Added/modified test-cases

controllers/pod.go

@@ -650,7 +650,7 @@ func (r *SingleClusterReconciler) cleanupDanglingPodsRack(sts *appsv1.StatefulSe
 
 // getIgnorablePods returns pods:
 // 1. From racksToDelete that are currently not running and can be ignored in stability checks.
-// 2. User given pods in ignorePodList that are currently not running and can be ignored from stability checks.
+// 2. Failed/pending pods identified using maxIgnorablePods field and can be ignored from stability checks.
 func (r *SingleClusterReconciler) getIgnorablePods(racksToDelete []asdbv1.Rack, configureRacks []RackState) (
 	sets.Set[string], error,
 ) {

controllers/rack.go

@@ -419,12 +419,11 @@ func (r *SingleClusterReconciler) upgradeOrRollingRestartRack(found *appsv1.Stat
 		}
 	}
 
-	podList, err := r.getOrderedRackPodList(rackState.Rack.ID)
-	if err != nil {
-		return found, reconcileError(fmt.Errorf("failed to list pods: %v", err))
-	}
-
 	if r.aeroCluster.Spec.RackConfig.MaxIgnorablePods != nil {
+		podList, err := r.getOrderedRackPodList(rackState.Rack.ID)
+		if err != nil {
+			return found, reconcileError(fmt.Errorf("failed to list pods: %v", err))
+		}
 		// Filter ignoredPods to update their dirtyVolumes in the status.
 		// IgnoredPods are skipped from upgrade/rolling restart, and as a result in case of device removal, dirtyVolumes
 		// are not updated in their pod status. This makes devices un-reusable as they cannot be cleaned up during init phase.