/spring-hadoop

Spring for Apache Hadoop is a framework for application developers to take advantage of the features of both Hadoop and Spring.

Primary LanguageJava

The Spring for Apache Hadoop provides extensions to Spring, Spring Batch, and Spring Integration to build manageable and robust pipeline solutions around Hadoop.

Spring for Apache Hadoop extends Spring Batch by providing support for reading from and writing to HDFS, running various types of Hadoop jobs (Java MapReduce, Streaming, Hive, Pig), HBase and Cascading interactions. An important goal is to provide excellent support for non-Java based developers to be productive using Spring Hadoop and not have to write any Java code to use the core feature set.

Spring for Apache Hadoop also applies the familiar Spring programming model to Java MapReduce jobs by providing support for dependency injection of simple jobs as well as a POJO based MapReduce programming model that decouples your MapReduce classes from Hadoop specific details such as base classes and data types.

Docs

You can find out more details from the user documentation or by browsing the javadocs. If you have ideas about how to improve or extend the scope, please feel free to contribute.

Artifacts

  • Maven:
<dependency>
  <groupId>org.springframework.data</groupId>
  <artifactId>spring-data-hadoop</artifactId>
  <version>${version}</version>
</dependency> 

<!-- used for nightly builds -->
<repository>
  <id>spring-maven-snapshot</id>
  <snapshots><enabled>true</enabled></snapshots>
  <name>Springframework Maven SNAPSHOT Repository</name>
  <url>http://repo.springsource.org/snapshot</url>
</repository> 

<!-- used for milestone/rc releases -->
<repository>
  <id>spring-maven-milestone</id>
  <name>Springframework Maven Milestone Repository</name>
  <url>http://repo.springsource.org/milestone</url>
</repository> 
  • Gradle:

Based on the artifact type, pick one of the repos below:

repositories {
  maven { url "http://repo.springsource.org/release" }
  maven { url "http://repo.springsource.org/milestone" }
  maven { url "http://repo.springsource.org/snapshot" }
}

dependencies {
   compile "org.springframework.data:spring-data-hadoop:${version}"
}

The dependency shown above is the standard one that includes the namespace support as well as core and batch support. If you don't use the namespace then you can use either the spring-data-hadoop-batch or spring-data-hadoop-core artifacts, depending on if you use any of the batch features or not.

For the Cascading support you need to include a dependency on the spring-cascading artifact.

The available releases can be seen in the SpringSource Repository

Building

Spring for Apache Hadoop uses Gradle as its build system. To build the system simply run:

gradlew

from the project root folder. This will compile the sources, run the tests and create the artifacts.

Supported distros

By default Spring for Apache Hadoop compiles against Apache Hadoop 1.0.x. Apache Hadoop 1.1.x (hadoop11) Apache Hadoop 1.2.x (hadoop12), Apache Hadoop 2.0.x Alpha (hadoop20), Cloudera CDH3 (cdh3), Cloudera CDH4 (cdh4), Hortonworks HDP 1.3 (hdp13) and Pivotal HD 1.0 (phd1) are also supported; to compile against them pass the -Pdistro=<label> project property, like so:

gradlew -Pdistro=hadoop11 build

Note that the chosen distro is displayed on the screen:

Using Apache Hadoop 1.1.x [1.1.2]

In this case, the specified Hadoop distribution (above Apache Hadoop 1.1.x) is used to create the project binaries. This option is useful when testing against Hadoop clusters incompatible with the Hadoop stable line.

CI Builds

The status of the CI builds are available at Status Summary Screen

We are currently running tests against the following distributions:

  • Apache Hadoop 1.0.4
  • Apache Hadoop 1.1.2
  • Cloudera CDH3
  • Cloudera CDH4
  • Greenplum HD 1.2
  • Hortonworks HDP 1.3

Testing

For its testing, Spring for Apache Hadoop expects a pseudo-distributed/local Hadoop instalation available on localhost configured with a port of 8020 for HDFS. The local Hadoop setup allows the project classpath to be automatically used by the Hadoop job tracker. These settings can be customized in two ways:

  • Build properties

From the command-line, use hd.fs for the file-system (to avoid confusion, specify the protocol such as 'hdfs://', 's3://', etc - if none is specified, hdfs:// will be used), hd.jt for the jobtracker and hd.hive for the Hive host/port information, to override the defaults. For example to run against HDFS at dumbo:8020 one would use:

gradlew -Phd.fs=hdfs://dumbo:8020 build
  • Properties file

Through the test.properties file under src/test/resources folder (further tweaks can be applied through hadoop-ctx.xml file under src/test/resources/org/springframework/data/hadoop).

Enabling Hbase/Hive/Pig/WebHdfs/Cascading Tests

Note that by default, only the vanilla Hadoop tests are running - you can enable additional tests (such as Hive or Pig) by adding the tasks enableHBaseTests, enableHiveTests, enablePigTests, enableWebHdfsTests or enableCascadingTests (or enableAllTests in short). Use test.properties file for customizing the default location for these services as well.

Disabling test execution

You can disable all tests by skipping the test task:

gradlew -x test

Contributing

Here are some ways for you to get involved in the community:

  • Get involved with the Spring community on the Spring Community Forums. Please help out on the forum by responding to questions and joining the debate.
  • Create JIRA tickets for bugs and new features and comment and vote on the ones that you are interested in.
  • Watch for upcoming articles on Spring by subscribing to springframework.org.

Github is for social coding: if you want to write code, we encourage contributions through pull requests from forks of this repository. If you want to contribute code this way, read the Spring Framework [contributor guidelines] (https://github.com/SpringSource/spring-framework/blob/master/CONTRIBUTING.md).

Staying in touch

Follow the project team (Costin, Mark, Thomas) on Twitter.

In-depth articles can be found at the SpringSource team blog, and releases are announced via our news feed.