Postgres is increasingly being run on production on containers and also on Kubernetes, thanks to the various existing Postgres operators. But running applications like databases in containers, with heavy requirements on the storage layer, definitely presents challenges.This talk will be focused on describing how Postgres interacts with the Storage subsystem, and in particular what advanced requirements it poses on the storage. This is easy to extrapolate to any other database or even other containerized data-intensive workloads.
In particular, this talk will analyze:
* What is Postgres storage architecture.
* What are usual performance expectations.
* How backups are/should be handled.
* How and what for snapshots can be used.
* The implications of database replication.
* Other advanced topics like tablespaces or WAL separation.