Automated patching approaches using Ansible

Patching and updating is a task that everyone who has to manage production systems has to deal with. There are two approaches that we will look are as follows:

• Rolling updates
• BlueGreen deployments

Rolling updates

Imagine that we have five web servers behind a load balancer. What we would like to do is a zero downtime upgrade of our web application. Using certain keywords available in Ansible, we can make this happen. 

In our example, we want to achieve the following: 

• Tell the load balancer that web server node is down
• Bring down the web server on that node
• Copy the updated application files to that node
• Bring up the web server on that node

The first keyword for us to look at is serial. Let's see this example from Ansible documentation:

- name: test play
  hosts: webservers
  serial: 1

The example is from http://docs.ansible.com/ansible/latest/playbooks_delegation.html#rolling-update-batch-size.

This ensures that the execution of the playbook is done serially...