Skip to content

Latest commit

 

History

History
228 lines (154 loc) · 7.28 KB

README.md

File metadata and controls

228 lines (154 loc) · 7.28 KB
Raw

PySpark Installation Instruction for Ubuntu 14.04,16.04

The scripts mentioned bellow will work for most of the versions of ubuntu, however we have tested it for only 14.04 and 16.04. F Pyspark installation requires you to setup

  1. Installing java properly

  2. Hadoop on single node

  3. Scala

  4. Spark

  5. Python

Java Setup

Step 1 : Installing Java

sudo apt-add-repository ppa:webupd8team/java
sudo apt-get install oracle-java7-installer

Step 2: Set the environment Variable

Put the bellow code at the end of .bashrc file. And inside the .profile file. It is located at '~' or 'home' location. Also for ubuntu 16.04 users, you should carefully append this line int /etc/environment file. Please take the backup of this /etc/environment file before making any changes.

JAVA_HOME=/usr/lib/jvm/java-7-oracle

Add the path of java's bin folder in PATH variable. First try to add the following code at the end of .bashrc

PATH=$JAVA_HOME/bin:$PATH
source ~/.bashrc

If it doesn't work, try updating PATH and JAVA_HOME in profilerc. For python, it seems that 16.04 has issues updating environment variable, therefore it is better that you edit /etc/environment file and add updated path in the location. Again, don't forget to take backup of this file, i.e /etc/environment

Do not use PATH=ABC:$PATH kind of format in /etc/environment, as .it can be fatal for your system. It will consider $ literally. Update PATH as follow in /etc/environment.

PATH=/usr/lib/jvm/java-7-oracle/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games

Now restart the system.

In case there are any reboot issues, open the machine without lightdm or display managers. Via cntl+alt+Fn+F2. Now revert the changes you did in /etc/environment via backup of the same file. And then type

sudo reboot

Please ping me at [email protected] , in case you have trouble setting up the environment variable.

Note : Do not install oracle-java-9 for any reasons, because it seems to be incompatible with latest scala version available

Hadoop Setup (Limited Use)

Step 1 : Download and unzip hadoop-x.x.x.tar.gz

mkdir ~/temp
cd ~/temp
wget https://dist.apache.org/repos/dist/release/hadoop/common/hadoop-2.7.2/hadoop-2.7.2.tar.gz
tar -xvzf hadoop-2.7.2.tar.gz
sudo mv hadoop-2.7.2 /usr/local/hadoop

Step 2 : Give permissions to hadoop. Replace <<username>> in bellow script with your computer name and execute

sudo chown -R <<username>>:<<username>> /usr/local/hadoop

Step 3 : Add an environment variable.

Follow steps similar to adding JAVA_HOME, just that instead of JAVA_HOME, you will have to add HADOOP_HOME. Also update the .bashrc and .profile with the following exports which also includes HADOOP HOME.

export HADOOP_HOME=/usr/local/hadoop
export HADOOP_OPTS="-Djava.library.path=$HADOOP_INSTALL/lib"
export HADOOP_MAPRED_HOME=$HADOOP_HOME
export HADOOP_COMMON_HOME=$HADOOP_HOME
export HADOOP_HDFS_HOME=$HADOOP_HOME
export YARN_HOME=$HADOOP_HOME
export HADOOP_COMMON_LIB_NATIVE_DIR=$HADOOP_HOME/lib/native
export PATH=$PATH:$HADOOP_HOME/sbin:$HADOOP_HOME/bin
export HADOOP_INSTALL=$HADOOP_HOME

The path environment variable needs to be updated in .bashrc,.profile as bellow :

PATH=$HADOOP_HOME/bin:$HADOOP_HOME/sbin:$PATH

Never update the PATH env variable in /etc/environment directly the way you modify other variables. Rather add the expanded path in the exisiting PATH env variable declaration as bellow

PATH=/usr/local/hadoop/bin:/usr/lib/jvm/java-7-oracle/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games

and finally reboot the system or source the ~/.bashrc and ~/.profile file, the get the changes reflected.

type 'hadoop', to see if the is returning any errors or not.

Updating *.conf files in Hadoop

Simplest way is to run the following code as shown bellow :

sudo mkdir -p /app/hadoop/tmp
sudo mkdir -p /app/hadoop/datanode
sudo mkdir -p /app/hadoop/namenode
sudo chown <<username>>:<<username>> /app/hadoop
cd ~/temp
wget -O hadoop-config.zip http://gdurl.com/Sw19/_/b5b85eb0cf49df586d54e4d060652f6b 
# can also download from https://drive.google.com/file/d/0B7bnnRB2aYVZWlNMbUpXd25fa1E/view?usp=sharing
#Be Careful while performing rm operations
rm -r /usr/local/hadoop/etc/hadoop 
unzip hadoop-config.zip 
mv hadoop /usr/local/hadoop/etc/hadoop
hadoop namenode -format
start-all.sh

Refer to bellow two blogs for further clarity :

http://www.michael-noll.com/tutorials/running-hadoop-on-ubuntu-linux-single-node-cluster/

http://www.tutorialspoint.com/hadoop/hadoop_enviornment_setup.htm

Scala Setup (Limited Use)

Step 1 : Download and unzip scala-x.x.x.tgz

cd ~/temp
wget http://downloads.lightbend.com/scala/2.11.8/scala-2.11.8.tgz
tar -xvzf scala-2.11.8.tgz scala-2.11.8
sudo mv scala-2.11.8 /usr/local/scala

Step 2 : Give permissions to scala. Replace <<username>> in bellow script with your computer name and execute

sudo chown -R <<username>>:<<username>> /usr/local/scala

Step 3 : Add an environment variable.

Follow steps similar to adding JAVA_HOME, just that instead of JAVA_HOME, you will have to add SCALA_HOME

SCALA_HOME=/usr/local/scala

and path needs to be updated in .bashrc,.profile as bellow :

PATH=$SCALA_HOME/bin:$PATH

Never update the PATH env variable in /etc/environment directly the way you modify other variables. Rather add the expanded path in the exisiting PATH env variable declaration as bellow

PATH=/usr/local/scala/bin:/usr/lib/jvm/java-7-oracle/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games

and finally reboot the system or source the ~/.bashrc and ~/.profile file, the get the changes reflected.

When you type 'scala', a command line prompt should come.

Spark Setup (Limited Use)

Step 1 : Download and unzip spark-x.x.x-bin-hadoopx.x.tgz

mkdir ~/temp
cd ~/temp
wget http://d3kbcqa49mib13.cloudfront.net/spark-2.0.0-bin-hadoop2.7.tgz
sudo mv spark-2.0.0-bin-hadoop2.7.tgz /usr/local/spark

Step 2 : Give permissions to spark. Replace <<username>> in bellow script with your computer name and execute

sudo chown -R <<username>>:<<username>> /usr/local/spark

Step 3 : Add an environment variable.

Follow steps similar to adding JAVA_HOME, just that instead of JAVA_HOME, you will have to add SPARK_HOME

SPARK_HOME=/usr/local/spark

and path needs to be updated in .bashrc,.profile as bellow :

PATH=$SPARK_HOME/bin:$PATH

Never update the PATH env variable in /etc/environment directly the way you modify other variables. Rather add the expanded path in the exisiting PATH env variable declaration as bellow

PATH=/usr/local/spark/bin:/usr/lib/jvm/java-7-oracle/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games

and finally reboot the system or source the ~/.bashrc and ~/.profile file, the get the changes reflected.

Now type spark-shell, to see if you get the required scala command prompt, with spark-context getting setup. You should also be able to get python open in spark context by typing pyspark.