equivalent of install-rp-rpm.yml to install the rpm. We'll assume they're installing from a local repo where the broker rpm already lives so we don't have to add another repo url. (and helps solve any weird airgap things on an MVP of the playbook)
probably need to manage truststore/cert files and turn them into appropriate java versions.
probably can stub out a quick unit file for launching connect or grab one. Here's an example. it's MIT-licensed. Not saying we use it directly, but it's an example to understand.
probably need to tune log4j.
equivalent of start-redpanda.yml to lay down configs, ensure that systemd units are there, enabled, and we can control the service.
need to have a way to wire in the KC host into console config. That'll require modification of start-console.yml (and probably it's associated jinja templates for configs)
need to probably create some specific host groups in hosts.ini for tagging which nodes are the KC nodes?
There's also separate concern about missing systemd unit files for KC startup, but that's a problem we'll temporarily solve with ansible deploying a custom setup until this is fixed to be provided more natively by the deployable package.
for ansible, I'm guessing we need the following:
There's also separate concern about missing systemd unit files for KC startup, but that's a problem we'll temporarily solve with ansible deploying a custom setup until this is fixed to be provided more natively by the deployable package.