First run MapReduce, recorded several problems encountered, Hadoop cluster is CDH version, but my Windows local jar package is directly with hadoop2.6.0 version, and did not specifically look for CDH version of the
1.Exception in thread "main" Java.lang.NullPointerException Atjava.lang.ProcessBuilder.start
Download Hadoop2 above version, in the Hadoop2 bin directory without Winutils.exe and Hadoop.dll, find the corresponding version of the two files on the web, placed under the bin of Hadoop, and add hadoop_ in the environment variables HOME and add%hadoop_home%/bin in path, restart the computer and then run it is not the problem.
2.atorg.apache.hadoop.io.nativeio.nativeio$windows.access (nativeio.java:557)
Download Hadoop-2.6.0-src.tar and unzip to find hadoop-2.6.0-src\hadoop-common-project\hadoop-common\src\main\java\org\apache\ Hadoop\io\nativeio under Nativeio.java copy into the project, the package path is Org.apache.hadoop.io.nativeio, then change the No. 557 behavior to return true.
3. operation result not error, only run result is false, view Hadoop activation node log found error: partialgroupnameexception the user name ' Ushio ' is not found. Id:ushio: No such user
Add the Hadoop_user_name variable to the environment variable, the value is the correct user name to execute Hadoop, Cloudera Manager installed CDH version of Hadoop, the value is HDFs, restart the computer and then normal operation.
In the following page to find the solution, the rest of the errors mentioned I did not meet.
http://blog.csdn.net/congcong68/article/details/42043093
Remote connection to Hadoop cluster debug MapReduce Error Record under Windows on Eclipse