当前位置: 首页>>代码示例>>Java>>正文


Java TaskAttemptInfo.getFinishTime方法代码示例

本文整理汇总了Java中org.apache.hadoop.mapreduce.jobhistory.JobHistoryParser.TaskAttemptInfo.getFinishTime方法的典型用法代码示例。如果您正苦于以下问题:Java TaskAttemptInfo.getFinishTime方法的具体用法?Java TaskAttemptInfo.getFinishTime怎么用?Java TaskAttemptInfo.getFinishTime使用的例子?那么, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在org.apache.hadoop.mapreduce.jobhistory.JobHistoryParser.TaskAttemptInfo的用法示例。


在下文中一共展示了TaskAttemptInfo.getFinishTime方法的3个代码示例,这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞,您的评价将有助于系统推荐出更棒的Java代码示例。

示例1: getMaxReduceTime

import org.apache.hadoop.mapreduce.jobhistory.JobHistoryParser.TaskAttemptInfo; //导入方法依赖的package包/类
/**
 * *  
 * @return This method returns the maximum time taken by reduce.
 */
@SuppressWarnings("deprecation")
private long getMaxReduceTime(Map<TaskAttemptID, TaskAttemptInfo> tasks, long referencedZeroTime) {
long maximumReduceTime = 0l;
for (Map.Entry<TaskAttemptID, TaskAttemptInfo> task : tasks
		.entrySet()) {
	if(!task.getKey().isMap()){
		TaskAttemptInfo taskAttemptInfo = (TaskAttemptInfo) (task.getValue());
		long reduceEnd = (taskAttemptInfo.getFinishTime()-referencedZeroTime)/CONVERSION_FACTOR_MILLISECS_TO_SECS;
		maximumReduceTime = reduceEnd > maximumReduceTime ? reduceEnd : maximumReduceTime;
	}}
return maximumReduceTime;
}
 
开发者ID:Impetus,项目名称:jumbune,代码行数:17,代码来源:YarnJobStatsUtility.java

示例2: compare

import org.apache.hadoop.mapreduce.jobhistory.JobHistoryParser.TaskAttemptInfo; //导入方法依赖的package包/类
@Override
public int compare(TaskAttemptInfo a, TaskAttemptInfo b) {
  long diff = a.getFinishTime() - b.getFinishTime();
  return diff == 0 ? 0 : (diff < 0 ? -1 : 1);
}
 
开发者ID:naver,项目名称:hadoop,代码行数:6,代码来源:TaskImpl.java

示例3: validateTaskAttemptLevelKeyValuesFormat

import org.apache.hadoop.mapreduce.jobhistory.JobHistoryParser.TaskAttemptInfo; //导入方法依赖的package包/类
private static void validateTaskAttemptLevelKeyValuesFormat(JobInfo job) {
  Map<TaskID, TaskInfo> tasks = job.getAllTasks();

  // For each task
  for (TaskInfo task : tasks.values()) {
    // validate info of each attempt
    for (TaskAttemptInfo attempt : task.getAllTaskAttempts().values()) {

      TaskAttemptID id = attempt.getAttemptId();
      assertNotNull(id);
      
      long startTime = attempt.getStartTime();
      assertTrue("Invalid Start time", startTime > 0);

      long finishTime = attempt.getFinishTime();
      assertTrue("Task FINISH_TIME is < START_TIME in history file",
          startTime < finishTime);

      // Make sure that the Task type exists and it is valid
      TaskType type = attempt.getTaskType();
      assertTrue("Unknown Task type \"" + type + "\" is seen in " +
                 "history file for task attempt " + id,
                 (type.equals(TaskType.MAP) || type.equals(TaskType.REDUCE) ||
                  type.equals(TaskType.JOB_CLEANUP) || 
                  type.equals(TaskType.JOB_SETUP)));

      // Validate task status
      String status = attempt.getTaskStatus();
      assertTrue("Unexpected TASK_STATUS \"" + status + "\" is seen in" +
                 " history file for task attempt " + id,
                 (status.equals(TaskStatus.State.SUCCEEDED.toString()) ||
                  status.equals(TaskStatus.State.FAILED.toString()) ||
                  status.equals(TaskStatus.State.KILLED.toString())));

      // Successful Reduce Task Attempts should have valid SHUFFLE_FINISHED
      // time and SORT_FINISHED time
      if (type.equals(TaskType.REDUCE) && 
          status.equals(TaskStatus.State.SUCCEEDED.toString())) {
        long shuffleFinishTime = attempt.getShuffleFinishTime();
        assertTrue(startTime < shuffleFinishTime);
        
        long sortFinishTime = attempt.getSortFinishTime();
        assertTrue(shuffleFinishTime < sortFinishTime);
      }
      else if (type.equals(TaskType.MAP) && 
          status.equals(TaskStatus.State.SUCCEEDED.toString())) {
        // Successful MAP Task Attempts should have valid MAP_FINISHED time
       long mapFinishTime = attempt.getMapFinishTime();
       assertTrue(startTime < mapFinishTime);
      }

      // check if hostname is valid
      String hostname = attempt.getHostname();
      Matcher m = hostNamePattern.matcher(hostname);
      assertTrue("Unexpected Host name of task attempt " + id, m.matches());

      // check if trackername is valid
      String trackerName = attempt.getTrackerName();
      m = trackerNamePattern.matcher(trackerName);
      assertTrue("Unexpected tracker name of task attempt " + id,
                 m.matches());

      if (!status.equals("KILLED")) {
        // check if http port is valid
        int httpPort = attempt.getHttpPort();
        assertTrue(httpPort > 0);
      }
      
      // check if counters are parsable
      Counters counters = attempt.getCounters();
      assertNotNull(counters);
    }
  }
}
 
开发者ID:rekhajoshm,项目名称:mapreduce-fork,代码行数:75,代码来源:TestJobHistory.java


注:本文中的org.apache.hadoop.mapreduce.jobhistory.JobHistoryParser.TaskAttemptInfo.getFinishTime方法示例由纯净天空整理自Github/MSDocs等开源代码及文档管理平台,相关代码片段筛选自各路编程大神贡献的开源项目,源码版权归原作者所有,传播和使用请参考对应项目的License;未经允许,请勿转载。