your management layer should be cattle too
2021-02-06, 13:00–13:30, D.infra

The "pet vs cattle" thing is getting pretty long in the tooth, but the one thing one keeps hearing is that the management layer somehow always remains a pet.

While there is certainly some truth to that - you don't want to rebuild your management plane every day - wouldn't it be cool if you could?

In this talk we will show how to fully automate the deployment and configuration of Foreman using Ansible, which allows for reproducible setups, redeployments and testing environments that actually match production.

The presented solution should not (and can not) replace working backups, but it might still save you when the restore isn't going as expected or someone asks to start "fresh" while keeping all the important details the same as before.

We will show how to deploy Foreman from scratch in an reproducible way, reusing the same bits the project uses itself. This will also include plugins like Katello and Ansible for best possible client management later.

We'll continue with integrating Foreman with compute resources and managing content in Katello, which will result in trivial deployments of further cattle systems from Foreman, which then can be managed via Ansible.

The result is a set of playbooks that can be stored in a git repository and deployed at will.

Debian Developer, Red Hat Engineer, ♥ automation