Tackling Ansible Scheduling with the at Command

Introduction

In this hands-on lab, we will use a playbook with the at module for job scheduling purposes.

Note: Ansible has been set up and configured already for use.

Solution

To access the lab environment, log in to the control node as cloud_user, using the IP address and login information provided in Credentials section of the hands-on lab page.

Sign onto the Ansible Control Node Server as cloud_user and Change to the ansible User

1. Sign in to the server called Ansible Control Node using the cloud_user and change to the ansible user via the su - ansible command.
2. Test that ansible is working via an ad-hoc command.

3. We can use the following:

   ansible all -m ping

Create and Run a Playbook Called install-at.yml that Installs, Enables and Starts the at Service on All Nodes

1. Create a playbook called install-at.yml that will install the at service on all nodes:

   vim install-at.yml

2. Enable that service atd and start the service as part of the playbook:

   hosts: all
   user: ansible
   become: yes
   gather_facts: no
   tasks:
   - name: install the at command for job scheduling
    action: yum name=at state=installed
   - name: Enable and Start service at if not started
    service:
      name: atd
      enabled: yes
      state: started

3. Save and exit.

4. Run the playbook and ensure that it installs, enables and starts the service correctly:

   ansible-playbook install-at.yml

5. Run the following to view more information about our at command:

   man at

6. We can clear our screen.

Create a Playbook Called at-scheduled-task.yml that Performs a Task on the Nodes in 20 Minutes

1. Using the at module, create a playbook called at-scheduled-task.yml that adds a scheduled task to the nodes:

   vim at-scheduled-task.yml  

   Note: The job should run in 20 minutes.

2. The task to run is df -h > /tmp/diskspace, which will appear like so:

   hosts: all
   user: ansible
   become: no
   gather_facts: no
   tasks:
   - name: Schedule a command to execute in 20 minutes as the ansible user
     at:
      command: df -h > /tmp/diskspace
      count: 20
      units: minutes

3. Save and exit.

4. Run the following:

   ansible-playbook at-scheduled-task.yml

Run the at-scheduled-task Playbook and Test Each Node to Ensure the Task has been Scheduled

1. On each node we should test the df -h task has been scheduled. We can use the following command on each node to see if it has been scheduled:

   atq
   at -c 1

2. We can clear the screen.

Create Playbook Called remove-at-task.yml That Will Remove the Previously Scheduled Task

1. We will create it with:

   vim remove-at-task.yml

2. The remove-at-task.yml playbook should remove the task that was set in the at-scheduled-task.yml playbook and the setup will look like this:

   hosts: all
   user: ansible
   become: no
   gather_facts: no
   tasks:
   
   - name: Match a command to an existing job and delete the job
   at:
     command: df -h > /tmp/diskspace
     state: absent

3. Save and exit.

4. Run our playbook with:

   ansible-playbook remove-at-task.yml

5. We can log in to each node and test with the atq command to ensure the tasks have been removed.

   Note: If they have not, we must troubleshoot why they have not been removed.

Conclusion

Congratulations — you've completed this hands-on lab!