This is a processor for heritrix that writes fetched pages to hbase.

Introduction

The layout of this contribution is modeled after Doug Judd's heritrix-hadoop-dfs-processor available off the heritrix home page.

This software is licensed under the LGPL. See accompanying LICENSE.txt document.

The hbase-writer is an extension to the Heritrix open source crawler written by the Internet Archive (http://crawler.archive.org/) that enables it to store crawled content directly into HBase tables (http://hbase.org/) running on the Hadoop Distributed FileSystem (http://lucene.apache.org/hadoop/). HBase-Writer writes crawled content into a given hbase table as records. In turn, these tables are directly supported by the Map/Reduce framework via HBase so Map/Reduce jobs can be done on them.

The current version of hbase-writer assumes version 2.0.x of Heritrix, version 0.19.x of Hadoop/HBase and Java 1.6. Newer versions of Hadoop and Heritrix may continue to work with this connector as long as the pertinent APIs have not changed. Just replace the jar files with the newer versions.

SETUP

• Start an instance of hbase.
• Install heritrix-2.0.x
• Copy the following jar files into the ${HERITRIX_HOME}/lib directory:

hbase-writer-x.x.x.jar

hbase-x.x.x.jar

zookeeper-x.x.x.jar

hadoop-x.x.x-core.jar

log4j-x.x.x.jar

• Start Heritrix

CONFIGURING HERITRIX

On the "Settings for sheet 'global'":

• Click on a 'details' link and add new processor com.powerset.heritrix.writer.HBaseWriterProcessor.
• Return to the global sheet and set this as your writer (the page will not draw completely if did not type in the name of the processor properly -- see heritrix_out.log for errors).
• Set at least the master and table configuration for HBaseWriterProcessor.

zkquorum

The zookeeper quroum that serves the hbase master address. Since hbase-0.20.0, the master server's address is returned by the zookeeper quorum. So this value is a comma seperated list of the zk quorum. i.e.: zkHost1,zkHost2,zkHost3

table

Which table in HBase to write the crawl to. This table will be created automatically if it doesnt exist. i.e.: Webtable

COMPILING THE SOURCE

The source is built using Maven2.

mvn clean compile

BUILDING THE JAR

mvn clean package

The hbase-writer-x.x.x.jar should be in the target/ directory.

OBTAINING THE DEPENDENCY JARS

You can get the hadoop (http://hadoop.apache.org/core/), hbase (http://hbase.org/) and log4j jars from downloading the releases from these 2 sites. You can also get the hadoop, hbase and log4j dependency jars from your ${HOME}/.m2/repository/ directory after you have built the project using maven. For example:

cp ${HOME}/.m2/repository/org/apache/hadoop/hbase/0.20.1/hbase-0.20.1.jar ${HERITRIX_HOME}/lib/ cp ${HOME}/.m2/repository/org/apache/hadoop/zookeeper/3.2.1/zookeeper-3.2.1.jar ${HERITRIX_HOME}/lib/ cp ${HOME}/.m2/repository/org/apache/hadoop/hadoop-core/0.20.1/hadoop-core-0.20.1.jar ${HERITRIX_HOME}/lib/ cp ${HOME}/.m2/repository/log4j/log4j/1.2.15/log4j-1.2.15.jar ${HERITRIX_HOME}/lib/

You can also get them by visiting the hbase-writer archive repository: http://repo1.opensourcemasters.org:8081/nexus/

UPGRADING TO NEW HADOOP/HBASE/HERITRIX VERSIONS

The hbase-writer project will be keeping up to date with the latest hadoop, hbase and heritrix versions, but if you would like to compile and build hbase-writer against different versions, you are able to do so with maven2 without changing any versioned files.

To build hbase-writer with new versions of hadoop, hbase or heritrix (or any of the dependencies), use a ${HOME}/.m2/settings.xml file.

A sample settings.xml file:

<?xml version="1.0" encoding="UTF-8"?>

<settings>

<profiles>

<profile>

<id>

myBuild

</id>

<properties>

<heritrix.version>2.0.2</heritrix.version> <hbase.version>0.20.2</hbase.version> <hadoop.version>0.20.2</hadoop.version> <zookeeper.version>3.2.1</zookeeper.version>

</properties>

</profile>

</profiles>

</settings>

BUILDING THE SITE REPORT

mvn clean site