Skip to content

CSV not removed after deletion when not in Success/Terminal state. #3130

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
tmshort opened this issue Dec 7, 2023 · 2 comments
Closed

CSV not removed after deletion when not in Success/Terminal state. #3130

tmshort opened this issue Dec 7, 2023 · 2 comments
Labels
kind/bug Categorizes issue or PR as related to a bug.

Comments

@tmshort
Copy link
Contributor

tmshort commented Dec 7, 2023

Bug Report

It appears that CSVs may not be deleted when they are "deleted". The CSV remains after a deletion attempt, which means that all resources associated with (owned by) the CSV remain.

See #3108

What did you do?
Happens during e2e (Install Plan creation with permission), it's a flake.

When a CSV is not in a (e.g.) success state, and it's deleted, it's not actually deleted.

What did you expect to see?
When a CSV is deleted, the CSV is deleted regardless of the state it's in.

What did you see instead? Under which circumstances?
When the CSV is deleted, it remains.

Environment

  • operator-lifecycle-manager version: master
  • e2e tests
  • Kubernetes version information: e2e test (most recent)
  • Kubernetes cluster kind: openshift and kind
@tmshort tmshort added the kind/bug Categorizes issue or PR as related to a bug. label Dec 7, 2023
@stevekuznetsov
Copy link
Contributor

stevekuznetsov commented Dec 7, 2023

This does not really have to do with the state that the CSV is in. The controller, as written today, is edge-driven. If it ever misses a deletion event, since there is no finalizer, you are not guaranteed that the clean-up code runs. It might be the case, in that end-to-end test, that waiting for the controller to see the CSV by not deleting until a specific state is enough to make sure the event is not missed, but it's not a generic fix to the issue.

@tmshort
Copy link
Contributor Author

tmshort commented Jan 3, 2024

What appears to be happening is that the CSV is being re-created by the Subscription, as the UIDs are different. With some of the performance enhancements, this is occurring a lot quicker now.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
kind/bug Categorizes issue or PR as related to a bug.
Projects
None yet
Development

No branches or pull requests

2 participants