Question

我有一个在配置单元13（YARN）上正常运行的脚本我正在试验tez。当我在大型数据集上运行查询时，遇到以下错误。

0 FATAL [Socket Reader #1 for port 55739] org.apache.hadoop.yarn.YarnUncaughtExceptionHandler: Thread Thread[Socket Reader #1 for port 55739,5,main] threw an Error.  Shutting down now...
            java.lang.OutOfMemoryError: GC overhead limit exceeded
                at java.nio.ByteBuffer.allocate(ByteBuffer.java:331)
                at org.apache.hadoop.ipc.Server$Connection.readAndProcess(Server.java:1510)
                at org.apache.hadoop.ipc.Server$Listener.doRead(Server.java:750)
                at org.apache.hadoop.ipc.Server$Listener$Reader.doRunLoop(Server.java:624)
                at org.apache.hadoop.ipc.Server$Listener$Reader.run(Server.java:595)
            2015-12-07 20:31:32,859 FATAL [AsyncDispatcher event handler] org.apache.hadoop.yarn.event.AsyncDispatcher: Error in dispatcher thread
            java.lang.OutOfMemoryError: GC overhead limit exceeded
            2015-12-07 20:31:30,590 WARN [IPC Server handler 0 on 55739] org.apache.hadoop.ipc.Server: IPC Server handler 0 on 55739, call heartbeat({  containerId=container_1449516549171_0001_01_000100, requestId=10184, startIndex=0, maxEventsToGet=0, taskAttemptId=null, eventCount=0 }), rpc version=2, client version=19, methodsFingerPrint=557389974 from 10.10.30.35:47028 Call#11165 Retry#0: error: java.lang.OutOfMemoryError: GC overhead limit exceeded
            java.lang.OutOfMemoryError: GC overhead limit exceeded
                at javax.security.auth.SubjectDomainCombiner.optimize(SubjectDomainCombiner.java:464)
                at javax.security.auth.SubjectDomainCombiner.combine(SubjectDomainCombiner.java:267)
                at java.security.AccessControlContext.goCombiner(AccessControlContext.java:499)
                at java.security.AccessControlContext.optimize(AccessControlContext.java:407)
                at java.security.AccessController.getContext(AccessController.java:501)
                at javax.security.auth.Subject.doAs(Subject.java:412)
                at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1548)
                at org.apache.hadoop.ipc.Server$Handler.run(Server.java:2007)
            2015-12-07 20:32:53,495 INFO [Thread-60] amazon.emr.metrics.MetricsSaver: Saved 4:3 records to /mnt/var/em/raw/i-782f08c8_20151207_7921_07921_raw.bin
            2015-12-07 20:32:53,495 INFO [AsyncDispatcher event handler] org.apache.hadoop.yarn.event.AsyncDispatcher: Exiting, bbye..
            2015-12-07 20:32:50,435 INFO [IPC Server handler 20 on 55739] org.apache.hadoop.ipc.Server: IPC Server handler 20 on 55739, call getTask(org.apache.tez.common.ContainerContext@409a6aa9), rpc version=2, client version=19, methodsFingerPrint=557389974 from 10.10.30.33:33644 Call#11094 Retry#0: error: java.io.IOException: java.lang.OutOfMemoryError: GC overhead limit exceeded
            java.io.IOException: java.lang.OutOfMemoryError: GC overhead limit exceeded
            2015-12-07 20:32:29,117 WARN [IPC Server handler 23 on 55739] org.apache.hadoop.ipc.Server: IPC Server handler 23 on 55739, call getTask(org.apache.tez.common.ContainerContext@7c7e6992), rpc version=2, client version=19, methodsFingerPrint=557389974 from 10.10.30.38:44218 Call#11260 Retry#0: error: java.lang.OutOfMemoryError: GC overhead limit exceeded
            java.lang.OutOfMemoryError: GC overhead limit exceeded
            2015-12-07 20:32:53,497 INFO [Thread-60] amazon.emr.metrics.MetricsSaver: Saved 1:1 records to /mnt/var/em/raw/i-782f08c8_20151207_7921_07921_raw.bin
            2015-12-07 20:32:53,498 INFO [Thread-61] amazon.emr.metrics.MetricsSaver: Saved 1:1 records to /mnt/var/em/raw/i-782f08c8_20151207_7921_07921_raw.bin
            2015-12-07 20:32:53,498 INFO [Thread-2] org.apache.tez.dag.app.DAGAppMaster: DAGAppMaster received a signal. Signaling TaskScheduler
            2015-12-07 20:32:53,498 INFO [Thread-2] org.apache.tez.dag.app.rm.TaskSchedulerEventHandler: TaskScheduler notified that iSignalled was : true
            2015-12-07 20:32:53,499 INFO [Thread-2] org.apache.tez.dag.history.HistoryEventHandler: Stopping HistoryEventHandler
            2015-12-07 20:32:53,499 INFO [Thread-2] org.apache.tez.dag.history.recovery.RecoveryService: Stopping RecoveryService
            2015-12-07 20:32:53,499 INFO [Thread-2] org.apache.tez.dag.history.recovery.RecoveryService: Closing Summary Stream
            2015-12-07 20:32:53,499 INFO [LeaseRenewer:hadoop@10.10.30.148:9000] org.apache.hadoop.util.ExitUtil: Halt with status -1 Message: HaltException

有关EMR集群的一些规范 - m1.xlarge主节点，4个r3.8xlarge核心节点，2个r3.8xlarge任务节点（约1.3 TB内存）

我尝试了以下设置，但它们无法正常工作。

set tez.task.resource.memory.mb=8000;
SET hive.tez.container.size=30208;
SET hive.tez.java.opts=-Xmx24168m;

因为亚马逊在EMR上提供0.4.1版本的tez，我现在正在运行（也许这就是问题？）

任何人都可以请求帮助解决它。我试图调整一些与内存相关的属性，如mapreduce.map.memory.mb，但没有运气

Answer 1

尝试更改tez.task.resource.memory.mb

这是在Tez上下文中用于地图内存和减少内存的参数。

Answer 2

使用以下属性检查 yarn-site.xml ：

<configuration>
  <property>
   <name>yarn.nodemanager.vmem-check-enabled</name>
    <value>false</value>
    <description>Whether virtual memory limits will be enforced for containers</description>
  </property>
 <property>
   <name>yarn.nodemanager.vmem-pmem-ratio</name>
    <value>4</value>
    <description>Ratio between virtual memory to physical memory when setting memory limits for containers</description>
  </property>
 <property>
    <name>yarn.scheduler.minimum-allocation-mb</name>
    <value>1024</value>
  </property>
  <property>
    <name>yarn.scheduler.maximum-allocation-mb</name>
    <value>2048</value>
  </property>
  <property>
    <name>yarn.nodemanager.resource.memory-mb</name>
    <value>2048</value>
  </property>
</configuration>

Hz与Tez内存不足错误

2 个答案: