Moving from Docker to Kubernetes - Retool's PostgreSQL DB backup

I'm running Retool on a single docker server (running on tag: latest).

My docker-compose.yml file is pretty old, circa late 2022 when i grabbed it from Retool so its missing a lot of the newer components like Workflows.

I'm building an on prem Kubernetes cluster (using helm) and I plan to port the database across to the new setup. Am I likely to face issues importing the PostgreSQL backup file onto a fresh database server given its never seen any of the newer features like workflows? Example no tables for workflows would exist on the backup file.

Will Retool build those missing tables automatically into the new database server or should I start with a fresh database and copy across the apps manually?

Hey there @rcanpolat :wave:

In this scenario, as long as you're keeping your Retool version the same before and after migrating, that removes risk for migration issues. Retool should take care of the rest. If you plan on upgrading, do that before you begin.

Additionally, to avoid any trouble, before you begin this process I suggest hard-coding the specific Retool version you wish to use. In using the :latest tag, your Retool instance can upgrade without you necessarily being aware.

For additional resources, here are the docs on externalizing your db for Kubernetes and our suggested best practices for self-hosted upgrades.

Best of luck with the new deployment! :star2: Please feel free the reach out with any lingering questions as you execute.

1 Like

Thanks @becca

So in my Kubernetes test environment I ran the PostgreSQL subchart to confirm Retool was fully working (including workflows).

I've since:

Uninstalled the Helm chart
Disabled the PostgreSQL subchart
Spun up an external PostgreSQL db
Imported the .sql backup from my old docker environment to the new PostgreSQL db
Installed the Helm chart pointing to the new external db.

When i check my pods on Kubernetes everything is running except for the workflow worker. I'm seeing this error on the logs for the worker pod. I can't get the workflow worker to start, it keeps going to Error and CrashLoopBackOff

Scrap that, issue was with the encryption key and special characters.

Ah, glad you found the issue. :clap: Hopefully all is up and running. If not, let us know. :blush: