Open gustavosr98 opened 1 year ago
@gustavosr98 thank you for the request, somehow we missed to reply it.
It was a design decision to minimize amount of charm corner cases and provide a high quality operator. The scaling capabilities of the current charm is well-tested and optimized from performance point of view.
I see the further possible optimizations for MAAS/VM installations where Juju unit setup/init can take some time. I will keep this open for the future. Thank you!
This is not a bug. It is more of a feature request / whishlist.
I see one of the requirements for postgresql charm to be able to restore from a backup is to scale in to a single unit
Would there be a way to be able to recover from a backup without having to scale in the cluster?
I am mostly thinking about escenarios where scaling in and out again can take quite some time, thus affecting availability, like baremetal deployments and deployments with many replicas