Installation

Dependencies

• The runner and web server are intended to run on a Hadoop edge node. However, they should work on nearly any modern Linux distribution, and can also run on Mac OS using Homebrew. This is because all of the Hadoop dependencies are currently in the checks only.
• The runner and checks require Python 2.7
• The existing checks have all been written and tested against Impala 2.3. They can be easily modified to run against Hive.

Installing the Check Runner:

• $ pip install hadoopinspector

Installing the Web Server:

• $ git clone https://github.com/willzfarmer/HadoopInspector.git

Installing the Checks:

• $ git clone https://github.com/willzfarmer/hadoopinspector_plugins.git

Build a check directory:

• ex: $ mkdir /hapinsp/checks
• Use the checks from hadoopinspector_plugins, and or write your own. But they must be located within the check directory - that you provide to the runner as an argument.

Build a log directory:

• ex: $ mkdir /hapinsp/logs

Create registry config files for each instance & database to be checked:

• See the wiki page on the registry for more info

Run the checker

hadoopinspector_runner.py --instance [your inst name] 
                      --database [your db name] 
                      --registry-filename /hapinsp/registry.json 
                      --results-filename /hapinsp/results.sqlite 
                      --check-dir /hapinsp/checks 
                      --log-dir /hapinsp/logs 
                      --report >> /hapinsp/logs/run.log

Run web server

server/hadinsp_httpserver.py