makes the replica count **configurable** and sets a sensible default of **2 replicas**

Author: jianzhangbjz

helm/olmv1/templates/deployment-olmv1-system-catalogd-controller-manager.yml

@@ -12,11 +12,11 @@ metadata:
   namespace: {{ .Values.namespaces.olmv1.name }}
 spec:
   minReadySeconds: 5
-  replicas: 1
+  replicas: {{ .Values.options.catalogd.deployment.replicas }}
   strategy:
     type: RollingUpdate
     rollingUpdate:
-      maxSurge: 1          # Allow temporary 2 pods (1 + 1) for zero-downtime updates
+      maxSurge: 1          # Allow temporary extra pod (e.g., 3 pods with 2 replicas) for zero-downtime updates
       maxUnavailable: 0    # Never allow pods to be unavailable during updates
   selector:
     matchLabels:

helm/olmv1/templates/deployment-olmv1-system-operator-controller-controller-manager.yml

@@ -11,11 +11,11 @@ metadata:
   name: operator-controller-controller-manager
   namespace: {{ .Values.namespaces.olmv1.name }}
 spec:
-  replicas: 1
+  replicas: {{ .Values.options.operatorController.deployment.replicas }}
   strategy:
     type: RollingUpdate
     rollingUpdate:
-      maxSurge: 1          # Allow temporary 2 pods (1 + 1) for zero-downtime updates
+      maxSurge: 1          # Allow temporary extra pod (e.g., 3 pods with 2 replicas) for zero-downtime updates
       maxUnavailable: 0    # Never allow pods to be unavailable during updates
   selector:
     matchLabels:

helm/olmv1/values.yaml

@@ -8,6 +8,7 @@ options:
     enabled: true
     deployment:
       image: quay.io/operator-framework/operator-controller:devel
+      replicas: 2
       extraArguments: []
     features:
       enabled: []
@@ -19,6 +20,7 @@ options:
     enabled: true
     deployment:
       image: quay.io/operator-framework/catalogd:devel
+      replicas: 2
       extraArguments: []
     features:
       enabled: []

manifests/experimental-e2e.yaml

@@ -2107,11 +2107,11 @@ metadata:
   namespace: olmv1-system
 spec:
   minReadySeconds: 5
-  replicas: 1
+  replicas: 2
   strategy:
     type: RollingUpdate
     rollingUpdate:
-      maxSurge: 1          # Allow temporary 2 pods (1 + 1) for zero-downtime updates
+      maxSurge: 1          # Allow temporary extra pod (e.g., 3 pods with 2 replicas) for zero-downtime updates
       maxUnavailable: 0    # Never allow pods to be unavailable during updates
   selector:
     matchLabels:
@@ -2258,11 +2258,11 @@ metadata:
   name: operator-controller-controller-manager
   namespace: olmv1-system
 spec:
-  replicas: 1
+  replicas: 2
   strategy:
     type: RollingUpdate
     rollingUpdate:
-      maxSurge: 1          # Allow temporary 2 pods (1 + 1) for zero-downtime updates
+      maxSurge: 1          # Allow temporary extra pod (e.g., 3 pods with 2 replicas) for zero-downtime updates
       maxUnavailable: 0    # Never allow pods to be unavailable during updates
   selector:
     matchLabels:

manifests/experimental.yaml

@@ -2032,11 +2032,11 @@ metadata:
   namespace: olmv1-system
 spec:
   minReadySeconds: 5
-  replicas: 1
+  replicas: 2
   strategy:
     type: RollingUpdate
     rollingUpdate:
-      maxSurge: 1          # Allow temporary 2 pods (1 + 1) for zero-downtime updates
+      maxSurge: 1          # Allow temporary extra pod (e.g., 3 pods with 2 replicas) for zero-downtime updates
       maxUnavailable: 0    # Never allow pods to be unavailable during updates
   selector:
     matchLabels:
@@ -2170,11 +2170,11 @@ metadata:
   name: operator-controller-controller-manager
   namespace: olmv1-system
 spec:
-  replicas: 1
+  replicas: 2
   strategy:
     type: RollingUpdate
     rollingUpdate:
-      maxSurge: 1          # Allow temporary 2 pods (1 + 1) for zero-downtime updates
+      maxSurge: 1          # Allow temporary extra pod (e.g., 3 pods with 2 replicas) for zero-downtime updates
       maxUnavailable: 0    # Never allow pods to be unavailable during updates
   selector:
     matchLabels:

manifests/standard-e2e.yaml

@@ -1795,11 +1795,11 @@ metadata:
   namespace: olmv1-system
 spec:
   minReadySeconds: 5
-  replicas: 1
+  replicas: 2
   strategy:
     type: RollingUpdate
     rollingUpdate:
-      maxSurge: 1          # Allow temporary 2 pods (1 + 1) for zero-downtime updates
+      maxSurge: 1          # Allow temporary extra pod (e.g., 3 pods with 2 replicas) for zero-downtime updates
       maxUnavailable: 0    # Never allow pods to be unavailable during updates
   selector:
     matchLabels:
@@ -1945,11 +1945,11 @@ metadata:
   name: operator-controller-controller-manager
   namespace: olmv1-system
 spec:
-  replicas: 1
+  replicas: 2
   strategy:
     type: RollingUpdate
     rollingUpdate:
-      maxSurge: 1          # Allow temporary 2 pods (1 + 1) for zero-downtime updates
+      maxSurge: 1          # Allow temporary extra pod (e.g., 3 pods with 2 replicas) for zero-downtime updates
       maxUnavailable: 0    # Never allow pods to be unavailable during updates
   selector:
     matchLabels:

manifests/standard.yaml

@@ -1720,11 +1720,11 @@ metadata:
   namespace: olmv1-system
 spec:
   minReadySeconds: 5
-  replicas: 1
+  replicas: 2
   strategy:
     type: RollingUpdate
     rollingUpdate:
-      maxSurge: 1          # Allow temporary 2 pods (1 + 1) for zero-downtime updates
+      maxSurge: 1          # Allow temporary extra pod (e.g., 3 pods with 2 replicas) for zero-downtime updates
       maxUnavailable: 0    # Never allow pods to be unavailable during updates
   selector:
     matchLabels:
@@ -1857,11 +1857,11 @@ metadata:
   name: operator-controller-controller-manager
   namespace: olmv1-system
 spec:
-  replicas: 1
+  replicas: 2
   strategy:
     type: RollingUpdate
     rollingUpdate:
-      maxSurge: 1          # Allow temporary 2 pods (1 + 1) for zero-downtime updates
+      maxSurge: 1          # Allow temporary extra pod (e.g., 3 pods with 2 replicas) for zero-downtime updates
       maxUnavailable: 0    # Never allow pods to be unavailable during updates
   selector:
     matchLabels:

test/upgrade-e2e/post_upgrade_test.go

@@ -46,12 +46,12 @@ func TestClusterCatalogUnpacking(t *testing.T) {
 	}, time.Minute, time.Second)
 
 	var managerPod corev1.Pod
-	t.Log("Waiting for only one controller-manager pod to remain")
+	t.Log("Waiting for controller-manager pods to match the desired replica count")
 	require.EventuallyWithT(t, func(ct *assert.CollectT) {
 		var managerPods corev1.PodList
 		err := c.List(ctx, &managerPods, client.MatchingLabels(managerLabelSelector))
 		require.NoError(ct, err)
-		require.Len(ct, managerPods.Items, 1)
+		require.Len(ct, managerPods.Items, int(*managerDeployment.Spec.Replicas))
 		managerPod = managerPods.Items[0]
 	}, time.Minute, time.Second)
 
@@ -221,7 +221,7 @@ func waitForDeployment(t *testing.T, ctx context.Context, controlPlaneLabel stri
 	t.Logf("Ensure the number of remaining pods equal the desired number of replicas (%d)", desiredNumReplicas)
 	require.EventuallyWithT(t, func(ct *assert.CollectT) {
 		require.NoError(ct, c.List(ctx, &managerPods, client.MatchingLabelsSelector{Selector: deploymentLabelSelector}))
-		require.Len(ct, managerPods.Items, 1)
+		require.Len(ct, managerPods.Items, int(desiredNumReplicas))
 	}, time.Minute, time.Second)
 	return &managerPods.Items[0]
 }