Verifying installation

After HDP with IBM Spectrum® Scale service is deployed, verify the installation setup.

Note: To run the IBM Spectrum Scale commands, add the /usr/lpp/mmfs/bin directory to the environment PATH.
  1. Verify all uid/gid values for user to ensure that they are consistent across all IBM Spectrum Scale nodes. Check by using mmdsh -N all id<user-name> to see whether the UID is consistent across all nodes.
  2. Check the IBM Spectrum Scale installed packages on all nodes by using rpm -qa | grep gpfs to verify that all base IBM Spectrum Scale packages have been installed.
  3. As user, ambari-qa, check the user id and access to the file system.
    HDFS commands:
    $hadoop fs -ls /user
    
    POSIX commands:
    $pwd; ls -ltr
    
    #Create a file with entry
    $ echo "My test" > mytest
    $ cat mytest
    
    HDFS commands:
    $ hadoop fs -cat mytest
    
    POSIX commands:
    $ rm mytest
    $ls -ltr
  4. Run wordcount as user.
    1. Create a mywordcountfile.
    2. Use the mywordcountfile file to be used as input to the wordcount program.
      $ yarn jar
      /usr/hdp/3.0.0.0-1634/hadoop-mapreduce/ hadoop-mapreduce-examples-3.1.0.3.0.0.0-1634.jar
      wordcount mywordcountfile wc_output
    3. Check the output in directory
      $ hadoop fs -ls wc_output
  5. Run teragen/terasort. See Hortonworks Smoke Test Mapreduce.