ansible-hortonworks-configuration

Motivation

The idea is to clone ansible-hortonworks from GitHub and change the variables without actually changing any of the files cloned.

Introduction

The repository shows how to override variables in ansible-hortonworks/inventory/aws/group_vars/all and ansible-hortonworks/playbooks/group_vars/all using parameter extra-vars from the command line when running playbooks.

If ran out-of-the-box, the cluster - one namenode with Ambari and one datanode - is build in region Ireland (eu-west-1) with centos7 and the instance types are t2.xlarge.

secret key and access key

In /etc/profile.d/sh.local add secret key and access key export AWS_SECRET_ACCESS_KEY= export AWS_ACCESS_KEY_ID= (or some other way)

Clone ansible-hortonworks repository

git-clone folder has a playbook that clones https://github.com/hortonworks/ansible-hortonworks.git

Files in config folder:

• aws-cluster.yml - in eu-west-1, 2 instances are created of type t2.xlarge. Ami is centos 7.
• hdp-cluster-hadoop3.yml - creates an HDP 3.0 cluster based on the variables in this file
• hdp-cluster-minimal.yml - creates an HDP 2.6.5 (default by HDP in GitHub) with minimal services

Example of how to run

. run-playbooks-with-extra-vars.sh config/hdp-cluster-hadoop3.yml

Output

All hortonworks playbooks are ran with verbose -v and output directory is created at the start of the script. The name of the directory matches the name of the yml file used in extra-vars for HDP variables.