Easy installation of Spark under Ubuntu 14.04

Java installation first needs to download on Oracle's website

Create a JVM folder in the Lib directory

 
   
  

  
    
   
sudo mkdir /usr/lib/jvm
 
   
  

Then unzip the file into this folder

 
   
  

  
    
   
sudo tar zxvf jdk-8u40-linux-i586.tar.gz -C /usr/lib/jvm
 
   
  

Go to Unzip folder

 
   
  

  
    
   
cd /usr/lib/jvm
 
   
  

And then change a name for convenience.
sudo mv jdk1.8.0_40 Java

Open configuration file

 
   
  

  
    
   
sudo gedit ~/.bashrc
 
   
  

Add the following settings

  
 

   
  

  
export JAVA_HOME=/usr/lib/jvm/java 
   
  

  
export JRE_HOME=${JAVA_HOME}/jre  
   
  

  
    export CLASSPATH=.:${JAVA_HOME}/lib:${JRE_HOME}/lib  
   
  

  
    export PATH=${JAVA_HOME}/bin:$PATH 
  
 

To make it effective:

 
   
  

  
    
   
source ~/.bashrc 
 
   
  

Complete

Install spark download Prebuild's Spark installation package

Extract

 
   
  

  
    
   
tar -xzf spark-1.2.0-bin-hadoop2.4.tgz
 
   
  

Then, in fact, this time Spark's Python mode is ready to use. Just CD in and run the Pyspark in the bin directory. But the next thing we do is link and choose a better location to store the files

Link

Put the file in a different place for the first half of the year.

 
   
  

  
    
   
sudo mv spark-1.2.1-bin-hadoop2.4 /srv/
 
   
  

Then set up a link

 
   
  

  
    
   
sudo ln -s /srv/spark-1.2.1-bin-hadoop2.4/ /srv/spark
 
   
  

You can change the ~/.bash_profile.

 
   
  

  
    
   
export SPARK_HOME=/srv/spark
  
    
   
export PATH=$SPARK_HOME/bin:$PATH
 
   
  

Complete

The command line executes the Pyspark directly

From for notes (Wiz)

Easy installation of Spark under Ubuntu 14.04