Mark Machine healthy condition as unknown if we can't list wl nodes

Signed-off-by: Stefan Büringer buringerst@vmware.com
kubernetes-sigs · Dec 12, 2023 · ec91fbc · ec91fbc
1 parent 164e35f
commit ec91fbc
Show file tree

Hide file tree

Showing 2 changed files with 6 additions and 1 deletion.
diff --git a/api/v1beta1/condition_consts.go b/api/v1beta1/condition_consts.go
@@ -207,6 +207,11 @@ const (
 
 	// NodeConditionsFailedReason (Severity=Warning) documents a node is not in a healthy state due to the failed state of at least 1 Kubelet condition.
 	NodeConditionsFailedReason = "NodeConditionsFailed"
+
+	// NodeInspectionFailedReason documents a failure in inspecting the node.
+	// This reason is used when the Machine controller is unable to list Nodes to find
+	// the corresponding Node for a Machine by ProviderID.
+	NodeInspectionFailedReason = "NodeInspectionFailed"
 )
 
 // Conditions and condition Reasons for the MachineHealthCheck object.

diff --git a/internal/controllers/machine/machine_controller_noderef.go b/internal/controllers/machine/machine_controller_noderef.go
@@ -79,8 +79,8 @@ func (r *Reconciler) reconcileNode(ctx context.Context, s *scope) (ctrl.Result,
 			// No need to requeue here. Nodes emit an event that triggers reconciliation.
 			return ctrl.Result{}, nil
 		}
-		log.Error(err, "Failed to retrieve Node by ProviderID")
 		r.recorder.Event(machine, corev1.EventTypeWarning, "Failed to retrieve Node by ProviderID", err.Error())
+		conditions.MarkUnknown(machine, clusterv1.MachineNodeHealthyCondition, clusterv1.NodeInspectionFailedReason, "Failed to get the Node for this Machine by ProviderID")
 		return ctrl.Result{}, err
 	}