add-app-to-platform should support a variables-file like Ansible's

[MikeTaylor]

@wafschneider will send me an example for how this file looks in Ansible, and I will add support for it.

[MikeTaylor]

Here is a sample Ansible variables file, supplied by @wafschneider. sample.yml.txt

And here is the equivalent JSON file, made using https://onlineyamltools.com/convert-yaml-to-json . sample.json.txt

[MikeTaylor]

The {{ stuff }} sequences are templates that are interpreted by Ansible using Jinja2 expressions to bring in values from Ansible Vault. Rather than try to reimplement all that machinery, we will do the template substitution ahead of time, so that add-app-to-platform can use the values as-is.

[MikeTaylor]

I have questions about the Ansible variable-file format.

• I assume I should ignore the tenants section at the end, on the basis that I am only (at this stage) working on tenant-agnostic deployment?
• Should I ignore the labels namespaces and shared-prod-2 and only use the contents of whatever is beneath them?
• I guess pull_secret is irrelevant here?
• If okapi_url, okapi_admin and okapi_password are defined, should I make them override whatever is set in the environment?
• I guess that within folio_modules I should pick out only the stanza referring to the module being deployed, and ignore all the others?
• Should I use replicas in the Okapi deployment plugin, or is that only relevant for Kubernetes?
• The docker_env settings go in the env section of the launch descriptor, right?
• Should I translate docker_mem into a setting for the launch descriptor’s dockerArgs?
• Should anything else go into dockerArgs?
• What is need_db?