feat(HA): adding antiaffinity in the controller deployment

so that no two pods get scheduled on the same node. Also keeping
the default replica to 1, if HA feature is required, we can change
replica count to 2(or more).

Signed-off-by: Pawan <pawan@mayadata.io>

deploy/zfs-operator.yaml

@@ -115,13 +115,23 @@ spec:
       app: openebs-zfs-controller
       role: openebs-zfs
   serviceName: "openebs-zfs"
-  replicas: 2
+  replicas: 1
   template:
     metadata:
       labels:
         app: openebs-zfs-controller
         role: openebs-zfs
     spec:
+      affinity:
+        podAntiAffinity:
+          requiredDuringSchedulingIgnoredDuringExecution:
+          - labelSelector:
+              matchExpressions:
+              - key: app
+                operator: In
+                values:
+                - openebs-zfs-controller
+            topologyKey: "kubernetes.io/hostname"
       priorityClassName: system-cluster-critical
       serviceAccount: openebs-zfs-controller-sa
       containers: