Esempio n. 1
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());
      }
    }
  }
 public FSDir(File dir) throws IOException {
   this.dir = dir;
   this.children = null;
   if (!dir.exists()) {
     if (!dir.mkdirs()) {
       throw new IOException("Mkdirs failed to create " + dir.toString());
     }
   } else {
     File[] files = dir.listFiles();
     int numChildren = 0;
     for (int idx = 0; idx < files.length; idx++) {
       if (files[idx].isDirectory()) {
         numChildren++;
       } else if (Block.isBlockFilename(files[idx])) {
         numBlocks++;
       }
     }
     if (numChildren > 0) {
       children = new FSDir[numChildren];
       int curdir = 0;
       for (int idx = 0; idx < files.length; idx++) {
         if (files[idx].isDirectory()) {
           children[curdir] = new FSDir(files[idx]);
           curdir++;
         }
       }
     }
   }
 }
    FSVolume(File currentDir, Configuration conf) throws IOException {
      this.reserved = conf.getLong("dfs.datanode.du.reserved", 0);
      boolean supportAppends = conf.getBoolean("dfs.support.append", false);
      File parent = currentDir.getParentFile();

      this.detachDir = new File(parent, "detach");
      if (detachDir.exists()) {
        recoverDetachedBlocks(currentDir, detachDir);
      }

      // Files that were being written when the datanode was last shutdown
      // are now moved back to the data directory. It is possible that
      // in the future, we might want to do some sort of datanode-local
      // recovery for these blocks. For example, crc validation.
      //
      this.tmpDir = new File(parent, "tmp");
      if (tmpDir.exists()) {
        if (supportAppends) {
          recoverDetachedBlocks(currentDir, tmpDir);
        } else {
          FileUtil.fullyDelete(tmpDir);
        }
      }
      this.dataDir = new FSDir(currentDir);
      if (!tmpDir.mkdirs()) {
        if (!tmpDir.isDirectory()) {
          throw new IOException("Mkdirs failed to create " + tmpDir.toString());
        }
      }
      if (!detachDir.mkdirs()) {
        if (!detachDir.isDirectory()) {
          throw new IOException("Mkdirs failed to create " + detachDir.toString());
        }
      }
      this.usage = new DF(parent, conf);
      this.dfsUsage = new DU(parent, conf);
      this.dfsUsage.start();
    }
  /**
   * Method to move files from HDFS to local filesystem
   *
   * <p>localPath: Path on the machines filesystem fs:FileSystem object from HDFS pathList:List of
   * paths for files that might need to be backed up size:max size in bytes to be backed up
   *
   * <p>ReturnsDate of the last files backed up if reached size limit, else, zero
   */
  public long backupFiles(
      String localPath, String preservePath, FileSystem fs, ArrayList<Path> pathList, long size) {
    Path fsPath;
    long tmpSize = 0;
    long tmpDate = 0;

    // Start iterating over all paths
    for (Path hdfsPath : pathList) {
      try {
        long nFileSize = fs.getContentSummary(hdfsPath).getLength();
        tmpSize = tmpSize + nFileSize;

        if ((tmpSize <= size) || (size == 0)) {
          FileStatus stat = fs.getFileStatus(hdfsPath);

          System.err.println(
              "File "
                  + hdfsPath.toUri().getPath()
                  + " "
                  + nFileSize
                  + " bytes, "
                  + "perms: "
                  + stat.getOwner()
                  + "/"
                  + stat.getGroup()
                  + ", "
                  + stat.getPermission().toString());

          tmpDate = stat.getModificationTime() / 1000;

          String sFsPath = localPath + hdfsPath.toUri().getPath();
          fsPath = new Path(sFsPath);

          File f = new File(sFsPath);

          // COMMENTED OUT: until a few backup cycles run
          // and the mtime gets in fact set on all copied
          // files.
          //
          // ignore it if the file exists and has the same mtime
          // if (f.exists() && f.isFile() && f.lastModified() == stat.getModificationTime())
          // {
          // System.out.println("no need to backup " + f.toString() + ", mtime matches hdfs");
          // continue;
          // }

          if (false == m_bDryRun) {
            // check if we need to back up the local file
            // (not directory), if it already exists.
            if (f.exists() && f.isFile()) {
              // ignore files with substrings in the
              // no-preserve file
              if (true == doPreserveFile(sFsPath)) {
                // move it to the backup path
                String sNewPath = preservePath + hdfsPath.toUri().getPath();
                File newFile = new File(sNewPath);

                // create directory structure for new file?
                if (false == newFile.getParentFile().exists()) {
                  if (false == newFile.getParentFile().mkdirs()) {
                    System.err.println("Failed to mkdirs " + newFile.getParentFile().toString());
                    System.exit(1);
                  }
                }

                // rename existing file to new location
                if (false == f.renameTo(newFile)) {
                  System.err.println(
                      "Failed to renameTo " + f.toString() + " to " + newFile.toString());
                  System.exit(1);
                }

                System.out.println("preserved " + f.toString() + " into " + newFile.toString());
              } else {
                System.out.println("skipped preservation of " + f.toString());
              }
            }

            // copy from hdfs to local filesystem
            fs.copyToLocalFile(hdfsPath, fsPath);

            // set the mtime to match hdfs file
            f.setLastModified(stat.getModificationTime());

            // compare checksums on both files
            compareChecksums(fs, hdfsPath, sFsPath);
          }

          // don't print the progress after every file -- go
          // by at least 1% increments
          long nPercentDone = (long) (100 * tmpSize / m_nTotalBytes);
          if (nPercentDone > m_nLastPercentBytesDone) {
            System.out.println(
                "progress: copied "
                    + prettyPrintBytes(tmpSize)
                    + ", "
                    + nPercentDone
                    + "% done"
                    + ", tstamp="
                    + tmpDate);

            m_nLastPercentBytesDone = nPercentDone;
          }

          if (m_nSleepSeconds > 0) {
            try {
              Thread.sleep(1000 * m_nSleepSeconds);
            } catch (Exception e2) {
              // ignore
            }
          }
        } else {
          return tmpDate;
        }
      } catch (IOException e) {
        System.err.println("FATAL ERROR: Something wrong with the file");
        System.err.println(e);
        System.out.println(tmpDate);
        System.exit(1);

        return 0;
      }
    }

    return 0;
  }
Esempio n. 5
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();
      }
    }
  }