AFAIU, there are sound reasons to host containers with e.g. OpenStack VMs instead of a k8s scheduler with a proper SAN and just figure out how to redundantly and resiliently sync - replicate, synchronize, primary/secondary, nodeprocd - and tune given the CAP theorem and the given DB implementation(s)?

Here's the official Ansible role for Trove, which provisions various e.g. SQL databases on an OpenStack cloud: https://github.com/openstack/openstack-ansible-os_trove