Running Hadoop Job Remotely

Question

I am trying to run a MapReduce job from outside the cluster.

e.g. Hadoop cluster is running on Linux machines. We have one web application running on a Windows machine. We want to run the hadoop job from this remote web application. We want to retrieve the hadoop output directory and present it as a Graph.

We have written the following piece of code:

Configuration conf = new Configuration();

Job job = new Job(conf);

conf.set("mapred.job.tracker", "192.168.56.101:54311"); 

conf.set("fs.default.name", "hdfs://192.168.56.101:54310");

job.setJarByClass(Analysis.class) ;
//job.setOutputKeyClass(Text.class);
//job.setOutputValueClass(IntWritable.class);

job.setMapperClass(Map.class);
job.setReducerClass(Reduce.class);



//job.set

job.setInputFormatClass(CustomFileInputFormat.class);
job.setOutputFormatClass(TextOutputFormat.class);

FileInputFormat.addInputPath(job, new Path(args[0]));
FileOutputFormat.setOutputPath(job, new Path(args[1]));

job.setMapOutputKeyClass(Text.class);
job.setMapOutputValueClass(IntWritable.class);
job.setOutputKeyClass(Text.class);
job.setOutputValueClass(IntWritable.class);


job.waitForCompletion(true);

And this is the error we get. Even if we shut down the hadoop 1.1.2 cluster, the error is still the same.

14/03/07 00:23:37 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
14/03/07 00:23:37 ERROR security.UserGroupInformation: PriviledgedActionException as:user cause:java.io.IOException: Failed to set permissions of path: \tmp\hadoop-user\mapred\staging\user818037780\.staging to 0700
Exception in thread "main" java.io.IOException: Failed to set permissions of path: \tmp\hadoop-user\mapred\staging\user818037780\.staging to 0700
at org.apache.hadoop.fs.FileUtil.checkReturnValue(FileUtil.java:691)
at org.apache.hadoop.fs.FileUtil.setPermission(FileUtil.java:664)
at org.apache.hadoop.fs.RawLocalFileSystem.setPermission(RawLocalFileSystem.java:514)
at org.apache.hadoop.fs.RawLocalFileSystem.mkdirs(RawLocalFileSystem.java:349)
at org.apache.hadoop.fs.FilterFileSystem.mkdirs(FilterFileSystem.java:193)
at org.apache.hadoop.mapreduce.JobSubmissionFiles.getStagingDir(JobSubmissionFiles.java:126)
at org.apache.hadoop.mapred.JobClient$2.run(JobClient.java:942)
at org.apache.hadoop.mapred.JobClient$2.run(JobClient.java:936)
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:1190)
at org.apache.hadoop.mapred.JobClient.submitJobInternal(JobClient.java:936)
at org.apache.hadoop.mapreduce.Job.submit(Job.java:550)
at org.apache.hadoop.mapreduce.Job.waitForCompletion(Job.java:580)
at LineCounter.main(LineCounter.java:86)

Answer 1

While running from a remote system, you should run as remote user. You can do it in your main class as follows:

public static void main(String a[]) {
     UserGroupInformation ugi
     = UserGroupInformation.createRemoteUser("root");

     try {


        ugi.doAs(new PrivilegedExceptionAction<Void>() {

            public Void run() throws Exception {
                 Configuration conf = new Configuration();

                 Job job = new Job(conf);

                 conf.set("hadoop.job.ugi", "root");

                 // write your remaining piece of code here. 

                return null;
            }
        });

    } catch (Exception e) {
        e.printStackTrace();
    }

}

Also while submitting a mapreduce job, it should copy your java classes with their dependent jars to hadoop cluster, where it execute mapreduce job.You can read more here.

So you need to create a runnable jar of your code (with main class Analysis in your case) with all dependent jar files inits manifest classpath. Then run your jar file from your commandline using

java -jar job-jar-with-dependencies.jar arguments

HTH!