我是hadoop新手。
我正在尝试设置giraph在hadoop-2.6.5上运行。
当我提交giraph作业时，作业成功提交，但失败，我在容器syslog中得到以下日志：
2018-01-30 12:09:01190 info[main]org.apache.hadoop.mapreduce.v2.app.mrappmaster:为应用程序appattempt\u 1517293264136\u 0002\u000002 2018-01-30 12:09:01437 warn[main]org.apache.hadoop.util.nativecodeloader:无法为您的平台加载本机hadoop库。。。在适用的情况下使用内置java类2018-01-30 12:09:01471 info[main]org.apache.hadoop.mapreduce.v2.app.mrappmaster:使用令牌执行：2018-01-30 12:09:01471 info[main]org.apache.hadoop.mapreduce.v2.app.mrappmaster:种类：yarn\u am\u rm\u令牌，服务：，ident:（appattemptid{application\u id{id:2 cluster\u timestamp:1517293264136}attemptid:2}keyid:-1485907628）2018-01-30 12:09:01583 info[main]org.apache.hadoop.mapreduce.v2.app.mrappmaster:使用mapred newapicommitter。2018-01-30 12:09:02154 info[main]org.apache.hadoop.mapreduce.v2.app.mrappmaster:outputcommitter set in config null 2018-01-30 12:09:02，207 fatal[main]org.apache.hadoop.mapreduce.v2.app.mrappmaster:启动mrappmaster java.lang.noclassdeffounderror:io/netty/buffer/bytebufallocator at org.apache.giraph.bsp.bspoutputformat.getoutputcommitter（bspoutputformat）时出错。java:62)在org.apache.hadoop.mapreduce.v2.app.mrappmaster$1.call（mrappmaster。java:470)在org.apache.hadoop.mapreduce.v2.app.mrappmaster$1.call（mrappmaster。java:452)在org.apache.hadoop.mapreduce.v2.app.mrappmaster.callwithjobclassloader（mrappmaster。java:1541)在org.apache.hadoop.mapreduce.v2.app.mrappmaster.createoutputcommitter（mrappmaster。java:452)在org.apache.hadoop.mapreduce.v2.app.mrappmaster.serviceinit（mrappmaster。java:371) 在org.apache.hadoop.service.abstractservice.init（abstractservice。java:163)在org.apache.hadoop.mapreduce.v2.app.mrappmaster$4.run（mrappmaster。java:1499)位于javax.security.auth.subject.doas（subject）的java.security.accesscontroller.doprivileged（本机方法）。java:422)在org.apache.hadoop.security.usergroupinformation.doas（用户组信息。java:1692)在org.apache.hadoop.mapreduce.v2.app.mrappmaster.initandstartapmaster（mrappmaster。java:1496)位于org.apache.hadoop.mapreduce.v2.app.mrappmaster.main（mrappmaster.com）。java:1429)原因：java.lang.classnotfoundexception:io.netty.buffer.bytebufallocator atjava.net.urlclassloader.findclass（urlclassloader。java:381)在java.lang.classloader.loadclass（classloader。java:424)在sun.misc.launcher$appclassloader.loadclass（launcher。java:331)在java.lang.classloader.loadclass（classloader。java:357) ... 更多信息2018-01-30 12:09:02209信息[main]org.apache.hadoop.util.exitil:退出状态1
日志中的诊断显示以下日志：
应用程序应用程序\u 1517293264136 \u 0002失败2次，原因是appattempt的am容器\u 1517293264136 \u 0002 \u000002退出，exitcode:1有关更详细的输出，请检查应用程序跟踪page:http用法：/172.16.0.218:8088/proxy/application\u 1517293264136\u 0002/然后，单击指向每次尝试的日志的链接。诊断：容器启动异常。容器id:container\u 1517293264136\u 0002\u 02\u000001退出代码：1堆栈跟踪：exitcodeexception exitcode=1:at org.apache.hadoop.util.shell.runcommand（shell）。java:575)在org.apache.hadoop.util.shell.run（shell。java:478)在org.apache.hadoop.util.shell$shellcommandexecutor.execute（shell。java:766)在org.apache.hadoop.yarn.server.nodemanager.defaultcontainerexecutor.launchcontainer（defaultcontainerexecutor）。java:212)在org.apache.hadoop.yarn.server.nodemanager.containermanager.launcher.containerlaunch.call（containerlaunch。java:302)在org.apache.hadoop.yarn.server.nodemanager.containermanager.launcher.containerlaunch.call（containerlaunch。java:82)在java.util.concurrent.futuretask.run（futuretask。java:266)位于java.util.concurrent.threadpoolexecutor.runworker（threadpoolexecutor。java:1142)在java.util.concurrent.threadpoolexecutor$worker.run（threadpoolexecutor。java:617)在java.lang.thread.run（线程。java:745)容器退出，退出代码为非零，尝试失败。应用程序失败。
它不及格的班级是 io/netty/buffer/ByteBufAllocator 在netty所有的jar里：https://mvnrepository.com/artifact/io.netty/netty-all
从其他问题来看，我尝试过在hadoop\u类路径中添加jar。

Yogin-Patel:hadoop yoginpatel$ echo $HADOOP_CLASSPATH
/Users/yoginpatel/Downloads/gradle-4.3/caches/modules-2/files-2.1/io.netty/netty-all/4.0.43.Final/9781746a179070e886e1fb4b1971a6bbf02061a4/netty-all-4.0.43.Final.jar
Yogin-Patel:hadoop yoginpatel$

它出现在 hadoop classpath 也。

Yogin-Patel:hadoop yoginpatel$ hadoop classpath
/Users/yoginpatel/hadoop/etc/hadoop:/Users/yoginpatel/hadoop/share/hadoop/common/lib/*:/Users/yoginpatel/hadoop/share/hadoop/common/*:/Users/yoginpatel/hadoop/share/hadoop/hdfs:/Users/yoginpatel/hadoop/share/hadoop/hdfs/lib/*:/Users/yoginpatel/hadoop/share/hadoop/hdfs/*:/Users/yoginpatel/hadoop/share/hadoop/yarn/lib/*:/Users/yoginpatel/hadoop/share/hadoop/yarn/*:/Users/yoginpatel/hadoop/share/hadoop/mapreduce/lib/*:/Users/yoginpatel/hadoop/share/hadoop/mapreduce/*:/Users/yoginpatel/Downloads/gradle-4.3/caches/modules-2/files-2.1/io.netty/netty-all/4.0.43.Final/9781746a179070e886e1fb4b1971a6bbf02061a4/netty-all-4.0.43.Final.jar:/contrib/capacity-scheduler/*.jar
Yogin-Patel:hadoop yoginpatel$

我正在尝试在开发环境中设置。这是单节点设置。
我甚至试过

job.addFileToClassPath(new Path("/Users/yoginpatel/Downloads/gradle-4.3/caches/modules-2/files-2.1/io.netty/netty-all/4.0.43.Final/9781746a179070e886e1fb4b1971a6bbf02061a4/netty-all-4.0.43.Final.jar"));

这些方法都没用。如何让hadoop节点访问必要的jar？
这是一个giraphjob submit代码，它将map reduce job提交到集群：

@Test
    public void testPageRank() throws IOException, ClassNotFoundException, InterruptedException {

        GiraphConfiguration giraphConf = new GiraphConfiguration(getConf());
        giraphConf.setWorkerConfiguration(1,1,100);
        GiraphConstants.SPLIT_MASTER_WORKER.set(giraphConf, false);

        giraphConf.setVertexInputFormatClass(JsonLongDoubleFloatDoubleVertexInputFormat.class);
        GiraphFileInputFormat.setVertexInputPath(giraphConf,
                                                 new Path("/input/tiny-graph.txt"));
        giraphConf.setVertexOutputFormatClass(IdWithValueTextOutputFormat.class);

        giraphConf.setComputationClass(PageRankComputation.class);

        GiraphJob giraphJob = new GiraphJob(giraphConf, "page-rank");
        giraphJob.getInternalJob().addFileToClassPath(new Path("/Users/yoginpatel/Downloads/gradle-4.3/caches/modules-2/files-2.1/io.netty/netty-all/4.0.43.Final/9781746a179070e886e1fb4b1971a6bbf02061a4/netty-all-4.0.43.Final.jar"));

        FileOutputFormat.setOutputPath(giraphJob.getInternalJob(),
                                       new Path("/output/page-rank2"));
        giraphJob.run(true);
    }

    private Configuration getConf() {
        Configuration conf = new Configuration();
        conf.set("fs.defaultFS", "hdfs://localhost:9000");

        conf.set("yarn.resourcemanager.address", "localhost:8032");

        // framework is now "yarn", should be defined like this in mapred-site.xm
        conf.set("mapreduce.framework.name", "yarn");
        return conf;
    }