Setup Apache Spark/ Jupyter Notebook on MacOS

notes.md

Spark Setup MacOS Guide

Reference: https://spark.apache.org/docs/latest/

Overview:

• Install Java 8+, then add to PATH and JAVA_HOME

1. Install Java.

• Download and install Java 8 through brew.

brew cask install java8

• Validate Java version

brew cask info java8

java8: 1.8.0_192-b12,750e1c8617c5452694857ad95c3ee230
https://www.oracle.com/technetwork/java/javase/downloads/jdk8-downloads-2133151.html
/usr/local/Caskroom/java8/1.8.0_192-b12,750e1c8617c5452694857ad95c3ee230 (246.2MB)
From: https://github.com/Homebrew/homebrew-cask-versions/blob/master/Casks/java8.rb
==> Name
Java Standard Edition Development Kit
==> Artifacts
JDK 8 Update 192.pkg (Pkg)
...

• Add java environment variables to terminal

  • Open Vim

  # Depending on terminal one might be using
  vim ~/.zshrc

  • Add the following to .zshrc

  # For Apache Spark
  if which java > /dev/null; then export JAVA_HOME=$(/usr/libexec/java_home); fi

2. Install Apache Spark

Brew is Mac OS Package Manager, similar to apt (http://brew.sh/)

brew update 
brew install scala
brew install apache-spark

3. Setup Variables

• Assump Current spark version == 2.4.0.

# For a ipython notebook and pyspark integration
if which pyspark > /dev/null; then
  export SPARK_HOME="/usr/local/Cellar/apache-spark/2.4.0/libexec/"
  export PYTHONPATH=$SPARK_HOME/python:$SPARK_HOME/python/build:$PYTHONPATH
  export PYTHONPATH=$SPARK_HOME/python/lib/py4j-0.10.4-src.zip:$PYTHONPATH
fi

• Up to this point, you should be able to launch pyspark and scala-shell from terminal.

4. Integrate Spark and Jupyter Notebook

• Install Python Env through pyenv, a python versioning manager.

pyenv install 3.6.7 

# Set Python 3.6.7 as main python interpreter
pyenv global 3.6.7

# Update new python
source ~/.zshrc

# Update pip from 10.01 to 18.1
pip install --upgrade pip

• (Optional) If you received pyspark 2.4.0 requires py4j==0.10.7, which is not installed., fixed by:

pip install py4j==0.10.7

• Install Jupyter, Apache Toree

pip install jupyter toree

• Create a kernel in Jupyter for managing Spark

jupyter kernelspec list
# Available kernels:
#  python3   /Users/dat/.pyenv/versions/3.6.7/share/jupyter/kernels/python3

jupyter toree install --replace --spark_home=$SPARK_HOME
jupyter kernelspec list

# Available kernels:
#  apache_toree_scala    /Users/dat/Library/Jupyter/kernels/apache_toree_scala
#  python3               /Users/dat/.pyenv/versions/3.6.7/share/jupyter/kernels/python3

Launch Jupyter Notebook and Test Our First Spark Application

jupyter notebook

• Remember to Select Toree as your main kernel