Re: Containers
What works for me is to put the data onto an OpenStack data volume which can be attached to an instance, have the DB engine (Postgres, Mongo, whatever you're having yourself) run in a container on the instance and mount the data directory from the volume as a container volume. This gives me the DB engine container which can be freely recreated at any time, the instance VM is also relatively transient as it can be easily recreated and the volume containing the data reattached to the new instance.
This is on OpenStack which just happens to be the environment I'm using, but similar approaches seem to be applicable on AWS, Azure etc.
Biting the hand that feeds IT
