Ex-05-Pseudo-Node-Configuration-for-Hadoop-on-Ubuntu

AIM

To implement Pseudo Node configuration for Hadoop on ubuntu

Pre-requisites

a) jdk

Single-Node Configuration

Create a dedicated user account for hadoop

    $sudo addgroup hadoop

    $sudo adduser --ingroup hadoop hduser

    $sudo usermod -a -G sudo hduser

    $su - hduser

Install java1.8 in folder /usr/local

 $sudo chmod 777 /usr/local

 $cd /usr/local

 $sudo tar xvzf $HOME/Downloads/jdk1.8.tar.gz

Install Hadoop

    $cd /usr/local

    $tar xvzf $HOME/Downloads/hadoop-2.5.1.tar.gz

    $sudo chmod 777 hadoop-2.5.1

Set the hadoop environment variables: Include the following lines in the $HOME/.bashrc file #Set Hadoop-related environment variables

export HADOOP_HOME=/usr/local/hadoop-2.5.1

#Set JAVA home directory

    export JAVA_HOME=/usr/local/jdk1.8.0_31

#Add Hadoop bin/ directory to PATH

    export PATH=$PATH:$HADOOP_HOME/bin

Set hadoop environment variables: Include the following lines /etc/profile file #--insert JAVA_HOME

    JAVA_HOME=/usr/local/jdk1.8.0_31

#--insert HADOOP_PREFIX

    HADOOP_PREFIX=/usr/local/hadoop-2.5.1

#--in PATH variable just append at the end of the line

    PATH=$PATH: $JAVA_HOME/bin:$HADOOP_PREFIX/bin

#--Append HADOOP_PREFIX at end of the export statement export

    PATH JAVA_HOME HADOOP_PREFIX

Run the.bashrc & profile files from the $ prompt for updating the changes

    $ source $HOME/.bashrc

    $ source /etc/profile

Verify java & hadoop installation using

    $ java -version
    $ echo $HADOOP_PREFIX
    $ cd $HADOOP_PREFIX

    $ bin/hadoop version

Configuration of the hadoop files: hadoop-env.sh, core-site.xml, mapred-site.xml, hdfs- site.xml and yarn-site.xml

   path ::	/usr/local/hadoop-2.5.1/etc/hadoop

a) hadoop-env.sh Include the following lines in hadoop-env.sh file

    export JAVA_HOME=/usr/local/jdk1.8.0_31
    export HADOOP_PREFIX=/usr/local/hadoop-2.5.1

b) core-site.xml Configure the directory for Hadoop to store its data files, the network ports it listens to, etc. Setup will use Hadoop’s Distributed File System (HDFS-single local machine)

    $ mkdir -p /app/hadoop/tmp
    $ chown hduser:hadoop /app/hadoop/tmp

Include the following lines in core-site.xml file between and tags

    <property>
            <name>hadoop.tmp.dir</name>
            <value>/app/hadoop/tmp</value>
    </property>
    <property>
            <name>fs.default.name</name>
            <value>hdfs://localhost:9000</value>
    </property>

c) mapred-site.xml

     $sudo cp mapred-site.xml.template mapred-site.xml

Include the following lines in mapred-site.xml file

    <property>
            <name>mapreduce.framework.name</name>
            <value>yarn</value>
    </property>

d) hdfs-site.xml Include the following lines in hdfs-site.xml file

    <property>
            <name>dfs.replication</name>
            <value>1</value>
    </property>

e) yarn-site.xml Include the following lines in yarn-site.xml file

    <property>
            <name>yarn.nodemanager.aux-services</name>
            <value>mapreduce_shuffle</value>
    </property>

Format the Hadoop File system implemented on top of the local file system using

    $/usr/local/hadoop-2.5.1/bin/hadoop namenode –format

Start Hadoop using

    $sbin/start-all.sh

Explore Hadoop using http://localhost:50070/ from the browser

The commonly used HDFS Commands are as follows:

### mkdir :
##### Creates a directory in the given path:
    syntax: bin/hdfs dfs -mkdir <paths>

### ls
##### Lists the files in a given path
    syntax:bin/hdfs dfs -ls <args>

### cp
##### Copies files from source to destination. This command allows multiple sources as well in which case the destination must be a directory.
    syntax:bin/hdfs dfs -cp <source> <dest>

Create a directory ‘/input’ in HDFS

    $bin/hdfs dfs -mkdir /input

Copy the input files into the distributed file system

    $cd $HOME/Downloads/

    $ tar xvzf $HOME/Downloads/mrsampledata.tar.gz

Run some of the examples provided

    $ bin/hadoop jar share/hadoop/mapreduce/hadoop-mapreduce- examples-2.5.1.jar grep /input /output '(CSE)'

Examine the output files


    $bin/hdfs dfs -put $HOME/Downloads/mrsampledata/* /input

Copy the output files from the distributed file system to the local file system and examine them:

    $ bin/hdfs dfs -get output output
    
    $ cat output/*

or View the output files on the distributed file system

    $ bin/hdfs dfs -cat /output/*

Result:

Thus, the implementation of Pseudo Node configuration for Hadoop on ubuntu is successfully executed.

ponnagarajanmurugan / ex-05-pseudo-node-configuration-for-hadoop-on-ubuntu Goto Github PK