Skip to content

scaleoutsoftware/hServer

Repository files navigation

Overview

ScaleOut hServer® is a MapReduce execution engine that runs on the ScaleOut StateServer (SOSS) in-memory compute engine to deliver fast execution times with linearly scalable speed-up for standard, unchanged Hadoop MapReduce applications. Additionally, the hServer library enables Hadoop MapReduce applications to continuously analyze live, fast-changing, in-memory data as well as HDFS files. A single line code change is all that is required to run a standard Apache Hadoop MapReduce application with ScaleOut hServer without YARN support. If ScaleOut hServer is co-located on a YARN-enabled Hadoop cluster, you can simply set the "mapreduce.framework.name" property to "hserver-yarn" (e.g., adding the -Dmapreduce.framework.name=hserver-yarn argument from the command-line) to run a Hadoop MapReduce application unchanged with ScaleOut hServer. For more information regarding ScaleOut hServer, please visit the ScaleOut hServer product page which explains ScaleOut hServer's architecture and usage in detail.

Documentation

The Javadoc for ScaleOut hServer can be found on our website under 'ScaleOut hServer V2 Documentation'. The soss-jnc-5.3-javadoc.jar JAR is also located on our website under 'API Documentation for ScaleOut StateServer and ScaleOut Analytics Server' and provides documentation for SOSS API usage. All additional ScaleOut hServer and ScaleOut StateServer documentation can be found on the ScaleOut Software Support website under Product Documentation.

Building

Prerequisites

  1. Download ScaleOut StateServer
  2. Install ScaleOut StateServer using hServer's quick start guide
  3. Download and Install Apache Ant

Linux

  • Clone the ScaleOut hServer project to a local directory.
  • Copy the SOSS Java Named Cache JAR (soss-jnc-5.3.jar) from the Linux installation Java API directory, (typically /usr/local/soss/java_api/) to the soss_lib subdirectory of the ScaleOut hServer source location.
  • Copy the third party library JARs from the Linux installation Java API lib directories, (typically /usr/local/soss/java_api/lib and /usr/local/soss/java_api/hslib) to the third_party_lib subdirectory of the ScaleOut hServer source location.
  • Download and copy the third party Hadoop distribution library JARs from our website under Optional Downloads to the appropriate subdirectory of the third_party_lib subdirectory of the ScaleOut hServer source location. For example, download the Apache Hadoop 2.4.1 JARs and copy them to the "/path/to/source/third_party_lib/hadoop-2.4.1/" directory.

Windows

  • Clone the ScaleOut hServer project to a local directory.
  • Copy the SOSS Java Named Cache JAR (soss-jnc-5.3.jar) from the Windows installation Java API directory, (typically C:\Program Files\ScaleOut_Software\StateServer\JavaAPI) to the soss_lib subdirectory of the ScaleOut hServer source location.
  • Copy the third party library JARs from the Windows installation Java API lib directory, (typically C:\Program Files\ScaleOut_Software\StateServer\JavaAPI\lib and C:\Program Files\ScaleOut_Software\StateServer\JavaAPI\hslib) to the third_party_lib subdirectory of the ScaleOut hServer source location.
  • Download and copy the third party Hadoop distribution library JARs from our website under Optional Downloads to the appropriate subdirectory of the third_party_lib subdirectory of the ScaleOut hServer source location. For example, download the Apache Hadoop 2.4.1 JARs and copy them to the "C:\path\to\source\third_party_lib\hadoop-2.4.1" directory.

To build all ScaleOut hServer JARs from source

Open a terminal or command prompt inside the hServer base directory and run the following command: ant all

All hServer JARs will be located in the build_output directory.

Supported Hadoop Distributions

To build the ScaleOut hServer libraries compiled against a specific supported Hadoop distribution (for example, Apache Hadoop version 2.4.1): ant build-hadoop-2.4.1

The following Hadoop distributions are currently supported:

Distribution Name Build Target Output Subdirectory
Cloudera CDH 4.4.0 build-cdh4.4.0 cdh4.4.0
Cloudera CDH 5.0.2 build-cdh5.0.2 cdh5.0.2
Cloudera CDH 5.0.2 (with YARN support) build-cdh5.0.2-yarn cdh5.0.2-yarn
Cloudera CDH 5.2.1 build-cdh5.2.1 cdh5.2.1
Cloudera CDH 5.2.1 (with YARN support) build-cdh5.2.1-yarn cdh5.2.1-yarn
Apache Hadoop 1.2.1 build-hadoop-1.2.1 hadoop-1.2.1
Apache Hadoop 2.4.1 build-hadoop-2.4.1 hadoop-2.4.1
Hortonworks Data Platform 2.1 (with YARN support) build-hdp2.1-yarn hdp2.1-yarn
Hortonworks Data Platform 2.2 (with YARN support) build-hdp2.2-yarn hdp2.2-yarn
IBM BigInsights 3.0.0 build-ibm-bi-3.0.0 ibm-bi-3.0.0

Getting Help

Need help building or deploying ScaleOut hServer? Post your questions on our forum.

About

ScaleOut hServer(R) is an integrated in-memory data grid and computation engine that executes standard Hadoop MapReduce code in-parallel, in-memory. Please visit our website for more information.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages