当前位置: 首页>>代码示例>>Java>>正文


Java ClusterStory类代码示例

本文整理汇总了Java中org.apache.hadoop.tools.rumen.ClusterStory的典型用法代码示例。如果您正苦于以下问题:Java ClusterStory类的具体用法?Java ClusterStory怎么用?Java ClusterStory使用的例子?那么, 这里精选的类代码示例或许可以为您提供帮助。


ClusterStory类属于org.apache.hadoop.tools.rumen包,在下文中一共展示了ClusterStory类的2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: startTaskTrackers

import org.apache.hadoop.tools.rumen.ClusterStory; //导入依赖的package包/类
/**
 * Start simulated task trackers based on topology.
 * @param clusterStory The cluster topology.
 * @param now
 *    time stamp when the simulator is started, {@link SimulatorTaskTracker}s
 *    are started shortly after this time stamp
 */
void startTaskTrackers(ClusterStory clusterStory, long now) {
  /** port assigned to TTs, incremented by 1 for each TT */
  int port = 10000;
  long ms = now + 100;

  for (MachineNode node : clusterStory.getMachines()) {
    String hostname = node.getName();
    RackNode rackNode = node.getRackNode();
    StaticMapping.addNodeToRack(hostname, rackNode.getName());
    String taskTrackerName = "tracker_" + hostname + ":localhost/127.0.0.1:"
        + port;
    port++;
    SimulatorTaskTracker tt = new SimulatorTaskTracker(jt, taskTrackerName,
        hostname, node.getMapSlots(), node.getReduceSlots());
    queue.addAll(tt.init(ms++));
  }
}
 
开发者ID:rhli,项目名称:hadoop-EAR,代码行数:25,代码来源:SimulatorEngine.java

示例2: startTaskTrackers

import org.apache.hadoop.tools.rumen.ClusterStory; //导入依赖的package包/类
/**
 * Start simulated task trackers based on topology.
 * @param clusterStory the cluster topology.
 * @param jobConf configuration object.
 * @param now
 *    time stamp when the simulator is started, {@link SimulatorTaskTracker}s
 *    are started uniformly randomly spread in [now,now+startDuration).
 * @return time stamp by which the entire cluster is booted up and all task
 *    trackers are sending hearbeats in their steady rate.
 */
long startTaskTrackers(ClusterStory cluster, JobConf jobConf, long now) {
  /** port assigned to TTs, incremented by 1 for each TT */
  int port = 10000;
  int numTaskTrackers = 0;

  Random random = new Random(RandomSeedGenerator.getSeed(
     "forStartTaskTrackers()", masterRandomSeed));

  final int startDuration = jobConf.getInt("mumak.cluster.startup.duration",
      DEFAULT_CLUSTER_STARTUP_DURATION);
  
  for (MachineNode node : cluster.getMachines()) {
    jobConf.set("mumak.tasktracker.host.name", node.getName());
    jobConf.set("mumak.tasktracker.tracker.name",
        "tracker_" + node.getName() + ":localhost/127.0.0.1:" + port);
    long subRandomSeed = RandomSeedGenerator.getSeed(
       "forTaskTracker" + numTaskTrackers, masterRandomSeed);
    jobConf.setLong("mumak.tasktracker.random.seed", subRandomSeed);
    numTaskTrackers++;
    port++;
    SimulatorTaskTracker tt = new SimulatorTaskTracker(jt, jobConf);
    long firstHeartbeat = now + random.nextInt(startDuration);
    queue.addAll(tt.init(firstHeartbeat));
  }
  
  // In startDuration + heartbeat interval of the full cluster time each 
  // TT is started up and told on its 2nd heartbeat to beat at a rate 
  // corresponding to the steady state of the cluster    
  long clusterSteady = now + startDuration + jt.getNextHeartbeatInterval();
  return clusterSteady;
}
 
开发者ID:rekhajoshm,项目名称:mapreduce-fork,代码行数:42,代码来源:SimulatorEngine.java


注:本文中的org.apache.hadoop.tools.rumen.ClusterStory类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。