Configuring and running Spark on Amazon Elastic Map Reduce

Launch a Hadoop cluster with Spark installed using the Amazon Elastic Map Reduce. Perform the following steps to create an EMR cluster with Spark installed:

1. Launch an Amazon EMR Cluster.
2. Open the Amazon EMR UI console at https://console.aws.amazon.com/elasticmapreduce/.
3. Choose Create cluster:

4. Choose appropriate Amazon AMI Version 3.9.0 or later as shown in the following screenshot:

5. For the applications to be installed field, choose Spark 1.5.2 or later from the list shown on the User Interface and click on Add.
6. Select other hardware options as necessary:
   • The Instance Type
   • The keypair to be used with SSH
   • Permissions
   • IAM roles (Default orCustom)

Refer to the following screenshot:

7. Click on Create cluster. The cluster will start instantiating as shown in the following screenshot:

8. Log in into the master. Once the EMR cluster is ready, you can SSH into the master:

   $ ssh -i rd_spark-user1.pem
   [email protected]    

The output will be similar to following listing:

     Last login: Wed Jan 13 10:46:26 2016

          __|  __|_  )
          _|  (     /   Amazon Linux AMI
         ___|___|___|

     https://aws.amazon.com/amazon-linux-ami/2015.09-release-notes/
     23 package(s) needed for security, out of 49 available
     Run "sudo yum update" to apply all updates.
     [hadoop@ip-172-31-2-31 ~]$ 

9. Start the Spark Shell:

      [hadoop@ip-172-31-2-31 ~]$ spark-shell
      16/01/13 10:49:36 INFO SecurityManager: Changing view acls to: 
          hadoop
      16/01/13 10:49:36 INFO SecurityManager: Changing modify acls to: 
          hadoop
      16/01/13 10:49:36 INFO SecurityManager: SecurityManager: 
          authentication disabled; ui acls disabled; users with view 
          permissions: Set(hadoop); users with modify permissions: 
          Set(hadoop)
      16/01/13 10:49:36 INFO HttpServer: Starting HTTP Server
      16/01/13 10:49:36 INFO Utils: Successfully started service 'HTTP 
          class server' on port 60523.
      Welcome to
            ____              __
           / __/__  ___ _____/ /__
          _ / _ / _ `/ __/  '_/
         /___/ .__/_,_/_/ /_/_   version 1.5.2
            /_/
      scala> sc

10. Run Basic Spark sample from the EMR:

    scala> val textFile = sc.textFile("s3://elasticmapreduce/samples
      /hive-ads/tables/impressions/dt=2009-04-13-08-05
      /ec2-0-51-75-39.amazon.com-2009-04-13-08-05.log")
   scala> val linesWithCartoonNetwork = textFile.filter(line =>  
      line.contains("cartoonnetwork.com")).count()

Your output will be as follows:

     linesWithCartoonNetwork: Long = 9