I’ve been running PMS for a few years now in my homelab and love it.
These days my homelab is essentially 100% kubernetes. Initially deploying plex via kubernetes revealed that sqlite experiences data corruption without a fully functional file locking implementation (aka don’t use NFS). Eventually I solved this using a local disk and restricting the pod to a single workernode, and later on after deploying ceph, found that sqlite was happy running via a network share that lives on cephfs. Nextcloud suffers similar sqlite legacy issues.
I’m sure this question comes up now and then:
–> Any plans to implement any other database support?
It’s so easy to deploy a database via kubernetes. Allowing other databases to be supported lets those databases offer features such a HA, and faster speed by spreading the load across nodes. I know the answer is probably I should write my own plex using yugabyte, but thought I’d ask… don’t really have time less I manage to become financially independent.
SQLite3 is the perfect database choice for Plex Media server. See:
Note, in particular, that in this case, Plex Media Server is the client. Not Plex’s individual client apps. None of them ever interact with the database directly.
This idea of implementing other database support for PMS has come up many times in the past. And the driving factor for all of them has been to introduce enterprise features for a product which is, and has always been intended to be, a personal media server. Distributed load, high-availability, etc… are not relevant here.
i wonder if another personal media server system were to come about, similar to plex, which could scale … if suddenly plex would be come interested, but a little late to the game.
i’m just setting something up for myself, but i could see how it starts making folks think about an enterprise deployment, which PMS is not intended to be
