fix: Reload RedisCluster instance after status updates conflict (#1583)

* fix: Reload RedisCluster instance after status updates

Fix race condition where status checks used stale instance data,
preventing health checks from running. Reload instance after each
status update to ensure subsequent checks have fresh status values.

This fixes cluster state reporting issues where ReadyLeaderReplicas
and ReadyFollowerReplicas checks always failed due to stale data.

Signed-off-by: Damian Seredyn <[email protected]>

* retry on conflict

Signed-off-by: drivebyer <[email protected]>

* update concurrent

Signed-off-by: drivebyer <[email protected]>

---------

Signed-off-by: Damian Seredyn <[email protected]>
Signed-off-by: drivebyer <[email protected]>
Co-authored-by: Damian Seredyn <[email protected]>
Co-authored-by: drivebyer <[email protected]>

Author: 3 people

.github/workflows/ci.yaml

@@ -260,7 +260,7 @@ jobs:
       - name: Install Redis Operator
         run: |
           make deploy IMG=redis-operator:e2e
-          kubectl set env deployment/redis-operator-redis-operator -n redis-operator-system OPERATOR_IMAGE=redis-operator:e2e FEATURE_GATES=GenerateConfigInInitContainer=true
+          kubectl set env deployment/redis-operator-redis-operator -n redis-operator-system OPERATOR_IMAGE=redis-operator:e2e FEATURE_GATES=GenerateConfigInInitContainer=true MAX_CONCURRENT_RECONCILES=10
 
       - name: Wait for Redis Operator to be ready
         run: |

internal/controller/rediscluster/rediscluster_controller.go

@@ -32,6 +32,7 @@ import (
 	retry "github.com/avast/retry-go"
 	appsv1 "k8s.io/api/apps/v1"
 	corev1 "k8s.io/api/core/v1"
+	apierrors "k8s.io/apimachinery/pkg/api/errors"
 	"k8s.io/client-go/kubernetes"
 	"k8s.io/client-go/tools/record"
 	ctrl "sigs.k8s.io/controller-runtime"
@@ -300,7 +301,8 @@ func (r *Reconciler) Reconcile(ctx context.Context, req ctrl.Request) (ctrl.Resu
 	}
 
 	// Mark the cluster status as ready if all the leader and follower nodes are ready
-	if instance.Status.ReadyLeaderReplicas == leaderReplicas && instance.Status.ReadyFollowerReplicas == followerReplicas {
+	// and the cluster is not already in Ready state (to avoid unnecessary status updates)
+	if instance.Status.ReadyLeaderReplicas == leaderReplicas && instance.Status.ReadyFollowerReplicas == followerReplicas && instance.Status.State != rcvb2.RedisClusterReady {
 		monitoring.RedisClusterHealthy.WithLabelValues(instance.Namespace, instance.Name).Set(0)
 		if k8sutils.RedisClusterStatusHealth(ctx, r.K8sClient, instance) {
 			monitoring.RedisClusterHealthy.WithLabelValues(instance.Namespace, instance.Name).Set(1)
@@ -339,7 +341,22 @@ func (r *Reconciler) updateStatus(ctx context.Context, rc *rcvb2.RedisCluster, s
 	copy := rc.DeepCopy()
 	copy.Spec = rcvb2.RedisClusterSpec{}
 	copy.Status = status
-	return common.UpdateStatus(ctx, r.Client, copy)
+	err := common.UpdateStatus(ctx, r.Client, copy)
+	if err != nil && apierrors.IsConflict(err) {
+		log.FromContext(ctx).Info("conflict detected, reloading instance and retrying status update")
+		namespacedName := client.ObjectKey{
+			Namespace: rc.Namespace,
+			Name:      rc.Name,
+		}
+		if err := r.Get(ctx, namespacedName, rc); err != nil {
+			return err
+		}
+		copy = rc.DeepCopy()
+		copy.Spec = rcvb2.RedisClusterSpec{}
+		copy.Status = status
+		return common.UpdateStatus(ctx, r.Client, copy)
+	}
+	return nil
 }
 
 // SetupWithManager sets up the controller with the Manager.