Submitting jobs

About this task

When you submit a job for a MapReduce application, the MapReduce framework sources the environment configuration from the pmr-env.sh configuration file, located on the host at the $SOAM_HOME/mapreduce/conf/ directory, and submits the job.

Procedure

  1. Click Workload > MapReduce > Jobs.
  2. Click New.

    The Submit Job window displays.

  3. Specify parameters for the job.
    1. From the Application Name dropdown list, select the application for which you want to submit jobs.
    2. In the Job Priority text box, specify the priority for the job (between 1 and 10000) so that its workload takes priority over other jobs in the same application.
    3. Upload the application .jar file that the job must use either from your local host or from the server.

      • To add the .jar file from your local computer, click Add Local File, navigate to the file, and click Open. When you upload a file from disk, the file you specified is copied to the work directory ($SOAM_HOME/mapreduce/work/) on the web server.

      • To add the .jar file from the server, click Add Server File , navigate to the file, and click OK.

        The directory that is set as My Samples is the samples directory located by default under $SOAM_HOME/mapreduce/version/os_type/samples/.

    4. In the Main Class text box, enter the class that should be invoked for the job; for example, wordcount.
    5. In the Main Class Options text box, enter options that specify additional settings for the job; for example, HDFS input and output directories.
    6. (Optional) Add additional configuration parameters for the job run, click Add.
      1. Under Job Configuration, click Add.
      2. In the Name dropdown list, choose a parameter.
      3. Accept the value for the parameter configured at the application level or change the value in the Value text box.
      4. Repeat steps i to iii to add more parameters.
  4. Click Submit.
    The Submit Job popup displays, showing progress of the job.
    Attention: Closing the popup while the job is still running terminates the job. Keep the popup open till the job is done.