User credentials lost after pod restart in Kubernetes deployment

[gabrielpulga]

Checklist

• I have searched Budibase discussions and GitHub issues to check if my issue already exists

Hosting

• Self
  • Method: k8s
  • Budibase Version: latest as of 10/28/2024
  • App Version: 2.33.2

Describe the bug
After deploying Budibase on Kubernetes using Helm, everything works as expected: I can log in, set up an owner account, and all functionality appears normal. However, when I restart the pods or redeploy the app, my credentials stop working, and I’m unable to log in with the previously created owner account. It seems as though user data may not be persistent across restarts, possibly due to missing persistent storage.

To Reproduce
Steps to reproduce the behavior:

1. Deploy Budibase using Helm with mostly default values, changing only the ingress settings to use a custom host domain with Azure Application Gateway.
2. Set up an owner account and log in.
3. Restart the pods or redeploy the app.
4. Attempt to log in again with the same credentials created in step 2.
5. See error: credentials are no longer recognized.

Expected behavior
User credentials and configuration should persist across pod restarts or redeployments, allowing me to log back into the app with the previously created account.

Screenshots
N/A

App Export
N/A

Desktop (please complete the following information):

• OS: [Provide your OS]
• Browser: Chrome
• Version: [Browser version if known]

Additional context
It’s unclear whether this issue is related to a specific service in Budibase (perhaps the app-service), or if certain data is not configured to use persistent storage. Any assistance with debugging the root cause or setting up persistent storage for user data would be appreciated.

[linear]

BUDI-8789 User credentials lost after pod restart in Kubernetes deployment

[gabrielpulga]

Any updates on this issue? I am pretty sure it is missing persistent storage by default, which is confusing because I can't find anything in the documentation about it (Kubernetes self hosting in Azure Kubernetes Service)

[gabrielpulga]

Seems like this issue is also related to it #10411

[gabrielpulga]

Solved this by adding PVC to CouchDB subchart