Book Image

Learning Ansible 2 - Second Edition

Book Image

Learning Ansible 2 - Second Edition

Overview of this book

Ansible is an open source automation platform that assists organizations with tasks such as configuration management, application deployment, orchestration, and task automation. With Ansible, even complex tasks can be handled easier than before. In this book, you will learn about the fundamentals and practical aspects of Ansible 2 by diving deeply into topics such as installation (Linux, BSD, and Windows Support), playbooks, modules, various testing strategies, provisioning, deployment, and orchestration. In this book, you will get accustomed with the new features of Ansible 2 such as cleaner architecture, task blocks, playbook parsing, new execution strategy plugins, and modules. You will also learn how to integrate Ansible with cloud platforms such as AWS. The book ends with the enterprise versions of Ansible, Ansible Tower and Ansible Galaxy, where you will learn to interact Ansible with different OSes to speed up your work to previously unseen levels By the end of the book, you’ll able to leverage the Ansible parameters to create expeditious tasks for your organization by implementing the Ansible 2 techniques and paradigms.
Table of Contents (16 chapters)
Learning Ansible 2 Second Edition
Credits
About the Author
About the Reviewer
www.PacktPub.com
Preface

Deployment strategies


We have seen how to distribute software in your environment, let's now speak about deployment strategies; that is how to upgrade your application without your service suffering from it.

There are three different problems you might incur during an update:

  • Downtime during the update rollout

  • The new version has problems

  • The new version seems to work, until it fails

The first problem is known to every system administrator. During the update, you are probably going to restart some services, and for the time between the stop and the start of the service, your application will not be available on that machine. To avoid this also means that your application is not available at all; you will need to have at least some machines with the application available and a smart load balancer in front that will remove (and add them back when is the case) all nonfunctioning nodes.

The second problem can be prevented in multiple ways. The cleanest one would be testing in the CI/CD pipeline. In...