Deploying a Hadoop Job to Amazon Elastic MapReduce
You can deploy a regular Hadoop job to an Amazon Elastic MapReduce cluster by selecting a .jar file and deploying it to an Elastic MapReducecluster you have configured.
You can also deploy a job you've developed from the Job Developer in Hadoop Studio. For more on how to do this, see Deploying a Job from the Job Developer
To deploy a job to a an Amazon Elastic MapReduce cluster, you create a job object in the Hadoop Manager presented via NetBeans Services. Karmasphere Studio then enables you to run that job on one or more clusters. It also enables you to re-run the same job or to create a basic job whose parameters you change every time you run it. If you want to experiment with Amazon Elastic MapReduce clusters with different properties, just set up the cluster objects and every time you run the job, you can associate it with a different cluster.
First, let's create the job object by right clicking your mouse on "Hadoop Jobs" and selecting "New Job...".
Then you give the job a name and select the job type. Select "Hadoop Job from pre-existing JAR file".
Now you configure the JAR file and give Karmasphere Studio class directives.
First, browse to select your JAR file.
The check boxes direct Karmasphere Studio to interpret meta data stored in the jar file in standard and/or Hadoop-specific modes. These options tell Karmasphere Studio to attempt to build the class path intelligently.
If you need to configure the class path manually, now is your opportunity to do it
In the final step, we're going to associate a default cluster and arguments with the job. These can be changed when we come to run a job. If you've already set up a remote cluster, then go ahead and select it. If you haven't yet set up a cluster, click "Add..." and read Using an Amazon Elastic MapReduc Cluster.
Now you've finished creating your job object, it's time to deploy it. Select the job name you just created under "Hadoop Jobs", right click and choose "Run Job...".
Now you are presented with the parameters you used when you created the job object, including summary information about the Amazon Elastic MapReduct cluster you have selected. If you want to change any of these, go right ahead. When you're happy with them, click the "Run" button.
Your job will now be sent to the remote cluster you associated it with and Karmasphere Studio will automatically begin monitoring the job. It may take a few seconds, but you will see the graphical Elastic MapReduct Cluster Monitor window appear. To learn more about this, we suggest you take a look at Monitoring a Hadoop Job on Amazon Elastic MapReduct .