Find file History
Fetching latest commit…
Cannot retrieve the latest commit at this time.

Spring Yarn Examples

NOTE: These samples are currently configured to use Spring Hadoop 2.0.0.M4 release. You will get Hadoop dependencies through different maven artifacts for Spring Hadoop. Currently vanilla Hadoop 2.2.0-alpha is supported.

# gradlew -Pdistro=hadoop20 clean build -x test

This project provides a number of examples to get you started using Spring Yarn and Spring Hadoop. These examples are designed to work with [Spring Hadoop] ( 2.0 or higher and are organized into the following sub projects:

Yarn Examples

  • batch-amjob - Example of running Spring Batch job only on Application Master
  • batch-files - Example of running Spring Batch partitioned job with HDFS file processing
  • batch-partition - Example of running Spring Batch partitioned job
  • custom-amservice - Example of using Spring Integration as custom Application Master Service
  • kill-application - Example of killing an applicatoin
  • list-applications - Example of listing applications known to resource manager
  • multi-context - Example of running a spring application context on multiple containers
  • simple-command - Example of running a command with containers
  • restart-context - Example of custom failed container handling

Imporing Examples to IDE

This project is built with gradle and each example may be imported to your Java IDE. If you are using Eclipse or SpringSource Tool Suite, go to the directory where you downloaded this project and type:

    # gradlew eclipse

If you are using IDEA,

    # gradlew idea

Running Examples

Detailed instructions for each example may be found in its own README file. These README files has a section Quick Instructions which is handy if you're already familiar with example and just want to copy/paste commands.

General Prerequisites

Examples depends on libraries which are not on Hadoop classpath on default. These dependencies are collected together during the examples build process.

For example if project yarn-examples-simple-command is built:

# gradlew :yarn-examples-common:yarn-examples-simple-command:build -x test

Project files are collected under simple-command/build/libs and dependencies under simple-command/build/dependency-libs.

Every example jar file has its own directory location in HDFS.

[root@centos hadoop]# hadoop/bin/hdfs dfs -ls /lib

[root@centos hadoop]# hadoop/bin/hdfs dfs -ls /app/simple-command

JUnit tests are not dependant on a running cluster but you need to have Hadoop installed on a host where tests are executed. Also familiar enviroment variables shown below must exist.

export PATH=$PATH:/usr/local/javas/java/bin
export JAVA_HOME=/usr/local/javas/java
export JRE_HOME=/usr/local/javas/java/jre
export HADOOP_HOME=/usr/local/hadoops/hadoop