即使是一个简单的wordcount mapreduce也会出现同样的错误。
hadoop 2.6.0版
下面是Yarn原木。
在资源协商期间似乎发生了某种超时。
但我无法验证相同的，究竟是什么原因导致超时。
2016-11-11 15:38:09313 info org.apache.hadoop.yarn.server.resourcemanager.amlauncher.amlauncher:启动appattempt时出错\u 1478856936677 \u 0004 \u000002。获取异常：java.io.ioexception:本地异常失败：java.io.ioexception:java.net.sockettimeoutexception:等待通道准备好读取时超时60000毫秒。ch:java.nio.channels.socketchannel[connected local=/10.0.37.145:49054 remote=platform demo/10.0.37.145:60487]；主机详细信息：本地主机为：“平台演示/10.0.37.145”；目的主机为：“平台演示”：60487；在org.apache.hadoop.net.netutils.wrapexception（netutils。java:772)在org.apache.hadoop.ipc.client.call（client。java:1472)在org.apache.hadoop.ipc.client.call（client。java:1399)在org.apache.hadoop.ipc.protobufrpceengine$invoker.invoke（protobufrpceengine。java:232)在com.sun.proxy.$proxy79.startcontainers（未知源）上org.apache.hadoop.yarn.api.impl.pb.client.containermanagementprotocolpbclientimpl.startcontainers（containermanagementprotocolpbclientimpl。java:96)在org.apache.hadoop.yarn.server.resourcemanager.amlauncher.amlauncher.launch（amlauncher。java:119)在org.apache.hadoop.yarn.server.resourcemanager.amlauncher.amlauncher.run（amlauncher。java:254)在java.util.concurrent.threadpoolexecutor.runworker（threadpoolexecutor。java:1145)在java.util.concurrent.threadpoolexecutor$worker.run（threadpoolexecutor。java:615)在java.lang.thread.run（线程。java:745)原因：java.io.ioexception:java.net.sockettimeoutexception:60000毫秒超时等待通道准备好读取。ch:java.nio.channels.socketchannel[connected local=/10.0.37.145:49054 remote=platform demo/10.0.37.145:60487]位于org.apache.hadoop.ipc.client$connection$1.run（client。java:680)位于javax.security.auth.subject.doas（subject）的java.security.accesscontroller.doprivileged（本机方法）。java:415)在org.apache.hadoop.security.usergroupinformation.doas（用户组信息。java:1628)在org.apache.hadoop.ipc.client$connection.handlesaslconnectionfailure（client。java:643)在org.apache.hadoop.ipc.client$connection.setupiostreams（client。java:730)在org.apache.hadoop.ipc.client$connection.access$2800（client。java:368)在org.apache.hadoop.ipc.client.getconnection（客户端。java:1521)在org.apache.hadoop.ipc.client.call（client。java:1438) ... 9更多原因：java.net.sockettimeoutexception:等待通道准备好读取时超时60000毫秒。ch:java.nio.channels.socketchannel[connected local=/10.0.37.145:49054 remote=platform demo/10.0.37.145:60487]位于org.apache.hadoop.net.socketiowithtimeout.doio（socketiowithtimeout）。java:164)在org.apache.hadoop.net.socketinputstream.read（socketinputstream。java:161)在org.apache.hadoop.net.socketinputstream.read（socketinputstream。java:131)在java.io.filternputstream.read（filternputstream。java:133)在java.io.bufferedinputstream.fill（bufferedinputstream。java:235)在java.io.bufferedinputstream.read（bufferedinputstream。java:254)在java.io.datainputstream.readint（datainputstream。java:387)在org.apache.hadoop.security.saslrpclient.saslconnect（saslrpclient。java:367)在org.apache.hadoop.ipc.client$connection.setupsaslconnection（客户端。java:553)在org.apache.hadoop.ipc.client$connection.access$1800（client。java:368)在org.apache.hadoop.ipc.client$connection$2.run（client。java:722)在org.apache.hadoop.ipc.client$connection$2.run（client。java:718)位于java.security.accesscontroller.doprivileged（本机方法）javax.security.auth.subject.doas（主题。java:415)在org.apache.hadoop.security.usergroupinformation.doas（usergroupinformation。java:1628)在org.apache.hadoop.ipc.client$connection.setupiostreams（client。java:717) ... 12个以上
2016-11-11 15:38:09319 info org.apache.hadoop.yarn.server.resourcemanager.rmapp.attempt.rmappattemptimpl:更新应用程序attempt appattempt\u 1478856936677\u 0004\u000002，最终状态：失败，退出状态：-1000 2016-11-11 15:38:09，319 info org.apache.hadoop.yarn.server.resourcemanager.rmapp.attempt.rmappattemptimpl:appattempt\u 1478856936677\u 0004\u000002状态从已分配更改为最终保存\u
我试图改变以下属性
yarn.nodemanager.resource.memory-mb内存
2200可分配给容器的物理内存量（mb）。
yarn.scheduler.minimum-allocation-mb最小分配
500
dfs.datanode.socket.write.timeout超时
3000000
dfs.socket.timeout 300万