Skip to content

statefulset: make pod restart detection idempotent #1133

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.

Sign up for GitHub

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

Merged
birdayz merged 3 commits into from
Oct 9, 2025
Merged

statefulset: make pod restart detection idempotent #1133

Changes from 1 commit
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
bf800bd
statefulset: make pod restart detection idempotent
birdayz Oct 9, 2025
f868986
statefulset: add unit test for idempotent pod restart detection
birdayz Oct 9, 2025
c123afd
Fix formatting in test file
birdayz Oct 9, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
16 changes: 14 additions & 2 deletions operator/pkg/resources/statefulset_update.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -94,7 +94,7 @@ func (r *StatefulSetResource) runUpdate(
// Check if we should run update on this specific STS.

log.V(logger.DebugLevel).Info("Checking that we should update")
stsChanged, nodePoolRestarting, err := r.shouldUpdate(current, modified)
stsChanged, nodePoolRestarting, err := r.shouldUpdate(ctx, current, modified)
if err != nil {
return fmt.Errorf("unable to determine the update procedure: %w", err)
}
Expand Down Expand Up @@ -629,7 +629,7 @@ func (r *StatefulSetResource) updateStatefulSet(
// If no differences are found, the function returns `false` as first value. The second value
// depends on the Node Pool Status, indicating whether rolling restart is still in progress.
func (r *StatefulSetResource) shouldUpdate(
current, modified *appsv1.StatefulSet,
ctx context.Context, current, modified *appsv1.StatefulSet,
) (bool, bool, error) {
log := r.logger.WithName("shouldUpdate")

Expand All @@ -648,6 +648,18 @@ func (r *StatefulSetResource) shouldUpdate(
return false, true, nil
}

// Check if any pods have ClusterUpdate condition set.
// This makes the controller idempotent: if a previous reconcile set the condition
// but the status flag wasn't persisted (due to controller restart, stale cache, etc.),
// we'll still detect that a rolling update is needed by checking actual pod state.
allPodsUpdated, err := r.areAllPodsUpdated(ctx)
if err != nil {
log.V(logger.DebugLevel).Error(err, "error checking if pods need update")
} else if !allPodsUpdated {
log.Info("Found pods with ClusterUpdate condition, continuing rolling update")
return false, true, nil
}

return false, false, err
}
log.Info("Detected diff", "patchResult", string(patchResult.Patch))
Expand Down
Loading