The Initializer container Job is stuck and doesn't complete the Job because of the CloudSQL Proxy

[bruno561]

Be informative
When using Cloud SQL as a managed database, the Job related to Initializer isn't completed because your Sidecar is still running. Sidecar is cloudsql-proxy.

Bug description
As described in the informational section, the Initializer Job fails when using CloudSQL. NotReady status is displayed.

Image of the logs of the container named Initializer. Logs stop at this point:

Image of Pod related to Job with status of NotReady.

Steps to reproduce
Basically, just use Cloud SQL in the Values file.

  # Google CloudSQL support in GKE via gce-proxy
  cloudsql:
    # To use CloudSQL in GKE set 'enable: true'
    enabled: true

Expected behavior
Have some way that when the initializer container finishes the job the cloudsql-proxy container is deleted, in order to finish the Job.

Deployment method (select with an X)

• Docker Compose
• Kubernetes
• GoDojo

Environment information

• GKE (Google Kubernetes Engine) version 1.23.12-gke.100
• https://github.com/DefectDojo/django-DefectDojo/tree/2.16.2/helm/defectdojo

Logs

The cloudsql-proxy container hangs at this point:

[moxli]

This results in a failing Helm chart due to the fact that the helm installer waits for the initializer pod to finish (which it never does).
As soon as the initializer pod is deleted manually, the helm installer finishes.

[tellierd]

Has anyone worked on a fix for this? I'm having the very same issue. Thx

[mtesauro]

@tellierd TBH, Helm isn't an officially supported deployment method for DefectDojo so we rely on the community to submit PRs for the Helm chart in this repo.

There are far to many deployment choices for k8s/Helm for an open source project to able to test them all.

[jndeverteuil]

With PR #10824 the bug is now fixed from what I tested

Also, PR #11363 has fixed the helm charts not being published

This means, the bugfix is available in Helm Chart version 1.6.164

[mtesauro]

Given that we have a report that this is working, I'm going to go ahead and close this.

Reopen or create a new issue if that's not the case.