mapreduce job任务非常慢

最新推荐文章于 2024-06-09 22:28:28 发布

谷雨hadoop

最新推荐文章于 2024-06-09 22:28:28 发布

阅读量8k

点赞数

分类专栏： hadoop 文章标签： mapreduce

本文链接：https://blog.youkuaiyun.com/linux_ja/article/details/52187154

版权

hadoop 专栏收录该内容

19 篇文章

订阅专栏

1、application日志

2016-08-11 14:48:15,174 INFO [RMCommunicator Allocator] org.apache.hadoop.mapreduce.v2.app.rm.RMContainerAllocator: Ramping down all scheduled reduces:0
2016-08-11 14:48:15,174 INFO [RMCommunicator Allocator] org.apache.hadoop.mapreduce.v2.app.rm.RMContainerAllocator: Going to preempt 0
2016-08-11 14:48:15,174 INFO [RMCommunicator Allocator] org.apache.hadoop.mapreduce.v2.app.rm.RMContainerAllocator: Recalculating schedule, headroom=0
2016-08-11 14:48:15,174 INFO [RMCommunicator Allocator] org.apache.hadoop.mapreduce.v2.app.rm.RMContainerAllocator: Reduce slow start threshold not met. completedMapsForReduceSlowstart 1
<span style="color:#ff6666;">2016-08-11 14:48:16,176 INFO [RMCommunicator Allocator] org.apache.hadoop.mapreduce.v2.app.rm.RMContainerAllocator: Ramping down all scheduled reduces:0
</span>2016-08-11 14:48:16,177 INFO [RMCommunicator Allocator] org.apache.hadoop.mapreduce.v2.app.rm.RMContainerAllocator: Going to preempt 0
2016-08-11 14:48:16,177 INFO [RMCommunicator Allocator] org.apache.hadoop.mapreduce.v2.app.rm.RMContainerAllocator: Recalculating schedule, headroom=0
2016-08-11 14:48:16,177 INFO [RMCommunicator Allocator] org.apache.hadoop.mapreduce.v2.app.rm.RMContainerAllocator: Reduce slow start threshold not met. completedMapsForReduceSlowstart 1
2016-08-11 14:48:17,179 INFO [RMCommunicator Allocator] org.apache.hadoop.mapreduce.v2.app.rm.RMContainerAllocator: Ramping down all scheduled reduces:0
2016-08-11 14:48:17,179 INFO [RMCommunicator Allocator] org.apache.hadoop.mapreduce.v2.app.rm.RMContainerAllocator: Going to preempt 0
2016-08-11 14:48:17,179 INFO [RMCommunicator Allocator] org.apache.hadoop.mapreduce.v2.app.rm.RMContainerAllocator: Recalculating schedule, headroom=0
2016-08-11 14:48:17,179 INFO [RMCommunicator Allocator] org.apache.hadoop.mapreduce.v2.app.rm.RMContainerAllocator: Reduce slow start threshold not met. completedMapsForReduceSlowstart 1

程序中大量的上面日志出现，跟踪RMContainerAllocator代码：

 // availableMemForMap must be sufficient to run at least 1 map
      if (ResourceCalculatorUtils.computeAvailableContainers(availableResourceForMap,
        mapResourceRequest, getSchedulerResourceTypes()) <= 0) {
        // to make sure new containers are given to maps and not reduces
        // ramp down all scheduled reduces if any
        // (since reduces are scheduled at higher priority than maps)
        LOG.info("Ramping down all scheduled reduces:"
            + scheduledRequests.reduces.size());
        for (ContainerRequest req : scheduledRequests.reduces.values()) {
          pendingReduces.add(req);
        }
        scheduledRequests.reduces.clear();

computeAvailableContainers：

<pre name="code" class="html">  public static int computeAvailableContainers(Resource available,
                                                  Resource required, EnumSet<SchedulerResourceTypes> resourceTypes) {
    if (resourceTypes.contains(SchedulerResourceTypes.CPU)) {
      return Math.min(available.getMemory() / required.getMemory(),
              available.getVirtualCores() / required.getVirtualCores());
    }
    return available.getMemory() / required.getMemory();
  }

  // get available resources for this job
    Resource headRoom = getAvailableResources();
    if (headRoom == null) {
      headRoom = Resources.none();
    }
    LOG.info("Recalculating schedule, headroom=" + headRoom);
    //check for slow start
    if (!getIsReduceStarted()) {//not set yet
      int completedMapsForReduceSlowstart = (int)Math.ceil(reduceSlowStart * 
                      totalMaps);
      if(completedMaps < completedMapsForReduceSlowstart) {
        LOG.info("Reduce slow start threshold not met. " +
              "completedMapsForReduceSlowstart " + 
            completedMapsForReduceSlowstart);
        return;
      } else {
        LOG.info("Reduce slow start threshold reached. Scheduling reduces.");
        setIsReduceStarted(true);
      }
    }
    
    //if all maps are assigned, then ramp up all reduces irrespective of the
    //headroom
    if (scheduledMaps == 0 && numPendingReduces > 0) {
      LOG.info("All maps assigned. " +
          "Ramping up all remaining reduces:" + numPendingReduces);
      scheduleAllReduces();
      return;
    }

从上面两段代码说明资源以及不够用了

2、从nm日志看,请求的物理内存是25G，太多了，导致其他job获取不到资源，

2016-08-11 14:48:42,560 INFO org.apache.hadoop.yarn.server.nodemanager.containermanager.monitor.ContainersMonitorImpl: Memory usage of ProcessTree 30460 for container-id container_1470814634002_0543_01_000008: 410.9 MB of 25 GB physical memory used; 1.7 GB of 125 GB virtual memory used
2016-08-11 14:48:45,609 INFO org.apache.hadoop.yarn.server.nodemanager.containermanager.monitor.ContainersMonitorImpl: Memory usage of ProcessTree 30460 for container-id container_1470814634002_0543_01_000008: 410.8 MB of 25 GB physical memory used; 1.7 GB of 125 GB virtual memory used
2016-08-11 14:48:48,659 INFO org.apache.hadoop.yarn.server.nodemanager.containermanager.monitor.ContainersMonitorImpl: Memory usage of ProcessTree 30460 for container-id container_1470814634002_0543_01_000008: 410.8 MB of 25 GB physical memory used; 1.7 GB of 125 GB virtual memory used
2016-08-11 14:48:51,708 INFO org.apache.hadoop.yarn.server.nodemanager.containermanager.monitor.ContainersMonitorImpl: Memory usage of ProcessTree 30460 for container-id container_1470814634002_0543_01_000008: 410.8 MB of 25 GB physical memory used; 1.7 GB of 125 GB virtual memory used
2016-08-11 14:48:54,756 INFO org.apache.hadoop.yarn.server.nodemanager.containermanager.monitor.ContainersMonitorImpl: Memory usage of ProcessTree 30460 for container-id container_1470814634002_0543_01_000008: 410.8 MB of 25 GB physical memory used; 1.7 GB of 125 GB virtual memory used
2016-08-11 14:48:57,804 INFO org.apache.hadoop.yarn.server.nodemanager.containermanager.monitor.ContainersMonitorImpl: Memory usage of ProcessTree 30460 for container-id container_1470814634002_0543_01_000008: 410.8 MB of 25 GB physical memory used; 1.7 GB of 125 GB virtual memory used
2016-08-11 14:49:00,857 INFO org.apache.hadoop.yarn.server.nodemanager.containermanager.monitor.ContainersMonitorImpl: Memory usage of ProcessTree 30460 for container-id container_1470814634002_0543_01_000008: 410.8 MB of 25 GB physical memory used; 1.7 GB of 125 GB virtual memory used
2016-08-11 14:49:03,904 INFO org.apache.hadoo

3、于是修改配置

<property>
<name>yarn.scheduler.minimum-allocation-mb</name>
<value>2048</value>
</property>

问题解决