public void configure(JobConf conf) {
    /*
     * It reads all the configurations and distributed cache from outside.
     */

    // Read number of nodes in input layer and output layer from configuration
    inputNumdims = conf.get("numdims");
    inputNumhid = conf.get("numhid");

    // Read the weights from distributed cache
    Path[] pathwaysFiles = new Path[0];
    try {
      pathwaysFiles = DistributedCache.getLocalCacheFiles(conf);
      for (Path path : pathwaysFiles) {
        /*
         * this loop reads all the distributed cache files
         * In fact, the driver program ensures that there is only one distributed cache file
         */
        BufferedReader fis = new BufferedReader(new FileReader(path.toString()));
        weightline = fis.readLine();
      }
    } catch (Exception e) {
      e.printStackTrace();
    }
  }
  public void configure(JobConf conf) {
    numberOfCenters = Integer.valueOf(conf.get("numberOfCenters"));
    centersDirectory = conf.get("centersReadDirectory");

    try {
      Configuration c = new Configuration();
      FileSystem fs = FileSystem.get(c);

      for (int index = 0; index < numberOfCenters; ++index) {
        SequenceFile.Reader reader =
            new SequenceFile.Reader(fs, new Path(centersDirectory + "/centers/" + index), c);

        LongWritable key = new LongWritable();
        Point value = new Point();

        reader.next(key, value);

        Point center = (Point) value;

        centers.add(center);

        reader.close();
      }
    } catch (IOException e) {
      // do nothing
      // I hope this doesn't happen
      System.out.println("well, damn.");
      e.printStackTrace();
    }
  }
Beispiel #3
0
  /**
   * Driver to copy srcPath to destPath depending on required protocol.
   *
   * @param args arguments
   */
  static void copy(final Configuration conf, final Arguments args) throws IOException {
    LOG.info("srcPaths=" + args.srcs);
    LOG.info("destPath=" + args.dst);
    checkSrcPath(conf, args.srcs);

    JobConf job = createJobConf(conf);
    if (args.preservedAttributes != null) {
      job.set(PRESERVE_STATUS_LABEL, args.preservedAttributes);
    }
    if (args.mapredSslConf != null) {
      job.set("dfs.https.client.keystore.resource", args.mapredSslConf);
    }

    // Initialize the mapper
    try {
      setup(conf, job, args);
      JobClient.runJob(job);
      finalize(conf, job, args.dst, args.preservedAttributes);
    } finally {
      // delete tmp
      fullyDelete(job.get(TMP_DIR_LABEL), job);
      // delete jobDirectory
      fullyDelete(job.get(JOB_DIR_LABEL), job);
    }
  }
Beispiel #4
0
  // Mostly for setting up the symlinks. Note that when we setup the distributed
  // cache, we didn't create the symlinks. This is done on a per task basis
  // by the currently executing task.
  public static void setupWorkDir(JobConf conf) throws IOException {
    File workDir = new File(".").getAbsoluteFile();
    FileUtil.fullyDelete(workDir);
    if (DistributedCache.getSymlink(conf)) {
      URI[] archives = DistributedCache.getCacheArchives(conf);
      URI[] files = DistributedCache.getCacheFiles(conf);
      Path[] localArchives = DistributedCache.getLocalCacheArchives(conf);
      Path[] localFiles = DistributedCache.getLocalCacheFiles(conf);
      if (archives != null) {
        for (int i = 0; i < archives.length; i++) {
          String link = archives[i].getFragment();
          if (link != null) {
            link = workDir.toString() + Path.SEPARATOR + link;
            File flink = new File(link);
            if (!flink.exists()) {
              FileUtil.symLink(localArchives[i].toString(), link);
            }
          }
        }
      }
      if (files != null) {
        for (int i = 0; i < files.length; i++) {
          String link = files[i].getFragment();
          if (link != null) {
            link = workDir.toString() + Path.SEPARATOR + link;
            File flink = new File(link);
            if (!flink.exists()) {
              FileUtil.symLink(localFiles[i].toString(), link);
            }
          }
        }
      }
    }
    File jobCacheDir = null;
    if (conf.getJar() != null) {
      jobCacheDir = new File(new Path(conf.getJar()).getParent().toString());
    }

    // create symlinks for all the files in job cache dir in current
    // workingdir for streaming
    try {
      DistributedCache.createAllSymlink(conf, jobCacheDir, workDir);
    } catch (IOException ie) {
      // Do not exit even if symlinks have not been created.
      LOG.warn(StringUtils.stringifyException(ie));
    }
    // add java.io.tmpdir given by mapred.child.tmp
    String tmp = conf.get("mapred.child.tmp", "./tmp");
    Path tmpDir = new Path(tmp);

    // if temp directory path is not absolute
    // prepend it with workDir.
    if (!tmpDir.isAbsolute()) {
      tmpDir = new Path(workDir.toString(), tmp);
      FileSystem localFs = FileSystem.getLocal(conf);
      if (!localFs.mkdirs(tmpDir) && !localFs.getFileStatus(tmpDir).isDir()) {
        throw new IOException("Mkdirs failed to create " + tmpDir.toString());
      }
    }
  }
Beispiel #5
0
 public void configure(JobConf job) {
   this.jobConf = job;
   urlNormalizers = new URLNormalizers(job, URLNormalizers.SCOPE_INJECT);
   interval = jobConf.getInt("db.fetch.interval.default", 2592000);
   filters = new URLFilters(jobConf);
   scfilters = new ScoringFilters(jobConf);
   scoreInjected = jobConf.getFloat("db.score.injected", 1.0f);
   curTime = job.getLong("injector.current.time", System.currentTimeMillis());
 }
Beispiel #6
0
 /**
  * Calculate how many maps to run. Number of maps is bounded by a minimum of the cumulative size
  * of the copy / (distcp.bytes.per.map, default BYTES_PER_MAP or -m on the command line) and at
  * most (distcp.max.map.tasks, default MAX_MAPS_PER_NODE * nodes in the cluster).
  *
  * @param totalBytes Count of total bytes for job
  * @param job The job to configure
  * @return Count of maps to run.
  */
 private static void setMapCount(long totalBytes, JobConf job) throws IOException {
   int numMaps = (int) (totalBytes / job.getLong(BYTES_PER_MAP_LABEL, BYTES_PER_MAP));
   numMaps =
       Math.min(
           numMaps,
           job.getInt(
               MAX_MAPS_LABEL,
               MAX_MAPS_PER_NODE * new JobClient(job).getClusterStatus().getTaskTrackers()));
   job.setNumMapTasks(Math.max(numMaps, 1));
 }
Beispiel #7
0
  public void inject(Path crawlDb, Path urlDir) throws IOException {
    SimpleDateFormat sdf = new SimpleDateFormat("yyyy-MM-dd HH:mm:ss");
    long start = System.currentTimeMillis();
    if (LOG.isInfoEnabled()) {
      LOG.info("Injector: starting at " + sdf.format(start));
      LOG.info("Injector: crawlDb: " + crawlDb);
      LOG.info("Injector: urlDir: " + urlDir);
    }

    Path tempDir =
        new Path(
            getConf().get("mapred.temp.dir", ".")
                + "/inject-temp-"
                + Integer.toString(new Random().nextInt(Integer.MAX_VALUE)));

    // map text input file to a <url,CrawlDatum> file
    if (LOG.isInfoEnabled()) {
      LOG.info("Injector: Converting injected urls to crawl db entries.");
    }
    JobConf sortJob = new NutchJob(getConf());
    sortJob.setJobName("inject " + urlDir);
    FileInputFormat.addInputPath(sortJob, urlDir);
    sortJob.setMapperClass(InjectMapper.class);

    FileOutputFormat.setOutputPath(sortJob, tempDir);
    sortJob.setOutputFormat(SequenceFileOutputFormat.class);
    sortJob.setOutputKeyClass(Text.class);
    sortJob.setOutputValueClass(CrawlDatum.class);
    sortJob.setLong("injector.current.time", System.currentTimeMillis());
    RunningJob mapJob = JobClient.runJob(sortJob);

    long urlsInjected = mapJob.getCounters().findCounter("injector", "urls_injected").getValue();
    long urlsFiltered = mapJob.getCounters().findCounter("injector", "urls_filtered").getValue();
    LOG.info("Injector: total number of urls rejected by filters: " + urlsFiltered);
    LOG.info(
        "Injector: total number of urls injected after normalization and filtering: "
            + urlsInjected);

    // merge with existing crawl db
    if (LOG.isInfoEnabled()) {
      LOG.info("Injector: Merging injected urls into crawl db.");
    }
    JobConf mergeJob = CrawlDb.createJob(getConf(), crawlDb);
    FileInputFormat.addInputPath(mergeJob, tempDir);
    mergeJob.setReducerClass(InjectReducer.class);
    JobClient.runJob(mergeJob);
    CrawlDb.install(mergeJob, crawlDb);

    // clean up
    FileSystem fs = FileSystem.get(getConf());
    fs.delete(tempDir, true);

    long end = System.currentTimeMillis();
    LOG.info(
        "Injector: finished at "
            + sdf.format(end)
            + ", elapsed: "
            + TimingUtil.elapsedTime(start, end));
  }
Beispiel #8
0
    /**
     * Produce splits such that each is no greater than the quotient of the total size and the
     * number of splits requested.
     *
     * @param job The handle to the JobConf object
     * @param numSplits Number of splits requested
     */
    public InputSplit[] getSplits(JobConf job, int numSplits) throws IOException {
      int cnfiles = job.getInt(SRC_COUNT_LABEL, -1);
      long cbsize = job.getLong(TOTAL_SIZE_LABEL, -1);
      String srcfilelist = job.get(SRC_LIST_LABEL, "");
      if (cnfiles < 0 || cbsize < 0 || "".equals(srcfilelist)) {
        throw new RuntimeException(
            "Invalid metadata: #files("
                + cnfiles
                + ") total_size("
                + cbsize
                + ") listuri("
                + srcfilelist
                + ")");
      }
      Path src = new Path(srcfilelist);
      FileSystem fs = src.getFileSystem(job);
      FileStatus srcst = fs.getFileStatus(src);

      ArrayList<FileSplit> splits = new ArrayList<FileSplit>(numSplits);
      LongWritable key = new LongWritable();
      FilePair value = new FilePair();
      final long targetsize = cbsize / numSplits;
      long pos = 0L;
      long last = 0L;
      long acc = 0L;
      long cbrem = srcst.getLen();
      SequenceFile.Reader sl = null;
      try {
        sl = new SequenceFile.Reader(fs, src, job);
        for (; sl.next(key, value); last = sl.getPosition()) {
          // if adding this split would put this split past the target size,
          // cut the last split and put this next file in the next split.
          if (acc + key.get() > targetsize && acc != 0) {
            long splitsize = last - pos;
            splits.add(new FileSplit(src, pos, splitsize, (String[]) null));
            cbrem -= splitsize;
            pos = last;
            acc = 0L;
          }
          acc += key.get();
        }
      } finally {
        checkAndClose(sl);
      }
      if (cbrem != 0) {
        splits.add(new FileSplit(src, pos, cbrem, (String[]) null));
      }

      return splits.toArray(new FileSplit[splits.size()]);
    }
  public void testFailAbort() throws IOException {
    JobConf job = new JobConf();
    job.set(FileSystem.FS_DEFAULT_NAME_KEY, "faildel:///");
    job.setClass("fs.faildel.impl", FakeFileSystem.class, FileSystem.class);
    setConfForFileOutputCommitter(job);
    JobContext jContext = new JobContextImpl(job, taskID.getJobID());
    TaskAttemptContext tContext = new TaskAttemptContextImpl(job, taskID);
    FileOutputCommitter committer = new FileOutputCommitter();
    FileOutputFormat.setWorkOutputPath(job, committer.getTempTaskOutputPath(tContext));

    // do setup
    committer.setupJob(jContext);
    committer.setupTask(tContext);
    String file = "test.txt";

    // A reporter that does nothing
    Reporter reporter = Reporter.NULL;
    // write output
    FileSystem localFs = new FakeFileSystem();
    TextOutputFormat theOutputFormat = new TextOutputFormat();
    RecordWriter theRecordWriter = theOutputFormat.getRecordWriter(localFs, job, file, reporter);
    writeOutput(theRecordWriter, reporter);

    // do abort
    Throwable th = null;
    try {
      committer.abortTask(tContext);
    } catch (IOException ie) {
      th = ie;
    }
    assertNotNull(th);
    assertTrue(th instanceof IOException);
    assertTrue(th.getMessage().contains("fake delete failed"));
    File jobTmpDir = new File(new Path(outDir, FileOutputCommitter.TEMP_DIR_NAME).toString());
    File taskTmpDir = new File(jobTmpDir, "_" + taskID);
    File expectedFile = new File(taskTmpDir, file);
    assertTrue(expectedFile + " does not exists", expectedFile.exists());

    th = null;
    try {
      committer.abortJob(jContext, JobStatus.State.FAILED);
    } catch (IOException ie) {
      th = ie;
    }
    assertNotNull(th);
    assertTrue(th instanceof IOException);
    assertTrue(th.getMessage().contains("fake delete failed"));
    assertTrue("job temp dir does not exists", jobTmpDir.exists());
  }
Beispiel #10
0
  private static void finalize(
      Configuration conf, JobConf jobconf, final Path destPath, String presevedAttributes)
      throws IOException {
    if (presevedAttributes == null) {
      return;
    }
    EnumSet<FileAttribute> preseved = FileAttribute.parse(presevedAttributes);
    if (!preseved.contains(FileAttribute.USER)
        && !preseved.contains(FileAttribute.GROUP)
        && !preseved.contains(FileAttribute.PERMISSION)) {
      return;
    }

    FileSystem dstfs = destPath.getFileSystem(conf);
    Path dstdirlist = new Path(jobconf.get(DST_DIR_LIST_LABEL));
    SequenceFile.Reader in = null;
    try {
      in = new SequenceFile.Reader(dstdirlist.getFileSystem(jobconf), dstdirlist, jobconf);
      Text dsttext = new Text();
      FilePair pair = new FilePair();
      for (; in.next(dsttext, pair); ) {
        Path absdst = new Path(destPath, pair.output);
        updatePermissions(pair.input, dstfs.getFileStatus(absdst), preseved, dstfs);
      }
    } finally {
      checkAndClose(in);
    }
  }
Beispiel #11
0
  private static IntWritable deduceInputFile(JobConf job) {
    Path[] inputPaths = FileInputFormat.getInputPaths(job);
    Path inputFile = new Path(job.get("map.input.file"));

    // value == one for sort-input; value == two for sort-output
    return (inputFile.getParent().equals(inputPaths[0])) ? sortInput : sortOutput;
  }
Beispiel #12
0
  /** Start the JobTracker process, listen on the indicated port */
  JobTracker(Configuration conf) throws IOException {
    //
    // Grab some static constants
    //
    maxCurrentTasks = conf.getInt("mapred.tasktracker.tasks.maximum", 2);
    RETIRE_JOB_INTERVAL = conf.getLong("mapred.jobtracker.retirejob.interval", 24 * 60 * 60 * 1000);
    RETIRE_JOB_CHECK_INTERVAL = conf.getLong("mapred.jobtracker.retirejob.check", 60 * 1000);
    TASK_ALLOC_EPSILON = conf.getFloat("mapred.jobtracker.taskalloc.loadbalance.epsilon", 0.2f);
    PAD_FRACTION = conf.getFloat("mapred.jobtracker.taskalloc.capacitypad", 0.1f);
    MIN_SLOTS_FOR_PADDING = 3 * maxCurrentTasks;

    // This is a directory of temporary submission files.  We delete it
    // on startup, and can delete any files that we're done with
    this.conf = conf;
    JobConf jobConf = new JobConf(conf);
    this.systemDir = jobConf.getSystemDir();
    this.fs = FileSystem.get(conf);
    FileUtil.fullyDelete(fs, systemDir);
    fs.mkdirs(systemDir);

    // Same with 'localDir' except it's always on the local disk.
    jobConf.deleteLocalFiles(SUBDIR);

    // Set ports, start RPC servers, etc.
    InetSocketAddress addr = getAddress(conf);
    this.localMachine = addr.getHostName();
    this.port = addr.getPort();
    this.interTrackerServer = RPC.getServer(this, addr.getPort(), 10, false, conf);
    this.interTrackerServer.start();
    Properties p = System.getProperties();
    for (Iterator it = p.keySet().iterator(); it.hasNext(); ) {
      String key = (String) it.next();
      String val = (String) p.getProperty(key);
      LOG.info("Property '" + key + "' is " + val);
    }

    this.infoPort = conf.getInt("mapred.job.tracker.info.port", 50030);
    this.infoServer = new JobTrackerInfoServer(this, infoPort);
    this.infoServer.start();

    this.startTime = System.currentTimeMillis();

    new Thread(this.expireTrackers).start();
    new Thread(this.retireJobs).start();
    new Thread(this.initJobs).start();
  }
Beispiel #13
0
      public void configure(JobConf job) {
        // 'key' == sortInput for sort-input; key == sortOutput for sort-output
        key = deduceInputFile(job);

        if (key == sortOutput) {
          partitioner = new HashPartitioner<WritableComparable, Writable>();

          // Figure the 'current' partition and no. of reduces of the 'sort'
          try {
            URI inputURI = new URI(job.get("map.input.file"));
            String inputFile = inputURI.getPath();
            partition =
                Integer.valueOf(inputFile.substring(inputFile.lastIndexOf("part") + 5)).intValue();
            noSortReducers = job.getInt("sortvalidate.sort.reduce.tasks", -1);
          } catch (Exception e) {
            System.err.println("Caught: " + e);
            System.exit(-1);
          }
        }
      }
Beispiel #14
0
  // Job configuration
  private static JobConf createJobConf(Configuration conf) {
    JobConf jobconf = new JobConf(conf, DistCp.class);
    jobconf.setJobName(NAME);

    // turn off speculative execution, because DFS doesn't handle
    // multiple writers to the same file.
    jobconf.setMapSpeculativeExecution(false);

    jobconf.setInputFormat(CopyInputFormat.class);
    jobconf.setOutputKeyClass(Text.class);
    jobconf.setOutputValueClass(Text.class);

    jobconf.setMapperClass(CopyFilesMapper.class);
    jobconf.setNumReduceTasks(0);
    return jobconf;
  }
Beispiel #15
0
 @Override
 public void configure(JobConf job) {
   super.configure(job);
   this.localDir = job.get("mapred.child.tmp");
   try {
     FileAndWriter faw = openFileWriter(localDir, 1);
     s1File = faw.file;
     s1FileWriter = faw.writer;
     faw = openFileWriter(localDir, 2);
     s2File = faw.file;
     s2FileWriter = faw.writer;
   } catch (IOException e) {
     e.printStackTrace();
   }
 }
Beispiel #16
0
  public int run(String[] args) throws Exception {
    JobConf conf = new JobConf(getConf(), XiangLi1_exercise3.class);
    conf.setJobName("xiangli1_exercise3");
    conf.setNumReduceTasks(0);
    conf.setOutputKeyClass(Text.class);
    conf.setOutputValueClass(Text.class);

    conf.setMapperClass(Map.class);
    // conf.setCombinerClass(Reduce.class);
    // conf.setReducerClass(Reduce.class);

    conf.setInputFormat(TextInputFormat.class);
    conf.setOutputFormat(TextOutputFormat.class);

    FileInputFormat.setInputPaths(conf, new Path(args[0]));
    FileOutputFormat.setOutputPath(conf, new Path(args[1]));

    JobClient.runJob(conf);
    return 0;
  }
  @Override
  public int run(String[] args) throws Exception {
    Configuration conf = getConf();
    FileSystem fs = FileSystem.get(conf);
    JobConf job = new JobConf(conf);
    job.setJarByClass(Jacobi.class);

    fs.delete(new Path("curX"), true);
    job.setInputFormat(TextInputFormat.class);
    job.setOutputFormat(TextOutputFormat.class);
    job.setMapperClass(Map.class);
    job.setReducerClass(Reduce.class);
    job.setOutputKeyClass(IntWritable.class);
    job.setOutputValueClass(DoubleWritable.class);

    FileInputFormat.setInputPaths(job, new Path("preX"));
    FileOutputFormat.setOutputPath(job, new Path("curX"));

    JobClient.runJob(job);
    return 1;
  }
Beispiel #18
0
 /**
  * Mapper configuration. Extracts source and destination file system, as well as top-level paths
  * on source and destination directories. Gets the named file systems, to be used later in map.
  */
 public void configure(JobConf job) {
   destPath = new Path(job.get(DST_DIR_LABEL, "/"));
   try {
     destFileSys = destPath.getFileSystem(job);
   } catch (IOException ex) {
     throw new RuntimeException("Unable to get the named file system.", ex);
   }
   sizeBuf = job.getInt("copy.buf.size", 128 * 1024);
   buffer = new byte[sizeBuf];
   ignoreReadFailures = job.getBoolean(Options.IGNORE_READ_FAILURES.propertyname, false);
   preserve_status = job.getBoolean(Options.PRESERVE_STATUS.propertyname, false);
   if (preserve_status) {
     preseved = FileAttribute.parse(job.get(PRESERVE_STATUS_LABEL));
   }
   update = job.getBoolean(Options.UPDATE.propertyname, false);
   overwrite = !update && job.getBoolean(Options.OVERWRITE.propertyname, false);
   this.job = job;
 }
Beispiel #19
0
 /**
  * Map method. Copies one file from source file system to destination.
  *
  * @param key src len
  * @param value FilePair (FileStatus src, Path dst)
  * @param out Log of failed copies
  * @param reporter
  */
 public void map(
     LongWritable key,
     FilePair value,
     OutputCollector<WritableComparable<?>, Text> out,
     Reporter reporter)
     throws IOException {
   final FileStatus srcstat = value.input;
   final Path relativedst = new Path(value.output);
   try {
     copy(srcstat, relativedst, out, reporter);
   } catch (IOException e) {
     ++failcount;
     reporter.incrCounter(Counter.FAIL, 1);
     updateStatus(reporter);
     final String sfailure = "FAIL " + relativedst + " : " + StringUtils.stringifyException(e);
     out.collect(null, new Text(sfailure));
     LOG.info(sfailure);
     try {
       for (int i = 0; i < 3; ++i) {
         try {
           final Path tmp = new Path(job.get(TMP_DIR_LABEL), relativedst);
           if (destFileSys.delete(tmp, true)) break;
         } catch (Throwable ex) {
           // ignore, we are just cleaning up
           LOG.debug("Ignoring cleanup exception", ex);
         }
         // update status, so we don't get timed out
         updateStatus(reporter);
         Thread.sleep(3 * 1000);
       }
     } catch (InterruptedException inte) {
       throw (IOException) new IOException().initCause(inte);
     }
   } finally {
     updateStatus(reporter);
   }
 }
    public void configure(JobConf conf) {
      try {
        Path vInput;
        FileSystem fs;
        URI[] fvector;
        nsize = conf.getInt("DIMENTION", 0);
        sumVec = new double[nsize];
        resVec = new double[nsize];
        diaVec = new double[nsize];
        Arrays.fill(sumVec, 0);
        Arrays.fill(resVec, 0);
        Arrays.fill(diaVec, 0);

        fvector = DistributedCache.getCacheFiles(conf);
        vInput = new Path(fvector[0].getPath());
        fs = FileSystem.get(URI.create("hdfs://node17.cs.rochester.edu:9000"), conf);

        FSDataInputStream fdis = fs.open(vInput);

        String line;
        while ((line = fdis.readLine()) != null) {
          StringTokenizer tokenizer = new StringTokenizer(line);
          int rowIdx = Integer.parseInt(tokenizer.nextToken());
          int colIdx = Integer.parseInt(tokenizer.nextToken());
          double matVar = Double.parseDouble(tokenizer.nextToken());
          if (rowIdx == colIdx) {
            diaVec[rowIdx] = matVar;
          } else if (colIdx == nsize) {
            resVec[rowIdx] = matVar;
          } else {
            sumVec[rowIdx] += matVar;
          }
        }
      } catch (IOException e) {
        e.printStackTrace();
      }
    }
Beispiel #21
0
    static void checkRecords(
        Configuration defaults, int noMaps, int noReduces, Path sortInput, Path sortOutput)
        throws IOException {
      JobConf jobConf = new JobConf(defaults, RecordChecker.class);
      jobConf.setJobName("sortvalidate-record-checker");

      jobConf.setInputFormat(SequenceFileInputFormat.class);
      jobConf.setOutputFormat(SequenceFileOutputFormat.class);

      jobConf.setOutputKeyClass(BytesWritable.class);
      jobConf.setOutputValueClass(IntWritable.class);

      jobConf.setMapperClass(Map.class);
      jobConf.setReducerClass(Reduce.class);

      JobClient client = new JobClient(jobConf);
      ClusterStatus cluster = client.getClusterStatus();
      if (noMaps == -1) {
        noMaps = cluster.getTaskTrackers() * jobConf.getInt("test.sortvalidate.maps_per_host", 10);
      }
      if (noReduces == -1) {
        noReduces = (int) (cluster.getMaxReduceTasks() * 0.9);
        String sortReduces = jobConf.get("test.sortvalidate.reduces_per_host");
        if (sortReduces != null) {
          noReduces = cluster.getTaskTrackers() * Integer.parseInt(sortReduces);
        }
      }
      jobConf.setNumMapTasks(noMaps);
      jobConf.setNumReduceTasks(noReduces);

      FileInputFormat.setInputPaths(jobConf, sortInput);
      FileInputFormat.addInputPath(jobConf, sortOutput);
      Path outputPath = new Path("/tmp/sortvalidate/recordchecker");
      FileSystem fs = FileSystem.get(defaults);
      if (fs.exists(outputPath)) {
        fs.delete(outputPath, true);
      }
      FileOutputFormat.setOutputPath(jobConf, outputPath);

      // Uncomment to run locally in a single process
      // job_conf.set("mapred.job.tracker", "local");
      Path[] inputPaths = FileInputFormat.getInputPaths(jobConf);
      System.out.println(
          "\nSortValidator.RecordChecker: Running on "
              + cluster.getTaskTrackers()
              + " nodes to validate sort from "
              + inputPaths[0]
              + ", "
              + inputPaths[1]
              + " into "
              + FileOutputFormat.getOutputPath(jobConf)
              + " with "
              + noReduces
              + " reduces.");
      Date startTime = new Date();
      System.out.println("Job started: " + startTime);
      JobClient.runJob(jobConf);
      Date end_time = new Date();
      System.out.println("Job ended: " + end_time);
      System.out.println(
          "The job took " + (end_time.getTime() - startTime.getTime()) / 1000 + " seconds.");
    }
Beispiel #22
0
    /**
     * Copy a file to a destination.
     *
     * @param srcstat src path and metadata
     * @param dstpath dst path
     * @param reporter
     */
    private void copy(
        FileStatus srcstat,
        Path relativedst,
        OutputCollector<WritableComparable<?>, Text> outc,
        Reporter reporter)
        throws IOException {
      Path absdst = new Path(destPath, relativedst);
      int totfiles = job.getInt(SRC_COUNT_LABEL, -1);
      assert totfiles >= 0 : "Invalid file count " + totfiles;

      // if a directory, ensure created even if empty
      if (srcstat.isDir()) {
        if (destFileSys.exists(absdst)) {
          if (!destFileSys.getFileStatus(absdst).isDir()) {
            throw new IOException("Failed to mkdirs: " + absdst + " is a file.");
          }
        } else if (!destFileSys.mkdirs(absdst)) {
          throw new IOException("Failed to mkdirs " + absdst);
        }
        // TODO: when modification times can be set, directories should be
        // emitted to reducers so they might be preserved. Also, mkdirs does
        // not currently return an error when the directory already exists;
        // if this changes, all directory work might as well be done in reduce
        return;
      }

      if (destFileSys.exists(absdst) && !overwrite && !needsUpdate(srcstat, destFileSys, absdst)) {
        outc.collect(null, new Text("SKIP: " + srcstat.getPath()));
        ++skipcount;
        reporter.incrCounter(Counter.SKIP, 1);
        updateStatus(reporter);
        return;
      }

      Path tmpfile = new Path(job.get(TMP_DIR_LABEL), relativedst);
      long cbcopied = 0L;
      FSDataInputStream in = null;
      FSDataOutputStream out = null;
      try {
        // open src file
        try {
          in = srcstat.getPath().getFileSystem(job).open(srcstat.getPath());
        } catch (IOException e) {
          LOG.error("Failed to open src file " + srcstat.getPath() + ", ignore and return");
          in = null;
          return;
        }
        reporter.incrCounter(Counter.BYTESEXPECTED, srcstat.getLen());
        // open tmp file
        out = create(tmpfile, reporter, srcstat);
        // copy file
        for (int cbread; (cbread = in.read(buffer)) >= 0; ) {
          out.write(buffer, 0, cbread);
          cbcopied += cbread;
          reporter.setStatus(
              String.format("%.2f ", cbcopied * 100.0 / srcstat.getLen())
                  + absdst
                  + " [ "
                  + StringUtils.humanReadableInt(cbcopied)
                  + " / "
                  + StringUtils.humanReadableInt(srcstat.getLen())
                  + " ]");
        }
      } finally {
        checkAndClose(in);
        checkAndClose(out);
      }

      if (cbcopied != srcstat.getLen()) {
        if (srcstat.getLen() == 0 && cbcopied > 0) {
          LOG.info("most likely see a WAL file corruption: " + srcstat.getPath());
        } else {
          throw new IOException(
              "File size not matched: copied "
                  + bytesString(cbcopied)
                  + " to tmpfile (="
                  + tmpfile
                  + ") but expected "
                  + bytesString(srcstat.getLen())
                  + " from "
                  + srcstat.getPath());
        }
      } else {
        if (totfiles == 1) {
          // Copying a single file; use dst path provided by user as destination
          // rather than destination directory, if a file
          Path dstparent = absdst.getParent();
          if (!(destFileSys.exists(dstparent) && destFileSys.getFileStatus(dstparent).isDir())) {
            absdst = dstparent;
          }
        }
        if (destFileSys.exists(absdst) && destFileSys.getFileStatus(absdst).isDir()) {
          throw new IOException(absdst + " is a directory");
        }
        if (!destFileSys.mkdirs(absdst.getParent())) {
          throw new IOException("Failed to craete parent dir: " + absdst.getParent());
        }
        rename(tmpfile, absdst);

        FileStatus dststat = destFileSys.getFileStatus(absdst);
        if (dststat.getLen() != srcstat.getLen()) {
          destFileSys.delete(absdst, false);
          throw new IOException(
              "File size not matched: copied "
                  + bytesString(dststat.getLen())
                  + " to dst (="
                  + absdst
                  + ") but expected "
                  + bytesString(srcstat.getLen())
                  + " from "
                  + srcstat.getPath());
        }
        updatePermissions(srcstat, dststat);
      }

      // report at least once for each file
      ++copycount;
      reporter.incrCounter(Counter.BYTESCOPIED, cbcopied);
      reporter.incrCounter(Counter.COPY, 1);
      updateStatus(reporter);
    }
Beispiel #23
0
 public void configure(JobConf job) {
   interval = job.getInt("db.fetch.interval.default", 2592000);
   scoreInjected = job.getFloat("db.score.injected", 1.0f);
   overwrite = job.getBoolean("db.injector.overwrite", false);
   update = job.getBoolean("db.injector.update", false);
 }
 private void setConfForFileOutputCommitter(JobConf job) {
   job.set(JobContext.TASK_ATTEMPT_ID, attempt);
   job.setOutputCommitter(FileOutputCommitter.class);
   FileOutputFormat.setOutputPath(job, outDir);
 }
Beispiel #25
0
  @Override
  public final void run() {
    try {

      // before preparing the job localize
      // all the archives
      TaskAttemptID taskid = t.getTaskID();
      LocalDirAllocator lDirAlloc = new LocalDirAllocator("mapred.local.dir");
      File jobCacheDir = null;
      if (conf.getJar() != null) {
        jobCacheDir = new File(new Path(conf.getJar()).getParent().toString());
      }
      File workDir =
          new File(
              lDirAlloc
                  .getLocalPathToRead(
                      TaskTracker.getJobCacheSubdir()
                          + Path.SEPARATOR
                          + t.getJobID()
                          + Path.SEPARATOR
                          + t.getTaskID()
                          + Path.SEPARATOR
                          + MRConstants.WORKDIR,
                      conf)
                  .toString());

      URI[] archives = DistributedCache.getCacheArchives(conf);
      URI[] files = DistributedCache.getCacheFiles(conf);
      FileStatus fileStatus;
      FileSystem fileSystem;
      Path localPath;
      String baseDir;

      if ((archives != null) || (files != null)) {
        if (archives != null) {
          String[] archivesTimestamps = DistributedCache.getArchiveTimestamps(conf);
          Path[] p = new Path[archives.length];
          for (int i = 0; i < archives.length; i++) {
            fileSystem = FileSystem.get(archives[i], conf);
            fileStatus = fileSystem.getFileStatus(new Path(archives[i].getPath()));
            String cacheId = DistributedCache.makeRelative(archives[i], conf);
            String cachePath = TaskTracker.getCacheSubdir() + Path.SEPARATOR + cacheId;
            if (lDirAlloc.ifExists(cachePath, conf)) {
              localPath = lDirAlloc.getLocalPathToRead(cachePath, conf);
            } else {
              localPath = lDirAlloc.getLocalPathForWrite(cachePath, fileStatus.getLen(), conf);
            }
            baseDir = localPath.toString().replace(cacheId, "");
            p[i] =
                DistributedCache.getLocalCache(
                    archives[i],
                    conf,
                    new Path(baseDir),
                    fileStatus,
                    true,
                    Long.parseLong(archivesTimestamps[i]),
                    new Path(workDir.getAbsolutePath()),
                    false);
          }
          DistributedCache.setLocalArchives(conf, stringifyPathArray(p));
        }
        if ((files != null)) {
          String[] fileTimestamps = DistributedCache.getFileTimestamps(conf);
          Path[] p = new Path[files.length];
          for (int i = 0; i < files.length; i++) {
            fileSystem = FileSystem.get(files[i], conf);
            fileStatus = fileSystem.getFileStatus(new Path(files[i].getPath()));
            String cacheId = DistributedCache.makeRelative(files[i], conf);
            String cachePath = TaskTracker.getCacheSubdir() + Path.SEPARATOR + cacheId;
            if (lDirAlloc.ifExists(cachePath, conf)) {
              localPath = lDirAlloc.getLocalPathToRead(cachePath, conf);
            } else {
              localPath = lDirAlloc.getLocalPathForWrite(cachePath, fileStatus.getLen(), conf);
            }
            baseDir = localPath.toString().replace(cacheId, "");
            p[i] =
                DistributedCache.getLocalCache(
                    files[i],
                    conf,
                    new Path(baseDir),
                    fileStatus,
                    false,
                    Long.parseLong(fileTimestamps[i]),
                    new Path(workDir.getAbsolutePath()),
                    false);
          }
          DistributedCache.setLocalFiles(conf, stringifyPathArray(p));
        }
        Path localTaskFile = new Path(t.getJobFile());
        FileSystem localFs = FileSystem.getLocal(conf);
        localFs.delete(localTaskFile, true);
        OutputStream out = localFs.create(localTaskFile);
        try {
          conf.writeXml(out);
        } finally {
          out.close();
        }
      }

      if (!prepare()) {
        return;
      }

      String sep = System.getProperty("path.separator");
      StringBuffer classPath = new StringBuffer();
      // start with same classpath as parent process
      classPath.append(System.getProperty("java.class.path"));
      classPath.append(sep);
      if (!workDir.mkdirs()) {
        if (!workDir.isDirectory()) {
          LOG.fatal("Mkdirs failed to create " + workDir.toString());
        }
      }

      String jar = conf.getJar();
      if (jar != null) {
        // if jar exists, it into workDir
        File[] libs = new File(jobCacheDir, "lib").listFiles();
        if (libs != null) {
          for (int i = 0; i < libs.length; i++) {
            classPath.append(sep); // add libs from jar to classpath
            classPath.append(libs[i]);
          }
        }
        classPath.append(sep);
        classPath.append(new File(jobCacheDir, "classes"));
        classPath.append(sep);
        classPath.append(jobCacheDir);
      }

      // include the user specified classpath

      // archive paths
      Path[] archiveClasspaths = DistributedCache.getArchiveClassPaths(conf);
      if (archiveClasspaths != null && archives != null) {
        Path[] localArchives = DistributedCache.getLocalCacheArchives(conf);
        if (localArchives != null) {
          for (int i = 0; i < archives.length; i++) {
            for (int j = 0; j < archiveClasspaths.length; j++) {
              if (archives[i].getPath().equals(archiveClasspaths[j].toString())) {
                classPath.append(sep);
                classPath.append(localArchives[i].toString());
              }
            }
          }
        }
      }
      // file paths
      Path[] fileClasspaths = DistributedCache.getFileClassPaths(conf);
      if (fileClasspaths != null && files != null) {
        Path[] localFiles = DistributedCache.getLocalCacheFiles(conf);
        if (localFiles != null) {
          for (int i = 0; i < files.length; i++) {
            for (int j = 0; j < fileClasspaths.length; j++) {
              if (files[i].getPath().equals(fileClasspaths[j].toString())) {
                classPath.append(sep);
                classPath.append(localFiles[i].toString());
              }
            }
          }
        }
      }

      classPath.append(sep);
      classPath.append(workDir);
      //  Build exec child jmv args.
      Vector<String> vargs = new Vector<String>(8);
      File jvm = // use same jvm as parent
          new File(new File(System.getProperty("java.home"), "bin"), "java");

      vargs.add(jvm.toString());

      // Add child (task) java-vm options.
      //
      // The following symbols if present in mapred.child.java.opts value are
      // replaced:
      // + @taskid@ is interpolated with value of TaskID.
      // Other occurrences of @ will not be altered.
      //
      // Example with multiple arguments and substitutions, showing
      // jvm GC logging, and start of a passwordless JVM JMX agent so can
      // connect with jconsole and the likes to watch child memory, threads
      // and get thread dumps.
      //
      //  <property>
      //    <name>mapred.child.java.opts</name>
      //    <value>-verbose:gc -Xloggc:/tmp/@[email protected] \
      //           -Dcom.sun.management.jmxremote.authenticate=false \
      //           -Dcom.sun.management.jmxremote.ssl=false \
      //    </value>
      //  </property>
      //
      String javaOpts = conf.get("mapred.child.java.opts", "-Xmx200m");
      javaOpts = javaOpts.replace("@taskid@", taskid.toString());
      String[] javaOptsSplit = javaOpts.split(" ");

      // Add java.library.path; necessary for loading native libraries.
      //
      // 1. To support native-hadoop library i.e. libhadoop.so, we add the
      //    parent processes' java.library.path to the child.
      // 2. We also add the 'cwd' of the task to it's java.library.path to help
      //    users distribute native libraries via the DistributedCache.
      // 3. The user can also specify extra paths to be added to the
      //    java.library.path via mapred.child.java.opts.
      //
      String libraryPath = System.getProperty("java.library.path");
      if (libraryPath == null) {
        libraryPath = workDir.getAbsolutePath();
      } else {
        libraryPath += sep + workDir;
      }
      boolean hasUserLDPath = false;
      for (int i = 0; i < javaOptsSplit.length; i++) {
        if (javaOptsSplit[i].startsWith("-Djava.library.path=")) {
          javaOptsSplit[i] += sep + libraryPath;
          hasUserLDPath = true;
          break;
        }
      }
      if (!hasUserLDPath) {
        vargs.add("-Djava.library.path=" + libraryPath);
      }
      for (int i = 0; i < javaOptsSplit.length; i++) {
        vargs.add(javaOptsSplit[i]);
      }

      // add java.io.tmpdir given by mapred.child.tmp
      String tmp = conf.get("mapred.child.tmp", "./tmp");
      Path tmpDir = new Path(tmp);

      // if temp directory path is not absolute
      // prepend it with workDir.
      if (!tmpDir.isAbsolute()) {
        tmpDir = new Path(workDir.toString(), tmp);
      }
      FileSystem localFs = FileSystem.getLocal(conf);
      if (!localFs.mkdirs(tmpDir) && !localFs.getFileStatus(tmpDir).isDir()) {
        throw new IOException("Mkdirs failed to create " + tmpDir.toString());
      }
      vargs.add("-Djava.io.tmpdir=" + tmpDir.toString());

      // Add classpath.
      vargs.add("-classpath");
      vargs.add(classPath.toString());

      // Setup the log4j prop
      long logSize = TaskLog.getTaskLogLength(conf);
      vargs.add(
          "-Dhadoop.log.dir=" + new File(System.getProperty("hadoop.log.dir")).getAbsolutePath());
      vargs.add("-Dhadoop.root.logger=INFO,TLA");
      vargs.add("-Dhadoop.tasklog.taskid=" + taskid);
      vargs.add("-Dhadoop.tasklog.totalLogFileSize=" + logSize);

      if (conf.getProfileEnabled()) {
        if (conf.getProfileTaskRange(t.isMapTask()).isIncluded(t.getPartition())) {
          File prof = TaskLog.getTaskLogFile(taskid, TaskLog.LogName.PROFILE);
          vargs.add(String.format(conf.getProfileParams(), prof.toString()));
        }
      }

      // Add main class and its arguments
      vargs.add(Child.class.getName()); // main of Child
      // pass umbilical address
      InetSocketAddress address = tracker.getTaskTrackerReportAddress();
      vargs.add(address.getAddress().getHostAddress());
      vargs.add(Integer.toString(address.getPort()));
      vargs.add(taskid.toString()); // pass task identifier

      String pidFile = null;
      if (tracker.isTaskMemoryManagerEnabled()) {
        pidFile =
            lDirAlloc
                .getLocalPathForWrite(
                    (TaskTracker.getPidFilesSubdir() + Path.SEPARATOR + taskid), this.conf)
                .toString();
      }

      // set memory limit using ulimit if feasible and necessary ...
      String[] ulimitCmd = Shell.getUlimitMemoryCommand(conf);
      List<String> setup = null;
      if (ulimitCmd != null) {
        setup = new ArrayList<String>();
        for (String arg : ulimitCmd) {
          setup.add(arg);
        }
      }

      // Set up the redirection of the task's stdout and stderr streams
      File stdout = TaskLog.getTaskLogFile(taskid, TaskLog.LogName.STDOUT);
      File stderr = TaskLog.getTaskLogFile(taskid, TaskLog.LogName.STDERR);
      stdout.getParentFile().mkdirs();
      tracker.getTaskTrackerInstrumentation().reportTaskLaunch(taskid, stdout, stderr);

      Map<String, String> env = new HashMap<String, String>();
      StringBuffer ldLibraryPath = new StringBuffer();
      ldLibraryPath.append(workDir.toString());
      String oldLdLibraryPath = null;
      oldLdLibraryPath = System.getenv("LD_LIBRARY_PATH");
      if (oldLdLibraryPath != null) {
        ldLibraryPath.append(sep);
        ldLibraryPath.append(oldLdLibraryPath);
      }
      env.put("LD_LIBRARY_PATH", ldLibraryPath.toString());
      jvmManager.launchJvm(
          this,
          jvmManager.constructJvmEnv(
              setup, vargs, stdout, stderr, logSize, workDir, env, pidFile, conf));
      synchronized (lock) {
        while (!done) {
          lock.wait();
        }
      }
      tracker.getTaskTrackerInstrumentation().reportTaskEnd(t.getTaskID());
      if (exitCodeSet) {
        if (!killed && exitCode != 0) {
          if (exitCode == 65) {
            tracker.getTaskTrackerInstrumentation().taskFailedPing(t.getTaskID());
          }
          throw new IOException("Task process exit with nonzero status of " + exitCode + ".");
        }
      }
    } catch (FSError e) {
      LOG.fatal("FSError", e);
      try {
        tracker.fsError(t.getTaskID(), e.getMessage());
      } catch (IOException ie) {
        LOG.fatal(t.getTaskID() + " reporting FSError", ie);
      }
    } catch (Throwable throwable) {
      LOG.warn(t.getTaskID() + " Child Error", throwable);
      ByteArrayOutputStream baos = new ByteArrayOutputStream();
      throwable.printStackTrace(new PrintStream(baos));
      try {
        tracker.reportDiagnosticInfo(t.getTaskID(), baos.toString());
      } catch (IOException e) {
        LOG.warn(t.getTaskID() + " Reporting Diagnostics", e);
      }
    } finally {
      try {
        URI[] archives = DistributedCache.getCacheArchives(conf);
        URI[] files = DistributedCache.getCacheFiles(conf);
        if (archives != null) {
          for (int i = 0; i < archives.length; i++) {
            DistributedCache.releaseCache(archives[i], conf);
          }
        }
        if (files != null) {
          for (int i = 0; i < files.length; i++) {
            DistributedCache.releaseCache(files[i], conf);
          }
        }
      } catch (IOException ie) {
        LOG.warn("Error releasing caches : Cache files might not have been cleaned up");
      }
      tracker.reportTaskFinished(t.getTaskID(), false);
      if (t.isMapTask()) {
        tracker.addFreeMapSlot();
      } else {
        tracker.addFreeReduceSlot();
      }
    }
  }
Beispiel #26
0
    static void checkRecords(Configuration defaults, Path sortInput, Path sortOutput)
        throws IOException {
      FileSystem inputfs = sortInput.getFileSystem(defaults);
      FileSystem outputfs = sortOutput.getFileSystem(defaults);
      FileSystem defaultfs = FileSystem.get(defaults);
      JobConf jobConf = new JobConf(defaults, RecordStatsChecker.class);
      jobConf.setJobName("sortvalidate-recordstats-checker");

      int noSortReduceTasks = outputfs.listStatus(sortOutput, sortPathsFilter).length;
      jobConf.setInt("sortvalidate.sort.reduce.tasks", noSortReduceTasks);
      int noSortInputpaths = inputfs.listStatus(sortInput).length;

      jobConf.setInputFormat(NonSplitableSequenceFileInputFormat.class);
      jobConf.setOutputFormat(SequenceFileOutputFormat.class);

      jobConf.setOutputKeyClass(IntWritable.class);
      jobConf.setOutputValueClass(RecordStatsChecker.RecordStatsWritable.class);

      jobConf.setMapperClass(Map.class);
      jobConf.setCombinerClass(Reduce.class);
      jobConf.setReducerClass(Reduce.class);

      jobConf.setNumMapTasks(noSortReduceTasks);
      jobConf.setNumReduceTasks(1);

      FileInputFormat.setInputPaths(jobConf, sortInput);
      FileInputFormat.addInputPath(jobConf, sortOutput);
      Path outputPath = new Path("/tmp/sortvalidate/recordstatschecker");
      if (defaultfs.exists(outputPath)) {
        defaultfs.delete(outputPath, true);
      }
      FileOutputFormat.setOutputPath(jobConf, outputPath);

      // Uncomment to run locally in a single process
      // job_conf.set("mapred.job.tracker", "local");
      Path[] inputPaths = FileInputFormat.getInputPaths(jobConf);
      System.out.println(
          "\nSortValidator.RecordStatsChecker: Validate sort "
              + "from "
              + inputPaths[0]
              + " ("
              + noSortInputpaths
              + " files), "
              + inputPaths[1]
              + " ("
              + noSortReduceTasks
              + " files) into "
              + FileOutputFormat.getOutputPath(jobConf)
              + " with 1 reducer.");
      Date startTime = new Date();
      System.out.println("Job started: " + startTime);
      JobClient.runJob(jobConf);
      Date end_time = new Date();
      System.out.println("Job ended: " + end_time);
      System.out.println(
          "The job took " + (end_time.getTime() - startTime.getTime()) / 1000 + " seconds.");

      // Check to ensure that the statistics of the
      // framework's sort-input and sort-output match
      SequenceFile.Reader stats =
          new SequenceFile.Reader(defaultfs, new Path(outputPath, "part-00000"), defaults);
      IntWritable k1 = new IntWritable();
      IntWritable k2 = new IntWritable();
      RecordStatsWritable v1 = new RecordStatsWritable();
      RecordStatsWritable v2 = new RecordStatsWritable();
      if (!stats.next(k1, v1)) {
        throw new IOException("Failed to read record #1 from reduce's output");
      }
      if (!stats.next(k2, v2)) {
        throw new IOException("Failed to read record #2 from reduce's output");
      }

      if ((v1.getBytes() != v2.getBytes())
          || (v1.getRecords() != v2.getRecords())
          || v1.getChecksum() != v2.getChecksum()) {
        throw new IOException(
            "("
                + v1.getBytes()
                + ", "
                + v1.getRecords()
                + ", "
                + v1.getChecksum()
                + ") v/s ("
                + v2.getBytes()
                + ", "
                + v2.getRecords()
                + ", "
                + v2.getChecksum()
                + ")");
      }
    }
Beispiel #27
0
  /**
   * Initialize DFSCopyFileMapper specific job-configuration.
   *
   * @param conf : The dfs/mapred configuration.
   * @param jobConf : The handle to the jobConf object to be initialized.
   * @param args Arguments
   */
  private static void setup(Configuration conf, JobConf jobConf, final Arguments args)
      throws IOException {
    jobConf.set(DST_DIR_LABEL, args.dst.toUri().toString());

    // set boolean values
    final boolean update = args.flags.contains(Options.UPDATE);
    final boolean overwrite = !update && args.flags.contains(Options.OVERWRITE);
    jobConf.setBoolean(Options.UPDATE.propertyname, update);
    jobConf.setBoolean(Options.OVERWRITE.propertyname, overwrite);
    jobConf.setBoolean(
        Options.IGNORE_READ_FAILURES.propertyname,
        args.flags.contains(Options.IGNORE_READ_FAILURES));
    jobConf.setBoolean(
        Options.PRESERVE_STATUS.propertyname, args.flags.contains(Options.PRESERVE_STATUS));

    final String randomId = getRandomId();
    JobClient jClient = new JobClient(jobConf);
    Path jobDirectory = new Path(jClient.getSystemDir(), NAME + "_" + randomId);
    jobConf.set(JOB_DIR_LABEL, jobDirectory.toString());

    FileSystem dstfs = args.dst.getFileSystem(conf);
    boolean dstExists = dstfs.exists(args.dst);
    boolean dstIsDir = false;
    if (dstExists) {
      dstIsDir = dstfs.getFileStatus(args.dst).isDir();
    }

    // default logPath
    Path logPath = args.log;
    if (logPath == null) {
      String filename = "_distcp_logs_" + randomId;
      if (!dstExists || !dstIsDir) {
        Path parent = args.dst.getParent();
        if (!dstfs.exists(parent)) {
          dstfs.mkdirs(parent);
        }
        logPath = new Path(parent, filename);
      } else {
        logPath = new Path(args.dst, filename);
      }
    }
    FileOutputFormat.setOutputPath(jobConf, logPath);

    // create src list, dst list
    FileSystem jobfs = jobDirectory.getFileSystem(jobConf);

    Path srcfilelist = new Path(jobDirectory, "_distcp_src_files");
    jobConf.set(SRC_LIST_LABEL, srcfilelist.toString());
    SequenceFile.Writer src_writer =
        SequenceFile.createWriter(
            jobfs,
            jobConf,
            srcfilelist,
            LongWritable.class,
            FilePair.class,
            SequenceFile.CompressionType.NONE);

    Path dstfilelist = new Path(jobDirectory, "_distcp_dst_files");
    SequenceFile.Writer dst_writer =
        SequenceFile.createWriter(
            jobfs, jobConf, dstfilelist, Text.class, Text.class, SequenceFile.CompressionType.NONE);

    Path dstdirlist = new Path(jobDirectory, "_distcp_dst_dirs");
    jobConf.set(DST_DIR_LIST_LABEL, dstdirlist.toString());
    SequenceFile.Writer dir_writer =
        SequenceFile.createWriter(
            jobfs,
            jobConf,
            dstdirlist,
            Text.class,
            FilePair.class,
            SequenceFile.CompressionType.NONE);

    // handle the case where the destination directory doesn't exist
    // and we've only a single src directory OR we're updating/overwriting
    // the contents of the destination directory.
    final boolean special = (args.srcs.size() == 1 && !dstExists) || update || overwrite;
    int srcCount = 0, cnsyncf = 0, dirsyn = 0;
    long fileCount = 0L, byteCount = 0L, cbsyncs = 0L;
    try {
      for (Iterator<Path> srcItr = args.srcs.iterator(); srcItr.hasNext(); ) {
        final Path src = srcItr.next();
        FileSystem srcfs = src.getFileSystem(conf);
        FileStatus srcfilestat = srcfs.getFileStatus(src);
        Path root = special && srcfilestat.isDir() ? src : src.getParent();
        if (srcfilestat.isDir()) {
          ++srcCount;
        }

        Stack<FileStatus> pathstack = new Stack<FileStatus>();
        for (pathstack.push(srcfilestat); !pathstack.empty(); ) {
          FileStatus cur = pathstack.pop();
          FileStatus[] children = srcfs.listStatus(cur.getPath());
          for (int i = 0; i < children.length; i++) {
            boolean skipfile = false;
            final FileStatus child = children[i];
            final String dst = makeRelative(root, child.getPath());
            ++srcCount;

            if (child.isDir()) {
              pathstack.push(child);
            } else {
              // skip file if the src and the dst files are the same.
              skipfile = update && sameFile(srcfs, child, dstfs, new Path(args.dst, dst));
              // skip file if it exceed file limit or size limit
              skipfile |=
                  fileCount == args.filelimit || byteCount + child.getLen() > args.sizelimit;

              if (!skipfile) {
                ++fileCount;
                byteCount += child.getLen();

                if (LOG.isTraceEnabled()) {
                  LOG.trace("adding file " + child.getPath());
                }

                ++cnsyncf;
                cbsyncs += child.getLen();
                if (cnsyncf > SYNC_FILE_MAX || cbsyncs > BYTES_PER_MAP) {
                  src_writer.sync();
                  dst_writer.sync();
                  cnsyncf = 0;
                  cbsyncs = 0L;
                }
              }
            }

            if (!skipfile) {
              src_writer.append(
                  new LongWritable(child.isDir() ? 0 : child.getLen()), new FilePair(child, dst));
            }

            dst_writer.append(new Text(dst), new Text(child.getPath().toString()));
          }

          if (cur.isDir()) {
            String dst = makeRelative(root, cur.getPath());
            dir_writer.append(new Text(dst), new FilePair(cur, dst));
            if (++dirsyn > SYNC_FILE_MAX) {
              dirsyn = 0;
              dir_writer.sync();
            }
          }
        }
      }
    } finally {
      checkAndClose(src_writer);
      checkAndClose(dst_writer);
      checkAndClose(dir_writer);
    }

    FileStatus dststatus = null;
    try {
      dststatus = dstfs.getFileStatus(args.dst);
    } catch (FileNotFoundException fnfe) {
      LOG.info(args.dst + " does not exist.");
    }

    // create dest path dir if copying > 1 file
    if (dststatus == null) {
      if (srcCount > 1 && !dstfs.mkdirs(args.dst)) {
        throw new IOException("Failed to create" + args.dst);
      }
    }

    final Path sorted = new Path(jobDirectory, "_distcp_sorted");
    checkDuplication(jobfs, dstfilelist, sorted, conf);

    if (dststatus != null && args.flags.contains(Options.DELETE)) {
      deleteNonexisting(dstfs, dststatus, sorted, jobfs, jobDirectory, jobConf, conf);
    }

    Path tmpDir =
        new Path(
            (dstExists && !dstIsDir) || (!dstExists && srcCount == 1)
                ? args.dst.getParent()
                : args.dst,
            "_distcp_tmp_" + randomId);
    jobConf.set(TMP_DIR_LABEL, tmpDir.toUri().toString());
    LOG.info("srcCount=" + srcCount);
    jobConf.setInt(SRC_COUNT_LABEL, srcCount);
    jobConf.setLong(TOTAL_SIZE_LABEL, byteCount);
    setMapCount(byteCount, jobConf);
  }