Example #1
0
  @Override
  /** start a new map-reduce job to do the merge, almost the same as ExecDriver. */
  public int execute(DriverContext driverContext) {
    HiveConf.setVar(job, HiveConf.ConfVars.HIVEINPUTFORMAT, CombineHiveInputFormat.class.getName());
    success = true;
    ShimLoader.getHadoopShims().setNullOutputFormat(job);
    job.setMapperClass(work.getMapperClass());

    Context ctx = driverContext.getCtx();
    boolean ctxCreated = false;
    try {
      if (ctx == null) {
        ctx = new Context(job);
        ctxCreated = true;
      }
    } catch (IOException e) {
      e.printStackTrace();
      console.printError(
          "Error launching map-reduce job",
          "\n" + org.apache.hadoop.util.StringUtils.stringifyException(e));
      return 5;
    }

    job.setMapOutputKeyClass(NullWritable.class);
    job.setMapOutputValueClass(NullWritable.class);
    if (work.getNumMapTasks() != null) {
      job.setNumMapTasks(work.getNumMapTasks());
    }

    // zero reducers
    job.setNumReduceTasks(0);

    if (work.getMinSplitSize() != null) {
      HiveConf.setLongVar(
          job, HiveConf.ConfVars.MAPREDMINSPLITSIZE, work.getMinSplitSize().longValue());
    }

    if (work.getInputformat() != null) {
      HiveConf.setVar(job, HiveConf.ConfVars.HIVEINPUTFORMAT, work.getInputformat());
    }

    String inpFormat = HiveConf.getVar(job, HiveConf.ConfVars.HIVEINPUTFORMAT);
    if ((inpFormat == null) || (!StringUtils.isNotBlank(inpFormat))) {
      inpFormat = ShimLoader.getHadoopShims().getInputFormatClassName();
    }

    LOG.info("Using " + inpFormat);

    try {
      job.setInputFormat((Class<? extends InputFormat>) (Class.forName(inpFormat)));
    } catch (ClassNotFoundException e) {
      throw new RuntimeException(e.getMessage());
    }

    String outputPath = this.work.getOutputDir();
    Path tempOutPath = Utilities.toTempPath(new Path(outputPath));
    try {
      FileSystem fs = tempOutPath.getFileSystem(job);
      if (!fs.exists(tempOutPath)) {
        fs.mkdirs(tempOutPath);
      }
    } catch (IOException e) {
      console.printError("Can't make path " + outputPath + " : " + e.getMessage());
      return 6;
    }

    RCFileBlockMergeOutputFormat.setMergeOutputPath(job, new Path(outputPath));

    job.setOutputKeyClass(NullWritable.class);
    job.setOutputValueClass(NullWritable.class);

    HiveConf.setBoolVar(
        job,
        HiveConf.ConfVars.HIVEMERGECURRENTJOBHASDYNAMICPARTITIONS,
        work.hasDynamicPartitions());

    int returnVal = 0;
    RunningJob rj = null;
    boolean noName = StringUtils.isEmpty(HiveConf.getVar(job, HiveConf.ConfVars.HADOOPJOBNAME));

    String jobName = null;
    if (noName && this.getQueryPlan() != null) {
      int maxlen = conf.getIntVar(HiveConf.ConfVars.HIVEJOBNAMELENGTH);
      jobName = Utilities.abbreviate(this.getQueryPlan().getQueryStr(), maxlen - 6);
    }

    if (noName) {
      // This is for a special case to ensure unit tests pass
      HiveConf.setVar(
          job,
          HiveConf.ConfVars.HADOOPJOBNAME,
          jobName != null ? jobName : "JOB" + Utilities.randGen.nextInt());
    }

    try {
      addInputPaths(job, work);

      Utilities.setMapRedWork(job, work, ctx.getMRTmpFileURI());

      // remove the pwd from conf file so that job tracker doesn't show this
      // logs
      String pwd = HiveConf.getVar(job, HiveConf.ConfVars.METASTOREPWD);
      if (pwd != null) {
        HiveConf.setVar(job, HiveConf.ConfVars.METASTOREPWD, "HIVE");
      }
      JobClient jc = new JobClient(job);

      String addedJars = Utilities.getResourceFiles(job, SessionState.ResourceType.JAR);
      if (!addedJars.isEmpty()) {
        job.set("tmpjars", addedJars);
      }

      // make this client wait if job trcker is not behaving well.
      Throttle.checkJobTracker(job, LOG);

      // Finally SUBMIT the JOB!
      rj = jc.submitJob(job);

      returnVal = jobExecHelper.progress(rj, jc);
      success = (returnVal == 0);

    } catch (Exception e) {
      e.printStackTrace();
      String mesg = " with exception '" + Utilities.getNameMessage(e) + "'";
      if (rj != null) {
        mesg = "Ended Job = " + rj.getJobID() + mesg;
      } else {
        mesg = "Job Submission failed" + mesg;
      }

      // Has to use full name to make sure it does not conflict with
      // org.apache.commons.lang.StringUtils
      console.printError(mesg, "\n" + org.apache.hadoop.util.StringUtils.stringifyException(e));

      success = false;
      returnVal = 1;
    } finally {
      try {
        if (ctxCreated) {
          ctx.clear();
        }
        if (rj != null) {
          if (returnVal != 0) {
            rj.killJob();
          }
          HadoopJobExecHelper.runningJobKillURIs.remove(rj.getJobID());
          jobID = rj.getID().toString();
        }
        RCFileMergeMapper.jobClose(outputPath, success, job, console);
      } catch (Exception e) {
      }
    }

    return (returnVal);
  }