fail provision with clear message when runtime-automation dir is missing

[prakhar1985]

Problem

When a content author sets showroom_ansible_runner_api: true in their AgV catalog but forgets to add a runtime-automation/ directory to their content repo, podman-compose fails silently at startup with:

Error: statfs /opt/showroom/content/runtime-automation: no such file or directory

This causes showroom.service to exit 125. The provision appears to succeed but the showroom is unreachable — the OCP pod is marked not-ready and the router returns 503.

Root cause: The default for showroom_ansible_runner_api is false. The failure only occurs when a dev explicitly enables it in their AgV config without providing the required runtime-automation/ directory in the content repo.

Fix

Add an early check in 30-showroom-clone-and-inject.yml (after clone, before compose is rendered) that fails the provision immediately with a clear, actionable message:

showroom_ansible_runner_api is true but the content repo
'https://github.com/...' (ref: main)
does not contain a runtime-automation/ directory.
Add runtime-automation/ with at least module-XX/solve.yml and
module-XX/validation.yml stubs, or set showroom_ansible_runner_api: false.

The stat and fail tasks are both gated on showroom_ansible_runner_api | bool, so there is zero overhead for the common case where it is false.

Behaviour

showroom_ansible_runner_api	runtime-automation/ in repo	Result
false (default)	—	both tasks skipped, no change
true	✅ present	stat passes, no failure
true	❌ missing	provision fails with clear message