# Molecule 6, drivers have become scenarios

**URL:** <https://forum.ansible.com/t/molecule-6-drivers-have-become-scenarios/6679>\
**Category:** Project Discussions\
**Tags:** devtools, molecule6, molecule\
**Created:** [June 20, 2024, 12:43pm UTC](https://forum.ansible.com/t/molecule-6-drivers-have-become-scenarios/6679 "2024-06-20T12:43:51Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![ebbex](https://avatars.discourse-cdn.com/v4/letter/e/dfb087/32.png) [@ebbex](https://forum.ansible.com/u/ebbex)\
**Post date:** [June 20, 2024, 12:43pm UTC](https://forum.ansible.com/t/molecule-6-drivers-have-become-scenarios/6679/1 "2024-06-20T12:43:51Z")

</div>

First off, I think the direction of eliminating drivers and just using ansible itself with collections to set up the “infrastructure” to test against is probably a good ting.

The downside is you’re now picking a tech/collection in your `create.yml`, and there’s (currently?) no good way to pick a different tech/collection for a given scenario, which I think is rather unfortunate.

So now scenarios like `deploy_app_with_apache` and `deploy_app_with_nginx` have to become `deploy_app_with_apache_using_docker` and if I want to provide testing for a different driver/backend it’d be `deploy_app_with_apache_using_libvirt`.

Would it be reasonable to reuse `--driver` (or just a new flag `--backend`) to look for keys in the playbooks dict? (As is already done for drivers)

Something like;

```auto
provisioner:
  name: ansible
  playbooks:
    docker:
      create.yml: create-docker.yml
      destroy.yml: destroy-docker.yml
    libvirt:
      create.yml: create-libvirt.yml
      ...

```