Hadoop Map/Reduce内存限制

如何设置Hadoop  Map/Reduce任务的内存限制?

Parameter   Type   Meaning  
mapred.cluster.map.memory.mb   set by admin, cluster-wide   Cluster definition of memory per map slot. The maximum amount of memory, in MB, each map task on a tasktracker can consume.  
mapred.cluster.reduce.memory.mb   set by admin, cluster-wide   Cluster definition of memory per reduce slot. The maximum amount of memory, in MB, each reduce task on a tasktracker can consume.  
mapred.job.map.memory.mb   set by user, per-job   Job requirement for map tasks. The maximum amount of memory each map task of a job can consume, in MB.  
mapred.job.reduce.memory.mb   set by user, per-job   job requirement for reduce tasks. The maximum amount of memory each reduce task of a job can consume, in MB.  
mapred.cluster.max.map.memory.mb   set by admin, cluster-wide   Max limit on jobs. The maximum value that can be specified by a user via mapred.job.map.memory.mb, in MB. A job that asks for more than this number will be failed at submission itself.  
mapred.cluster.max.reduce.memory.mb   set by admin, cluster-wide   Max limit on jobs. The maximum value that can be specified by a user via mapred.job.reduce.memory.mb, in MB. A job that asks for more than this number will be failed at submission itself.  

不设置时默认都是-1,无限制

相关介绍请参考Hadoop-0.20.2 作业内存控制策略分析

设置时请注意其大小关系。比如你设置了mapred.cluster.map.memory.mb为1024 ,然后你提交任务时没有设置mapred.job.map.memory.mb(默认为-1,无限制),此时便会报如下错误:

2012-06-13 16:18:10,951 ERROR exec.Task (SessionState.java:printError(380)) - Job Submission failed with exception 'org.apache.hadoop.ipc.RemoteException(java.io.IOException: job_201206131602_0003(-1 memForMapTasks -1 memForReduceTasks): Invalid job requirements.           at org.apache.hadoop.mapred.JobTracker.checkMemoryRequirements(JobTracker.java:5160)           at org.apache.hadoop.mapred.JobTracker.submitJob(JobTracker.java:3949)           at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)           at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)           at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)           at java.lang.reflect.Method.invoke(Method.java:597)           at org.apache.hadoop.ipc.RPC$Server.call(RPC.java:523)           at org.apache.hadoop.ipc.Server$Handler$1.run(Server.java:1383)           at org.apache.hadoop.ipc.Server$Handler$1.run(Server.java:1379)           at java.security.AccessController.doPrivileged(Native Method)           at javax.security.auth.Subject.doAs(Subject.java:396)           at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1059)           at org.apache.hadoop.ipc.Server$Handler.run(Server.java:1377)   )'   org.apache.hadoop.ipc.RemoteException: java.io.IOException: job_201206131602_0003(-1 memForMapTasks -1 memForReduceTasks): Invalid job requirements.           at org.apache.hadoop.mapred.JobTracker.checkMemoryRequirements(JobTracker.java:5160)           at org.apache.hadoop.mapred.JobTracker.submitJob(JobTracker.java:3949)           at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)           at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)           at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)           at java.lang.reflect.Method.invoke(Method.java:597)           at org.apache.hadoop.ipc.RPC$Server.call(RPC.java:523)           at org.apache.hadoop.ipc.Server$Handler$1.run(Server.java:1383)           at org.apache.hadoop.ipc.Server$Handler$1.run(Server.java:1379)           at java.security.AccessController.doPrivileged(Native Method)           at javax.security.auth.Subject.doAs(Subject.java:396)           at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1059)           at org.apache.hadoop.ipc.Server$Handler.run(Server.java:1377)              at org.apache.hadoop.ipc.Client.call(Client.java:1030)           at org.apache.hadoop.ipc.RPC$Invoker.invoke(RPC.java:224)           at org.apache.hadoop.mapred.$Proxy7.submitJob(Unknown Source)           at org.apache.hadoop.mapred.JobClient$2.run(JobClient.java:862)           at org.apache.hadoop.mapred.JobClient$2.run(JobClient.java:791)           at java.security.AccessController.doPrivileged(Native Method)           at javax.security.auth.Subject.doAs(Subject.java:396)           at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1059)           at org.apache.hadoop.mapred.JobClient.submitJobInternal(JobClient.java:791)           at org.apache.hadoop.mapred.JobClient.submitJob(JobClient.java:765)           at org.apache.hadoop.hive.ql.exec.ExecDriver.execute(ExecDriver.java:452)           at org.apache.hadoop.hive.ql.exec.MapRedTask.execute(MapRedTask.java:136)           at org.apache.hadoop.hive.ql.exec.Task.executeTask(Task.java:133)           at org.apache.hadoop.hive.ql.exec.TaskRunner.runSequential(TaskRunner.java:57)           at org.apache.hadoop.hive.ql.Driver.launchTask(Driver.java:1332)           at org.apache.hadoop.hive.ql.Driver.execute(Driver.java:1123)           at org.apache.hadoop.hive.ql.Driver.run(Driver.java:931)           at org.apache.hadoop.hive.service.HiveServer$HiveServerHandler.execute(HiveServer.java:191)           at org.apache.hadoop.hive.service.ThriftHive$Processor$execute.getResult(ThriftHive.java:629)           at org.apache.hadoop.hive.service.ThriftHive$Processor$execute.getResult(ThriftHive.java:617)           at org.apache.thrift.ProcessFunction.process(ProcessFunction.java:32)           at org.apache.thrift.TBaseProcessor.process(TBaseProcessor.java:34)           at org.apache.thrift.server.TThreadPoolServer$WorkerProcess.run(TThreadPoolServer.java:176)  

内容版权声明:除非注明,否则皆为本站原创文章。

转载注明出处:http://www.heiqu.com/26dda34466dedafbb0ee06f293a5d0bb.html