(bug) Fix Helm conflict race when chart ownership transfers between ClusterSummaries#1741
Merged
gianlucam76 merged 1 commit intoprojectsveltos:mainfrom Apr 23, 2026
Conversation
…lusterSummaries When a new ClusterSummary (B) is created before an existing one (A) is deleted, and both reference the same Helm chart, B may receive a NonRetriableError for chart ownership conflict while A is still being withdrawn. The existing recovery path handles the normal case correctly: when A is deleted, it unregisters from the chart manager and requeues all conflicting ClusterSummaries so they retry. The race occurs when B's deploy worker is still running at that moment. IsInProgress(B) returns true, the requeue is skipped, and nobody notifies B once A is gone and B has become the rightful owner. B's deploy worker eventually finishes and stores the NonRetriableError. When B's reconciler reads it, the conflict is already resolved, A has been removed from the chart manager and B now owns the chart. Without this fix, B is permanently marked FailedNonRetriable and never retried, since NonRetriableError intentionally suppresses further reconciliation. The fix detects this at the moment B processes the stale result: if all charts that were in conflict can now be managed by B, the error is treated as resolved, status is reset to Provisioning, and a requeue is returned. The next reconciliation cycle submits a fresh deploy.
gianlucam76
force-pushed
the
handle-helm-conflict-race
branch
from
d510fab to
83d2813
Compare
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
When a new ClusterSummary (B) is created before an existing one (A) is deleted, and both reference the same Helm chart, B may receive a NonRetriableError for chart ownership conflict while A is still being withdrawn.
The existing recovery path handles the normal case correctly: when A is deleted, it unregisters from the chart manager and requeues all conflicting ClusterSummaries so they retry.
The race occurs when B's deploy worker is still running at that moment. IsInProgress(B) returns true, the requeue is skipped, and nobody notifies B once A is gone and B has become the rightful owner.
B's deploy worker eventually finishes and stores the NonRetriableError. When B's reconciler reads it, the conflict is already resolved, A has been removed from the chart manager and B now owns the chart. Without this fix, B is permanently marked FailedNonRetriable and never retried, since NonRetriableError intentionally suppresses further reconciliation.
The fix detects this at the moment B processes the stale result: if all charts that were in conflict can now be managed by B, the error is treated as resolved, status is reset to Provisioning, and a requeue is returned. The next reconciliation cycle submits a fresh deploy.
Fixes 463