Deploy and remove a web server with Ansible

​Ansible is an agentless IT automation tool that unravels the mystery of how work gets done. Many sysadmins run Ansible ad hoc commands or playbooks daily to handle their automated tasks. They rely on Ansible's default command-line interface (CLI) output to visualize and understand what happens after their command or playbook finishes running.

Since Ansible's CLI standard output is clean and lean, one option for tracking runtimes, getting metrics, understanding resource consumption, and gathering other information is to modify the output to include more details through using callback plugins. This article explores how to use two different Ansible callback plugins to profile system activity and maximum memory usage of tasks and full execution using cgroups.

There is a lot of ground to cover with this topic. This article sets the stage by defining the environment and necessary components. A second article will cover playbook execution tests using the callback plugins.

Setting up the test environment

This scenario needs only a managed node and a target machine. It uses the managed node (ServerA) to set the configuration files and playbooks. It runs the tests in the target machine (ServerB), where the httpd service is installed and configured to act as a test web server. Both ServerA and ServerB have the following configuration:

• CPU: 1 vCPU
• Memory: 1GB
• Disk: 10GB
• OS: RHEL 8.4
• Ansible version: 2.9.26

[ Register for our free Ansible essentials online course. ]

All commands are executed on the managed node. Start by creating an inventory file with just the target machine's hostname:

[servera]$ cat << EOF > hosts
[demo]
serverb
EOF

Next, create a simple ansible.cfg file with some access configuration information:

[servera]$ cat << EOF > ansible.cfg
[defaults]
inventory=hosts
remote_user=alexon

[privilege_escalation]
become=True
become_method=sudo
become_user=root
become_ask_pass=False
EOF

As I mentioned earlier, the target machine is a basic web server. Create two simple playbooks. The first, deploy-webserver.yml, installs, configures, and starts the httpd service on the target with a custom index.html page. The second, remove-webserver.yml, removes all of that configuration.

Here is the first playbook:

[servera]$ cat << EOF > deploy-webserver.yml
---
- name: Demo playbook to test callback plugin - Deployment
  hosts: serverb
  tasks:
    
    - name: Install httpd package
      dnf:
        name: httpd
        state: latest
    
    - name: Start and enable httpd service
      service:
        name: httpd
        enabled: true
        state: started

    - name: Create a custom index.html file
      copy:
        dest: /var/www/html/index.html
        content: |
          Enable SysAdmin Demo:
          Ansible Profiling with Callback Plugin
          Custom Web Page
EOF

And here is the second playbook:

[servera]$ cat << EOF > remove-webserver.yml
---
- name: Demo playbook to test callback plugin - Remove
  hosts: serverb
  tasks:

    - name: Stop and disable httpd service
      service:
        name: httpd
        enabled: false
        state: stopped

    - name: Remove httpd package
      dnf:
        name: httpd
        state: absent
EOF

You should have the following four files in your demo directory:

[servera]$ ls -1
ansible.cfg
deploy-webserver.yml
hosts
remove-webserver.yml

Testing the environment

First, test communication with the target machine by pinging it through an Ansible ad-hoc command:

[servera]$ ansible -i hosts -m ping serverb

serverb | SUCCESS => {
    "ansible_facts": {
        "discovered_interpreter_python": "/usr/libexec/platform-python"
    },
    "changed": false,
    "ping": "pong"
}

Run the playbooks to verify that they work as expected. First, run the deploy-webserver.yml playbook:

[servera]$ ansible-playbook deploy-webserver.yml

PLAY [Demo playbook to test callback plugin - Deployment] *******************************************************************************************

TASK [Gathering Facts] ******************************************************************************************************************************
ok: [serverb]

TASK [Install httpd package] ************************************************************************************************************************
changed: [serverb]

TASK [Start and enable httpd service] ***************************************************************************************************************
changed: [serverb]

TASK [Create a custom index.html file] **************************************************************************************************************
ok: [serverb]

PLAY RECAP ******************************************************************************************************************************************
serverb                    : ok=4    changed=2    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0

Success! What about the remove-webserver.yml playbook?

[servera]$ ansible-playbook remove-webserver.yml

PLAY [Demo playbook to test callback plugin - Remove] ***********************************************************************************************

TASK [Gathering Facts] ******************************************************************************************************************************
ok: [serverb]

TASK [Stop and disable httpd service] ***************************************************************************************************************
changed: [serverb]

TASK [Remove httpd package] ************************************************************************************************************************
changed: [serverb]

PLAY RECAP ******************************************************************************************************************************************
serverb                    : ok=3    changed=2    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0

It works like a charm. But how do you know answers to things like: How long does it take each task to run? How many resources (memory and CPU) were consumed? What was the playbooks' total runtime? And how much of the total resources were consumed from the target machine? These look like questions for the second article!

Wrapping up

In this article, I defined the issue of tracking runtimes, getting metrics, and consuming resources in Ansible's CLI output when running playbooks. I covered basic adjustments for both Ansible configuration files and playbooks. I also established the infrastructure you'll use in the next article in the series, which goes into the tests for profiling resources and executing tasks with Ansible callback plugins.