {@link JobHistory}.
   */
  private TaskType getTaskType(TaskInProgress tip) {
    if (tip.isJobCleanupTask()) {
      return TaskType.JOB_CLEANUP;
    } else if (tip.isJobSetupTask()) {
      return TaskType.JOB_SETUP;
    } else if (tip.isMapTask()) {
      return TaskType.MAP;
    } else {
      return TaskType.REDUCE;
    }
  }
 
  /**
   * Get the level of locality that a given task would have if launched on
   * a particular TaskTracker. Returns 0 if the task has data on that machine,
   * 1 if it has data on the same rack, etc (depending on number of levels in
   * the network hierarchy).
   */
  int getLocalityLevel(TaskInProgress tip, TaskTrackerStatus tts) {
    Node tracker = jobtracker.getNode(tts.getHost());
    int level = this.maxLevel;
    // find the right level across split locations
    for (String local : maps[tip.getIdWithinJob()].getSplitLocations()) {
      Node datanode = jobtracker.getNode(local);
      int newLevel = this.maxLevel;
      if (tracker != null && datanode != null) {
        newLevel = getMatchingLevelForNodes(tracker, datanode);
      }
      if (newLevel < level) {
        level = newLevel;
        // an optimization
        if (level == 0) {
          break;
        }
      }
    }
    return level;
  }
 
  /**
   * Test method to set the cluster sizes
   */
  void setClusterSize(int clusterSize) {
    this.clusterSize = clusterSize;
  }

  static class JobSummary {
    static final Log LOG = LogFactory.getLog(JobSummary.class);
   
    // Escape sequences
    static final char EQUALS = '=';
    static final char[] charsToEscape =
      {StringUtils.COMMA, EQUALS, StringUtils.ESCAPE_CHAR};
   
    /**
     * Log a summary of the job's runtime.
     *
     * @param job {@link JobInProgress} whose summary is to be logged, cannot
     *            be <code>null</code>.
     * @param cluster {@link ClusterStatus} of the cluster on which the job was
     *                run, cannot be <code>null</code>
     */
    public static void logJobSummary(JobInProgress job, ClusterStatus cluster) {
      JobStatus status = job.getStatus();
      JobProfile profile = job.getProfile();
      String user = StringUtils.escapeString(profile.getUser(),
                                             StringUtils.ESCAPE_CHAR,
                                             charsToEscape);
      String queue = StringUtils.escapeString(profile.getQueueName(),
                                              StringUtils.ESCAPE_CHAR,
                                              charsToEscape);
      Counters jobCounters = job.getJobCounters();
      long mapSlotSeconds =
        (jobCounters.getCounter(JobCounter.SLOTS_MILLIS_MAPS) +
         jobCounters.getCounter(JobCounter.FALLOW_SLOTS_MILLIS_MAPS)) / 1000;
      long reduceSlotSeconds =
        (jobCounters.getCounter(JobCounter.SLOTS_MILLIS_REDUCES) +
         jobCounters.getCounter(JobCounter.FALLOW_SLOTS_MILLIS_REDUCES)) / 1000;

      LOG.info("jobId=" + job.getJobID() + StringUtils.COMMA +
               "submitTime" + EQUALS + job.getStartTime() + StringUtils.COMMA +
               "launchTime" + EQUALS + job.getLaunchTime() + StringUtils.COMMA +
               "finishTime" + EQUALS + job.getFinishTime() + StringUtils.COMMA +
               "numMaps" + EQUALS + job.getTasks(TaskType.MAP).length +
                           StringUtils.COMMA +
               "numSlotsPerMap" + EQUALS + job.getNumSlotsPerMap() +
                                  StringUtils.COMMA +
               "numReduces" + EQUALS + job.getTasks(TaskType.REDUCE).length +
                              StringUtils.COMMA +
               "numSlotsPerReduce" + EQUALS + job.getNumSlotsPerReduce() +
                                     StringUtils.COMMA +
               "user" + EQUALS + user + StringUtils.COMMA +
               "queue" + EQUALS + queue + StringUtils.COMMA +
               "status" + EQUALS +
                          JobStatus.getJobRunState(status.getRunState()) +
                          StringUtils.COMMA +
               "mapSlotSeconds" + EQUALS + mapSlotSeconds + StringUtils.COMMA +
               "reduceSlotsSeconds" + EQUALS + reduceSlotSeconds  +
                                      StringUtils.COMMA +
               "clusterMapCapacity" + EQUALS + cluster.getMaxMapTasks() +
                                      StringUtils.COMMA +
               "clusterReduceCapacity" + EQUALS + cluster.getMaxReduceTasks()
      );
    }
  }
 
  /**
   * Creates the localized copy of job conf
   * @param jobConf
   * @param id
   */
  void setUpLocalizedJobConf(JobConf jobConf,
      org.apache.hadoop.mapreduce.JobID id) {
    String localJobFilePath = jobtracker.getLocalJobFilePath(id);
    File localJobFile = new File(localJobFilePath);
    FileOutputStream jobOut = null;
    try {
      jobOut = new FileOutputStream(localJobFile);
      jobConf.writeXml(jobOut);
      if (LOG.isDebugEnabled()) {
        LOG.debug("Job conf for " + id + " stored at "
            + localJobFile.getAbsolutePath());
      }
    } catch (IOException ioe) {
      LOG.error("Failed to store job conf on the local filesystem ", ioe);
    } finally {
      if (jobOut != null) {
        try {
          jobOut.close();
        } catch (IOException ie) {
          LOG.info("Failed to close the job configuration file "
              + StringUtils.stringifyException(ie));
        }
      }
    }
  }

  /**
   * Deletes localized copy of job conf
   */
  void cleanupLocalizedJobConf(org.apache.hadoop.mapreduce.JobID id) {
    String localJobFilePath = jobtracker.getLocalJobFilePath(id);
    File f = new File (localJobFilePath);
    LOG.info("Deleting localized job conf at " + f);
    if (!f.delete()) {
      LOG.debug("Failed to delete file " + f);
    }
  }
 
  /**
   * generate job token and save it into the file
   * @throws IOException
   */
  private void generateAndStoreTokens() throws IOException{
    Path jobDir = jobtracker.getSystemDirectoryForJob(jobId);
    Path keysFile = new Path(jobDir, TokenCache.JOB_TOKEN_HDFS_FILE);
    // we need to create this file using the jobtracker's filesystem
    FSDataOutputStream os = jobtracker.getFileSystem().create(keysFile);
   
    //create JobToken file and write token to it
    JobTokenIdentifier identifier = new JobTokenIdentifier(new Text(jobId
        .toString()));
    Token<JobTokenIdentifier> token = new Token<JobTokenIdentifier>(identifier,
        jobtracker.getJobTokenSecretManager());
    token.setService(identifier.getJobId());
   
    // add this token to the tokenStorage
    if(tokenStorage == null)
      tokenStorage = new TokenStorage();
   
    TokenCache.setJobToken(token, tokenStorage);
   
    // write TokenStorage out
    tokenStorage.write(os);
    os.close();
    LOG.info("jobToken generated and stored with users keys in "
        + keysFile.toUri().getPath());
  }

}
TOP

Related Classes of org.apache.hadoop.mapred.JobInProgress$FallowSlotInfo

TOP
Copyright © 2018 www.massapi.com. All rights reserved.
All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.
ga('create', 'UA-20639858-1', 'auto'); ga('send', 'pageview');