Machine has soul.

Debug Hadoop Applications With IntelliJ

| Comments

In my previous blog, I explained how to Create a Hadoop Compilation and Development Environment so that you can build the Hadoop source code you get from the Apache Hadoop git repository.

However, not everyone need to dig into the source code, normally more people will just call the Hadoop APIs to write a MapReduce program. When you write a MapReduce program, you rarely get it right for one time, then you need to debug your code.

This time I’ll show you how to debug your Hadoop applications using the IntelliJ IDE.

Environment: CentOS 6.6, Oracle JDK 1.7.0_75, Maven 3.2.5, IntelliJ Idea 14.0.3

1. Create a MapReduce Maven project

First we need to create a MapReduce program to debug. Let’s use the simplest example, WordCount, for demonstration. The source code is here WordCount – Hadoop Wiki.

1.1 Generate the Maven project

Use the mvn command to generate the scaffolding for the project.

mvn archetype:generate -DgroupId=me.soulmachine -DartifactId=wordcount -DarchetypeArtifactId=maven-archetype-quickstart -DinteractiveMode=false

Delete the src/test/java/me/soulmachine/ file, as it will not be used in this example, and rename the src/main/java/me/soulmachine/ file to

Now the structure of the project is as the following:

tree ./wordcount
├── pom.xml
└── src
    ├── main
    │   └── java
    │       └── me
    │           └── soulmachine
    │               └──
    └── test
        └── java
            └── me
                └── soulmachine

1.2 Write the code

Rename the src/main/java/me/soulmachine/ file to, edit the file and copy and paste the code from WordCount – Hadoop Wiki, remember to change the first line from package org.myorg to package me.soulmachine.

I made some improvements to this code, check it out:

1.3 Edit the pom.xml file

The content of the pom.xml file is as the following:

<project xmlns="" xmlns:xsi=""




1.4 Compile

mvn clean package

If this step succeeds, move to the next step.

2. Debug WordCount in IntelliJ Idea

2.1 Open the project in IntelliJ Idea

Launch IntelliJ Idea, click File->Open and open the pom.xml, thus you can import the project into IntelliJ Idea.

2.2 Create a Run/Debug Configuration

Select the project “WordCount”, Click menu Run->Edit Configurations ...,

  1. Click the button on the right of Main Class and select me.soulmachine.WordCount as the main class.
  2. Set Program Arguments to input/ output, and create a directory named input in the project directory, no need to create the output/ directory
  3. Leave other fields as default.
  4. Click OK and finish.

2.3 Run the program

  1. Create a text file named file1.txt in the input/ directory and set its content as Hello World Bye World, create another text file named file2.txt and set its content to Hello Hadoop Goodbye Hadoop.
  2. Click menu Run->Run 'Word Count', after it finishes, you will see a fold named output/ in the project root directory and there are two files named part-r-00000 and _SUCCESS.

2.4 Debug the program

Set a break point in the main function and click menu Run->Debug 'Word Count', then you can debug the code step by step.


  1. Debugging Hadoop Applications with IntelliJ |
  2. WordCount – Hadoop Wiki
  3. Maven in 5 Minutes
  4. Develop Java MapReduce programs for Hadoop in HDInsight | Azure