Cluster Node App Instance Affinity

Expand DesignatedNodeId to allow for multiple affinity types.
Preferred will attempt to place on requested node but not guaranteed.
	App will also failover upon node failure.
Required will only run app on requested node id, failover disabled.

Signed-off-by: Andrew Durbin <andrewd@zededa.com>

Author: andrewd-zededa

proto/config/appconfig.proto

@@ -92,6 +92,13 @@ message SnapshotConfig {
   repeated SnapshotDesc snapshots = 4;
 }
 
+// Type of app instance affinity
+enum AffinityType {
+  AFFINITY_TYPE_UNSPECIFIED = 0;
+  AFFINITY_TYPE_PREFERRED = 1; // Node ID preferred at app instance boot time
+  AFFINITY_TYPE_REQUIRED = 2; // Node ID required at app instance boot time
+}
+
 // The complete configuration for an Application Instance
 // When changing key fields such as the drives/volumeRefs or the number
 // of interfaces, the controller is required to issue a purge command i.e.,
@@ -217,6 +224,11 @@ message AppInstanceConfig {
 
   // AppRuntimeType - to indicate the types of Application Runtime Deployment
   AppRuntimeType runtime_type = 28;
+
+  // Affinity is used for cluster nodes to determine
+  // preferred or required node scheduling for app instances.
+  // Node Id for scheduling is defined in designated_node_id.
+  AffinityType affinity = 29;
 }
 
 message PatchEnvelopeRef {