MongoDB: Hadoop Integerateion 2-优快云博客

本文详细介绍了如何在现有的Hadoop集群上部署并运行MongoDB Hadoop示例，包括环境准备、下载所需依赖、配置集群、安装Hadoop和Hive等步骤，并提供了具体命令和操作细节。

Prepare ENV

download mongo-java-driver from http://central.maven.org/maven2/org/mongodb/mongo-java-driver

compile mongo-hadoop-connector for hadoop2.3.0

alter build.gradle to change hadoop-version to 2.3 and delete related download dependencies task

#./gradlew jar

distribute the aboves jars to hadoop clustert's nodes

#cp core/build/libs/mongo-hadoop-core-1.2.1-SNAPSHOT-hadoop_2.3.jar hadoop-2.3.0/share/hadoop/common/lib

#cp mongo-java-driver-2.12.2.jar hadoop-2.3.0/share/hadoop/common/lib

or add the above two jars into project lib dir and add them to build path

Note: the destination dir is not hadoop-2.3.0/lib

Run MongoDB Hadoop examples

#cd /path/to/mongodb-hadoop

#./gradlew historicalYield

The above command will download hadoop and install it. But,I want to run the example on my existed hadoop cluster.

To see what happens for this command, first find the files which execute this commands

#find . |xargs grep 'historicalYield' -sl

task historicalYield(dependsOn: 'configureCluster') << {
    exec() {
        commandLine "mongoimport", "-d", "mongo_hadoop", "-c", "yield_historical.in", "--drop",
                    "examples/treasury_yield/src/main/resources/yield_historical_in.json"
    }

    hadoop("examples/treasury_yield/build/libs/treasury_yield-${project(':core').version}-hadoop_${hadoop_version}.jar",
           "com.mongodb.hadoop.examples.treasury.TreasuryYieldXMLConfig", [
                "mongo.input.uri=mongodb://localhost:27017/mongo_hadoop.yield_historical.in",
                "mongo.output.uri=mongodb://localhost:27017/mongo_hadoop.yield_historical.out"
           ])
}

task configureCluster(dependsOn: ['copyFiles']) << {
}

task copyFiles(dependsOn: ['installHadoop', 'installHive']) << {
    def hadoopEtc
    def hadoopLib
    if (hadoop_version.startsWith("1")) {
        hadoopLib = "${hadoopHome}/lib"
        hadoopEtc = "${hadoopHome}/conf"
    } else {
        hadoopLib = "${hadoopHome}/share/hadoop/common"
        hadoopEtc = "${hadoopHome}/etc/hadoop"
    }

    println "Updating mongo jars"
    copy {
        from "core/build/libs/mongo-hadoop-core-${project(':core').version}-hadoop_${hadoop_version}.jar"
        into hadoopLib
        rename { "mongo-hadoop-core.jar" }
    }
    copy {
        from "hive/build/libs/mongo-hadoop-hive-${project(':core').version}-hadoop_${hadoop_version}.jar"
        into hiveHome + '/lib'
        rename { "mongo-hadoop-hive.jar" }
    }
    download {
        src "http://central.maven.org/maven2/org/mongodb/mongo-java-driver/${javaDriverVersion}/mongo-java-driver-${javaDriverVersion}.jar"
        dest "${hadoopLib}/mongo-java-driver.jar"
        onlyIfNewer true
    }
    println "Updating cluster configuration"
    copy {
        from 'clusterConfigs'
        into hadoopEtc
    }
}

def hadoop(jar, className, args) {
    def line = ["${hadoopHome}/bin/hadoop",
                "jar", jar, className,
                //Split settings
                "-Dmongo.input.split_size=8",
                "-Dmongo.job.verbose=true",
    ]
    args.each {
        line << "-D${it}"
    }
    println "Executing hadoop job:\n ${line.join(' \\\n\t')}"
    def hadoopEnv = [:]
    if (hadoop_version.startsWith("cdh")) {
        hadoopEnv.MAPRED_DIR = 'share/hadoop/mapreduce2'
    }
    exec() {
        environment << hadoopEnv
        commandLine line
    }
}

----------------------------

so,understood the details behind the screen. We can run the example by hand

1. load sample data into mongoDB

#mongoimport -d mongo_hadoop -c yield_historical.in --drop <examples/treasury_yield/src/main/resources/yield_historical_in.json

2.run example

#cd mongo-hadoop/examples/treasury_yield/build/libs

#hadoop jar treasury_yield-1.2.1-SNAPSHOT-hadoop_2.3.jar com.mongodb.hadoop.examples.treasury.TreasuryYieldXMLConfigV2 -Dmongo.input.uri=mongodb://localhost:27017/mongo_hadoop.yield_historical.in -Dmongo.output.uri=mongodb://localhost:27017/mongo_hadoop.yield_historical.out -Dmongo.input.split_size=8 -Dmongo.job.verbose=true

References

http://docs.mongodb.org/ecosystem/tutorial/getting-started-with-java-driver/

https://github.com/mongodb/mongo-hadoop/blob/master/CONFIG.md

https://github.com/mongodb/mongo-hadoop/blob/master/examples/README.md

http://docs.mongodb.org/ecosystem/tutorial/getting-started-with-hadoop/

http://mongodb-documentation.readthedocs.org/en/latest/ecosystem/tutorial/getting-started-with-hadoop.html

http://www.mongodb.com/press/integration-hadoop-and-mongodb-big-data%E2%80%99s-two-most-popular-technologies-gets-significant

http://blog.mortardata.com/post/43080668046/mongodb-hadoop-why-how

http://help.mortardata.com/data_apps/mongo_hadoop

http://www.severalnines.com/blog/big-data-integration-etl-clickstream-mongodb-hadoop-analytics