Ansible: IT automation for everybody

Ansible is an open source tool for software provisioning, application deployment, orchestration, configuration, and administration. Its purpose is to help you automate your configuration processes and simplify the administration of multiple systems. Thus, Ansible essentially pursues the same goals as Puppet, Chef, or Saltstack.

What I like about Ansible is that it’s flexible, lean, and easy to start with. In most use cases, it keeps the job simple.

I chose to use Ansible back in 2016 because no agent has to be installed on the managed nodes—a node is what Ansible calls a managed remote system. All you need to start managing a remote system with Ansible is SSH access to the system, and Python installed on it. Python is preinstalled on most Linux systems, and I was already used to managing my hosts via SSH, so I was ready to start right away. And if the day comes where I decide not to use Ansible anymore, I just have to delete my Ansible controller machine (control node) and I’m good to go. There are no agents left on the managed nodes that have to be removed.

Ansible offers two ways to control your nodes. The first one uses playbooks. These are simple ASCII files written in Yet Another Markup Language (YAML), which is easy to read and write. And second, there are the ad-hoc commands, which allow you to run a command or module without having to create a playbook first.

You organize the hosts you would like to manage and control in an inventory file, which offers flexible format options. For example, this could be an INI-like file that looks like:

mail.example.com

[webservers]
foo.example.com
bar.example.com

[dbservers]
one.example.com
two.example.com
three.example.com

[site1:children]
webservers
dbservers

Examples

I would like to give you two small examples of how to use Ansible. I started with these really simple tasks before I used Ansible to take control of more complex tasks in my infrastructure.

Ad-hoc: Check if Ansible can remote manage a system

As you might recall from the beginning of this article, all you need to manage a remote host is SSH access to it, and a working Python interpreter on it. To check if these requirements are fulfilled, run the following ad-hoc command against a host from your inventory:

[jkastning@ansible]$ ansible mail.example.com -m ping
mail.example.com | SUCCESS => {
    "changed": false, 
    "ping": "pong"
}

Playbook: Register a system and attach a subscription

This example shows how to use a playbook to keep installed packages up to date. The playbook is an ASCII text file which looks like this:

---
# Make sure all packages are up to date
- name: Update your system
  hosts: mail.example.com
  tasks:
  - name: Make sure all packages are up to date
    yum:
      name: "*"
      state: latest

Now, we are ready to run the playbook:

[jkastning@ansible]$ ansible-playbook yum_update.yml 

PLAY [Update your system] **************************************************************************

TASK [Gathering Facts] *****************************************************************************
ok: [mail.example.com]

TASK [Make sure all packages are up to date] *******************************************************
ok: [mail.example.com]

PLAY RECAP *****************************************************************************************
mail.example.com : ok=2    changed=0    unreachable=0    failed=0

Here everything is ok and there is nothing else to do. All installed packages are already the latest version. 

It’s simple: Try and use it

The examples above are quite simple and should only give you a first impression. But, from the start, it did not take me long to use Ansible for more complex tasks like the Poor Man's RHEL Mirror or the Ansible Role for RHEL Patchmanagment.

Today, Ansible saves me a lot of time and supports my day-to-day work tasks quite well. So what are you waiting for? Try it, use it, and feel a bit more comfortable at work.