Replace Vagrantfile with GitHub Action and virtual machines

[bastelfreak]

Right now this repository has a Vagrantfile. This lists a huge amount of virtual machines. All VMs execute a script to collect factsets for different facter versions. We can heavily improve the setup:

• Update the script to rely openvox-agent and openfact
• create a github workflow that will be triggered manually
• Generate a CI matrix based on a list of operating systems
• in each VM, generate the factsets
• publish the factsets as github artifact in each job
• final job collects the artifacts and publishes them as a PR

Some additional documentation:

• https://github.com/OpenVoxProject/openvox-agent/blob/main/.github/workflows/acceptance.yml ->
• https://github.com/OpenVoxProject/shared-actions/blob/main/.github/workflows/beaker_acceptance.yml ->
• https://github.com/jpartlow/kvm_automation_tooling

[yakatz]

Are all the operating systems we generate factsets for available in Github Actions?

[bastelfreak]

Github offers Ubuntu, Windows and MacOS. But that doesn't matter here. The linked openvox-agent workflow uses the Ubuntu runners to start different virtual machines, within the runner, via nested virtualization. This isn't the most efficient setup, but we don't need a lot of performance. And the virtual machine can use whatever cloudimages are available.

[ekohl]

I suppose we need VMs for proper kernel facts that containers wouldn't get right

[bastelfreak]

Yes. That's why I had the idea with the nested virtual machines.