About

This is a guide on local installation of Apache-zeppelin and configuration of its Hbase interpreter on MacOS (Version 10.13.1).

Versions

• Java 1.8.0_172
• Hadoop 2.7.6
• Hbase 1.2.0
• Apache-zeppelin 0.7.3

Install Homebrew

1. Copy and paste this command into Terminal:

   $ /usr/bin/ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)"
   

Install Java 8

1. Copy and paste these commands into Terminal (Zeppelin 0.7.3 only supports Java <= 8):

   brew tap caskroom/versions
   brew cask install java8
   export JAVA_HOME=`/usr/libexec/java_home -v 1.8`
   

Install Hadoop

1. Download a stable release from Apache Hadoop Releases.

2. Unpack it on your local file system, e.g. /usr/local/...

3. Put Hadoop Binary on the shell path:

   export HADOOP_HOME=/usr/local/hadoop-x.y.z
   export PATH=$PATH:$HADOOP_HOME/bin:$HADOOP_HOME/sbin
   

4. Use "hadoop version" to verify hadoop has been installed

   hadoop version
   

Install Hbase

1. Download a stable release from Apache Download Mirror.

2. Unpack it on your local file system, e.g.

   /usr/local/...
   

3. Specifying JAVA_HOME variable in Hbase's conf/hbase-env.sh file. e.g.

   export JAVA_HOME=/Library/Java/JavaVirtualMachines/jdk1.7.0_25.jdk/Contents/Home
   

4. Run Hbase Shell.

   cd /usr/local/hbase
   ./bin/start-hbase.sh
   cd bin
   ./hbase shell 
   

5. Run simple instructions such as creating a table to see if it works.

   create 'table', 'col1', 'col2'
   

Install Apache-zeppelin

1. Download the binary package with all interpreters from Apache Zeppelin

2. Unpack it on your local file system, e.g. /usr/local/...

3. Start zeppelin

   cd /usr/local/zeppelin
   bin/zeppelin-daemon.sh start
   

4. Open an browser and type:

   http://localhost:8080
   

   It should direct you to the zeppelin web interface

Configure Apache-zeppelin

1. Go to Interpreters and find the configurations of Hbase Interpreter

2. In properties, change hbase.home to be the home directory of your hbase. e.g.

   /usr/local/hbase-1.2.0/
   

3. In properties, change hbase.ruby.sources to be the relative path of the ruby files

   lib/ruby
   

4. Add the following dependenceis into the Hbase Interpreter

   /usr/local/hbase-1.2.0/lib/hbase-client-1.2.0.jar
   /usr/local/hbase-1.2.0/lib/hbase-protocol-1.2.0.jar
   /usr/local/hbase-1.2.0/lib/hbase-common-1.2.0.jar
   

5. Your Hbase Interpreter should look like this:

6. Create a Notebook with Hbase interpreter and see if the following command works:

   create 'table', 'col1', 'col2'
   

References

1. How To Locally Install & Configure Apache Spark & Zeppelin
2. Hadoop: The Definitive Guide
3. ADDING HBASE INTERPRETER TO ZEPPELIN
4. HBase - Installation
5. Apache Zeppelin 0.7.3 - http error 503 in browser