Hadoop集群：Hadoop2.6.0，系统：windows7，开发环境：eclipse
Eclipse调用Hadoop运行MapReduce程序其实就是普通的java程序可以提交MR任务到集群执行而已。
1、首先需要配置环境变量：
在系统变量中新增：
然后再Path中增加：%HADOOP_HOME%\bin;
2、需要在开发的MapReduce的main函数中指定配置如下：
Configuration conf = new Configuration();
conf.setBoolean("mapreduce.app-submission.cross-platform", true);// 配置使用跨平台提交任务
conf.set("fs.defaultFS", "hdfs://imageHandler1:9000/tmp"); // 指定namenode
conf.set("mapreduce.framework.name", "yarn"); // 指定使用yarn框架
conf.set("yarn.resourcemanager.address", "imageHandler1:8032"); // 指定ResourceManager
conf.set("yarn.resourcemanager.scheduler.address", "imageHandler1:8030");// 指定资源分配器3、在eclipse中运行main函数：
例如运行wordcount。
（1）首先是出现类似下面的错误：
1. 2014-04-03 21:20:21,568 ERROR [main] util.Shell (Shell.java:getWinUtilsPath(303)) - Failed to locate the winutils binary in the hadoop binary path
2. java.io.IOException: Could not locate executable null\bin\winutils.exe in the Hadoop binaries.
3. at org.apache.hadoop.util.Shell.getQualifiedBinPath(Shell.java:278)
4. at org.apache.hadoop.util.Shell.getWinUtilsPath(Shell.java:300)
5. at org.apache.hadoop.util.Shell.<clinit>(Shell.java:293)
6. at org.apache.hadoop.util.StringUtils.<clinit>(StringUtils.java:76)
7. at org.apache.hadoop.yarn.conf.YarnConfiguration.<clinit>(YarnConfiguration.java:345)
8. at org.fansy.hadoop.mr.WordCount.getConf(WordCount.java:104)
9. at org.fansy.hadoop.mr.WordCount.runJob(WordCount.java:84)
10. at org.fansy.hadoop.mr.WordCount.main(WordCount.java:47)这个错误可以不用管，也可以在配置环境变量中的hadoop的bin目录下加入winutils.exe文件。
然后会报一个权限错误，需要调整相应目录的权限，例如修改conf配置中namenode相应目录的权限，这里是/tmp，在Hadoop集群中执行：hdfs dfs -chmod 777 /tmp。
（2）然后出现类似下面的错误：
1. 2014-04-03 20:32:36,596 ERROR [main] security.UserGroupInformation (UserGroupInformation.java:doAs(1494)) - PriviledgedActionException as:Administrator (auth:SIMPLE) cause:java.io.IOException: Failed to run job : Application application_1396459813671_0001 failed 2 times due to AM Container for appattempt_1396459813671_0001_000002 exited with exitCode: 1 due to: Exception from container-launch:
2. org.apache.hadoop.util.Shell$ExitCodeException: /bin/bash: line 0: fg: no job control
3.
4. at org.apache.hadoop.util.Shell.runCommand(Shell.java:464)
5. at org.apache.hadoop.util.Shell.run(Shell.java:379)
6. at org.apache.hadoop.util.Shell$ShellCommandExecutor.execute(Shell.java:589)
7. at org.apache.hadoop.yarn.server.nodemanager.DefaultContainerExecutor.launchContainer(DefaultContainerExecutor.java:195)
8. at org.apache.hadoop.yarn.server.nodemanager.containermanager.launcher.ContainerLaunch.call(ContainerLaunch.java:283)
9. at org.apache.hadoop.yarn.server.nodemanager.containermanager.launcher.ContainerLaunch.call(ContainerLaunch.java:79)
10. at java.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:334)
11. at java.util.concurrent.FutureTask.run(FutureTask.java:166)
12. at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
13. at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
14. at java.lang.Thread.run(Thread.java:724)
15. .Failing this attempt.. Failing the application- .
这时基本成功了一半，用上面的错误去google，可以得到这个网页：https://issues.apache.org/jira/browse/MAPREDUCE-5655 。Hadoop2.6.0只需修改
YARNRunner.java即可，MRApps.java不需要修改。
YARNRunner.java修改如下：
在该类中搜索“// Setup the command to run the AM”，然后注释掉“vargs.add(MRApps.crossPlatformifyMREnv(jobConf,Environment.JAVA_HOME) + "/bin/java");”，在该行下面增加：
String remoteOs = conf.get("mapred.remote.os");
vvargs.add("Linux".equals(remoteOs) ? "$JAVA_HOME/bin/java" : MRApps.crossPlatformifyMREnv(jobConf, Environment.JAVA_HOME) + "/bin/java");
修改完该类后替换Eclipse中的jar包hadoop-mapreduce-client-jobclient-2.6.0.jar中相应的类。
修改hadoop-mapreduce-client-core-2.6.0.jar中的mapred-default.xml（只需修改eclipse中引入的jar包即可），增加：
1. <property>
2. <name>mapred.remote.os</name>
3. <value>Linux</value>
4. <description>
5. Remote MapReduce framework's OS, can be either Linux or Windows
6. </description>
7. </property>（题外话，添加了这个属性后，按说我new一个Configuration后，我使用conf.get("mapred.remote.os")的时候应该是可以得到Linux的，但是我得到的却是null，这个就不清楚是怎么了。）
（3）这时再运行程序，还是报错，登录yarn主监控页面，查看log日志，可以看到下面的错误：
- Error: Could not find or load main class org.apache.hadoop.mapreduce.v2.app.MRAppMaster
按照https://issues.apache.org/jira/browse/MAPREDUCE-5655中的解决办法，修改mapred-default.xml和yarn-default.xml，分别在hadoop-mapreduce-client-core-2.6.0.jar和hadoop-yarn-common-2.6.0.jar中（只需修改eclipse中引入的jar包即可）。
在mapred-default.xml找到mapreduce.application.classpath，修改如下：
<property>
<name>mapreduce.application.classpath</name>
<value>
$HADOOP_CONF_DIR,
$HADOOP_COMMON_HOME/share/hadoop/common/*,
$HADOOP_COMMON_HOME/share/hadoop/common/lib/*,
$HADOOP_HDFS_HOME/share/hadoop/hdfs/*,
$HADOOP_HDFS_HOME/share/hadoop/hdfs/lib/*,
$HADOOP_MAPRED_HOME/share/hadoop/mapreduce/*,
$HADOOP_MAPRED_HOME/share/hadoop/mapreduce/lib/*,
$HADOOP_YARN_HOME/share/hadoop/yarn/*,
$HADOOP_YARN_HOME/share/hadoop/yarn/lib/*,
</value>
</property>
在yarn-default.xml中找到yarn.application.classpath，修改如下：
<property>
<name>yarn.application.classpath</name>
<value>
$HADOOP_CONF_DIR,
$HADOOP_COMMON_HOME/share/hadoop/common/*,
$HADOOP_COMMON_HOME/share/hadoop/common/lib/*,
$HADOOP_HDFS_HOME/share/hadoop/hdfs/*,
$HADOOP_HDFS_HOME/share/hadoop/hdfs/lib/*,
$HADOOP_MAPRED_HOME/share/hadoop/mapreduce/*,
$HADOOP_MAPRED_HOME/share/hadoop/mapreduce/lib/*,
$HADOOP_YARN_HOME/share/hadoop/yarn/*,
$HADOOP_YARN_HOME/share/hadoop/yarn/lib/*
</value>
</property>（4）经过上面的修改再次运行报错类似：
1. Caused by: java.lang.ClassNotFoundException: Class org.fansy.hadoop.mr.WordCount$WCMapper not found
2. at org.apache.hadoop.conf.Configuration.getClassByName(Configuration.java:1626)
3. at org.apache.hadoop.conf.Configuration.getClass(Configuration.java:1718)
4. ... 8 more
需要上传wordcount程序的jar文件到$HADOOP_HOME/share/hadoop/mapreduce/lib下面（集群每台机器都要上传），然后再次运行，成功了。
