ROX-31829: Implement aggressive conflict resolution #64
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
mclasmeier
force-pushed
the
mc/no-ssa-4
branch
from
270ea6f to
542ad8a
Compare
mclasmeier
changed the title
porridge
reviewed
Nov 18, 2025
Collaborator
porridge
left a comment
porridge
left a comment
There was a problem hiding this comment.
Looks OK in general, some comments inline.
Also, the custom predicates bit should probably be a separate PR for ease of upstreaming later?
Contributor
Author
|
The |
added 16 commits
November 19, 2025 10:04
updater tests and use Ginkgo assertions instead of production error handling.
mclasmeier
force-pushed
the
mc/no-ssa-4
branch
from
542ad8a to
cc60c7d
Compare
Contributor
Author
|
@porridge Hope I have addressed everything. Thank you! |
porridge
reviewed
Nov 20, 2025
Collaborator
porridge
left a comment
porridge
left a comment
There was a problem hiding this comment.
Some questions and comments inline.
pkg/reconciler/reconciler.go
Outdated
| // This is a safety measure to ensure that the chart is fully uninstalled before the CR is deleted. | ||
| if obj.GetDeletionTimestamp() == nil && !controllerutil.ContainsFinalizer(obj, uninstallFinalizer) { | ||
| log.V(1).Info("Adding uninstall finalizer.") | ||
| patch := client.MergeFrom(obj.DeepCopy()) |
Collaborator
There was a problem hiding this comment.
Looking at the doc for MergeFrom, it seems like we want to use strategic merge patch for the list that finalizers is, no? 🤔
Contributor
Author
There was a problem hiding this comment.
- Agreed, there was a race, thanks for noticing.
- StrategicMergePatch() unfortunately doesn't fix it. I think this has to do with the fact that the
finalizerslist has no merge strategy associated. It would still just replace -- I tested this with a test using an interceptor to modify an object during aGet()call while still returning the old one. - I think a JSON merge patch might do the trick here, but thinking about this again, I think we can refrain from complicating this piece of code for two reasons:
3a. The reconciler code as it is now is not expecting any other finalizers to be in the picture, it's deliberately replacing them withSetFinalizers()anyway.
3b. Even if there was a conflict here, I think this is one of the few situations where we could keep failing with a conflict error, because it happens right at the beginning of the reconciliation, before any heave Helm work is being done. And also because this is not something that should happen frequently -- it is, as far as I can tell, not expected that users fiddle around with the finalizers of a helm-operator-managed CR.
In any case, modifying this behavior seems to be independent of this specific feature.
Hence, I have reverted this change so that we have the old behavior.
added 6 commits
November 21, 2025 10:24
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Implements a conflict resolution strategy for use in rhacs-operator.
Can be reviewed commit by commit.