Improve handling of failed updates - delete and new commits - #966
Conversation
|
PR is now waiting for a maintainer to run the acceptance tests. This PR will only perform build and linting. |
|
cc @blampe |
|
PR is now waiting for a maintainer to run the acceptance tests. This PR will only perform build and linting. |
|
/run-acceptance-tests |
|
Please view the PR build: https://github.com/pulumi/pulumi-kubernetes-operator/actions/runs/16628182496 |
| if stack.DeletionTimestamp != nil { // Marked for deletion (and has already been destroyed). | ||
| return true | ||
| } |
There was a problem hiding this comment.
It is incorrect to consider the DeletionTimestamp here, because the failed update might be the destroy operation that is triggered by the spec.destroyOnFinalize field and runs during finalization. The system must continue to retry the destroy op until it succeeds. That said, it does make sense to re-trigger a destroy if the commit has changed.
Also, if a stack was "stuck" in a failed destroy loop, the corrective action is to edit the spec and set spec.destroyOnFinalize to false. This causes the finalization routine to complete.
|
I took a minute to update the PR, please stand by. |
updated the tests for clarity
|
PR is now waiting for a maintainer to run the acceptance tests. This PR will only perform build and linting. |
|
/run-acceptance-tests |
|
Please view the PR build: https://github.com/pulumi/pulumi-kubernetes-operator/actions/runs/16633521733 |
4 participants
Proposed changes
If the last update failed check the following:
Related issues (optional)
Fixes #951