fixup! fixup! fixup! fixup! fixup! fixup! fixup! fixup! fixup! fixup! fixup! fixup! fixup! fixup! fixup! fixup! fixup! General improvements

chart/Chart.yaml

@@ -10,9 +10,9 @@ description: |
 
 type: application
 
-version: 0.2.40
+version: 0.2.42
 
-appVersion: "0.2.40"
+appVersion: "0.2.42"
 
 home: https://github.com/lukaszraczylo/kubernetes-images-sync-operator
 

chart/values.yaml

@@ -12,7 +12,7 @@ sa:
         - ALL
     image:
       repository: ghcr.io/lukaszraczylo/kubernetes-images-sync-operator
-      tag: 0.2.40
+      tag: 0.2.42
     resources:
       limits:
         cpu: 500m

docker-image-worker/s3_utils.py

@@ -69,16 +69,23 @@ def get_s3_client(use_role=False, role_name=None, use_current_role=False, aws_ac
                     logger.info(f"Environment: {key}={value}")
 
             # Get the AWS region from environment or parameter
-            aws_region = region or os.environ.get('AWS_REGION') or os.environ.get('AWS_DEFAULT_REGION')
+            aws_region = os.environ.get('AWS_REGION') or os.environ.get('AWS_DEFAULT_REGION')
-            if not aws_region:
+            if not aws_region and not region:
                 raise ValueError("AWS region must be specified either through region parameter or AWS_REGION environment variable")
 
+            # Use region from parameter only if not set in environment
+            if not aws_region:
+                aws_region = region
+                # Set it in environment for other AWS clients
+                os.environ['AWS_REGION'] = region
+
             logger.info(f"Using AWS region: {aws_region}")
 
             # Create an STS client in the correct region
-            sts = boto3.client('sts', 
+            sts_kwargs = {'endpoint_url': f'https://sts.{aws_region}.amazonaws.com'}
-                region_name=aws_region,
+            if not os.environ.get('AWS_REGION') and not os.environ.get('AWS_DEFAULT_REGION'):
-                endpoint_url=f'https://sts.{aws_region}.amazonaws.com')
+                sts_kwargs['region_name'] = aws_region
+            sts = boto3.client('sts', **sts_kwargs)
 
             # Read the web identity token
             token_file = os.environ.get('AWS_WEB_IDENTITY_TOKEN_FILE')
@@ -105,14 +112,17 @@ def get_s3_client(use_role=False, role_name=None, use_current_role=False, aws_ac
                 credentials = response['Credentials']
                 
                 # Create the S3 client with the temporary credentials
-                client = boto3.client(
+                s3_kwargs = {
-                    's3',
+                    'aws_access_key_id': credentials['AccessKeyId'],
-                    region_name=aws_region,
+                    'aws_secret_access_key': credentials['SecretAccessKey'],
-                    aws_access_key_id=credentials['AccessKeyId'],
+                    'aws_session_token': credentials['SessionToken']
-                    aws_secret_access_key=credentials['SecretAccessKey'],
+                }
-                    aws_session_token=credentials['SessionToken'],
+                # Only set region_name if not already in environment
-                    **client_kwargs
+                if not os.environ.get('AWS_REGION') and not os.environ.get('AWS_DEFAULT_REGION'):
-                )
+                    s3_kwargs['region_name'] = aws_region
+                # Add any additional kwargs
+                s3_kwargs.update(client_kwargs)
+                client = boto3.client('s3', **s3_kwargs)
 
                 logger.info(f"Successfully assumed role with web identity: {response['AssumedRoleUser']['Arn']}")
                 

internal/controller/raczylo.com/clusterimageexport_controller.go

@@ -308,7 +308,7 @@ func (r *ClusterImageExportReconciler) handleDeletion(ctx context.Context, clust
 			// Continue with deletion even if cleanup fails
 		}
 
-		// Delete existing cleanup job if it exists
+		// Create or recreate cleanup job
 		jobName := "cleanup-" + shared.NormalizeImageName(clusterImageExport.Name)
 		existingJob := &batchv1.Job{}
 		err := r.Get(ctx, client.ObjectKey{
@@ -317,26 +317,35 @@ func (r *ClusterImageExportReconciler) handleDeletion(ctx context.Context, clust
 		}, existingJob)
 
 		if err == nil {
-			// Job exists, delete it
+			// Job exists, delete it first
-			if err := r.Delete(ctx, existingJob); err != nil && !errors.IsNotFound(err) {
+			deletePolicy := metav1.DeletePropagationForeground
-				l.Error(err, "Failed to delete existing cleanup job")
+			deleteOptions := client.DeleteOptions{
-				// Continue with deletion even if cleanup job deletion fails
+				PropagationPolicy: &deletePolicy,
-			} else {
-				l.Info("Successfully deleted existing cleanup job", "job", jobName)
 			}
+			if err := r.Delete(ctx, existingJob, &deleteOptions); err != nil && !errors.IsNotFound(err) {
+				l.Error(err, "Failed to delete existing cleanup job")
+				return ctrl.Result{Requeue: true}, nil
+			}
+			l.Info("Successfully deleted existing cleanup job", "job", jobName)
+			// Requeue to wait for job deletion to complete
+			return ctrl.Result{Requeue: true}, nil
 		} else if !errors.IsNotFound(err) {
-			// Unexpected error, log but continue
+			// Unexpected error
 			l.Error(err, "Error checking for existing cleanup job")
+			return ctrl.Result{Requeue: true}, nil
 		}
 
 		// Create new cleanup job
-		r.runCleanupJob(ctx, clusterImageExport)
+		if err := r.runCleanupJob(ctx, clusterImageExport); err != nil {
+			l.Error(err, "Failed to create cleanup job")
+			return ctrl.Result{Requeue: true}, nil
+		}
 
-		// Remove the finalizer regardless of cleanup job status
+		// Only remove finalizer after cleanup job is created successfully
 		controllerutil.RemoveFinalizer(clusterImageExport, clusterImageExportFinalizer)
 		if err := r.Update(ctx, clusterImageExport); err != nil {
 			if errors.IsNotFound(err) {
-				// CRD is already gone, which is fine
+				// Object is already gone, which is fine
 				return ctrl.Result{}, nil
 			}
 			return ctrl.Result{}, err
@@ -367,7 +376,7 @@ func (r *ClusterImageExportReconciler) deleteAssociatedClusterImages(ctx context
 	return nil
 }
 
-func (r *ClusterImageExportReconciler) runCleanupJob(ctx context.Context, clusterImageExport *raczylocomv1.ClusterImageExport) {
+func (r *ClusterImageExportReconciler) runCleanupJob(ctx context.Context, clusterImageExport *raczylocomv1.ClusterImageExport) error {
 	l := log.FromContext(ctx)
 	normalisedImageName := "cleanup-" + shared.NormalizeImageName(clusterImageExport.Name)
 
@@ -413,11 +422,12 @@ func (r *ClusterImageExportReconciler) runCleanupJob(ctx context.Context, cluste
 
 	cleanupJob := shared.CreateJob(jobParams, func(raczylocomv1.ClusterImageExport) []string { return nil })
 
-	// Try to create the cleanup job but don't block on errors
+	// Try to create the cleanup job
 	if err := r.Create(ctx, cleanupJob); err != nil {
-		l.Error(err, "Failed to create cleanup job, continuing with deletion anyway")
+		l.Error(err, "Failed to create cleanup job")
-		return
+		return err
 	}
 
 	l.Info("Created cleanup job with retry limit and TTL")
+	return nil
 }