Refactor instance locking mechanism to use per-instance locks for concurrency

pkg/manager/manager.go

@@ -40,7 +40,8 @@ type instanceManager struct {
 	localNodeName   string // Name of the local node
 
 	// Synchronization
-	operationMu  sync.Mutex // Protects start/stop/update/delete/restart operations
+	operationMu   sync.Mutex // DEPRECATED: Use instanceLocks for per-instance operations
+	instanceLocks sync.Map   // map[string]*sync.Mutex - per-instance locks for concurrent operations
 	shutdownOnce  sync.Once
 }
 
@@ -287,3 +288,20 @@ func (im *instanceManager) getNodeForInstance(inst *instance.Instance) *config.N
 
 	return nil
 }
+
+// lockInstance returns the lock for a specific instance, creating one if needed.
+// This allows concurrent operations on different instances while preventing
+// concurrent operations on the same instance.
+func (im *instanceManager) lockInstance(name string) *sync.Mutex {
+	lock, _ := im.instanceLocks.LoadOrStore(name, &sync.Mutex{})
+	return lock.(*sync.Mutex)
+}
+
+// unlockAndCleanup unlocks the instance lock and removes it from the map.
+// This should only be called when deleting an instance to prevent memory leaks.
+func (im *instanceManager) unlockAndCleanup(name string) {
+	if lock, ok := im.instanceLocks.Load(name); ok {
+		lock.(*sync.Mutex).Unlock()
+		im.instanceLocks.Delete(name)
+	}
+}

pkg/manager/operations.go

@@ -240,9 +240,10 @@ func (im *instanceManager) UpdateInstance(name string, options *instance.Options
 		return nil, err
 	}
 
-	// Lock for local instance operations to prevent races
-	im.operationMu.Lock()
-	defer im.operationMu.Unlock()
+	// Lock this specific instance only
+	lock := im.lockInstance(name)
+	lock.Lock()
+	defer lock.Unlock()
 
 	// Handle port changes
 	oldPort := inst.GetPort()
@@ -330,9 +331,10 @@ func (im *instanceManager) DeleteInstance(name string) error {
 		return nil
 	}
 
-	// Lock for local instance operations to prevent races
-	im.operationMu.Lock()
-	defer im.operationMu.Unlock()
+	// Lock this specific instance and clean up the lock on completion
+	lock := im.lockInstance(name)
+	lock.Lock()
+	defer im.unlockAndCleanup(name)
 
 	if inst.IsRunning() {
 		return fmt.Errorf("instance with name %s is still running, stop it before deleting", name)
@@ -376,9 +378,10 @@ func (im *instanceManager) StartInstance(name string) (*instance.Instance, error
 		return inst, nil
 	}
 
-	// Lock for local instance operations to prevent races
-	im.operationMu.Lock()
-	defer im.operationMu.Unlock()
+	// Lock this specific instance only
+	lock := im.lockInstance(name)
+	lock.Lock()
+	defer lock.Unlock()
 
 	if inst.IsRunning() {
 		return inst, fmt.Errorf("instance with name %s is already running", name)
@@ -438,9 +441,10 @@ func (im *instanceManager) StopInstance(name string) (*instance.Instance, error)
 		return inst, nil
 	}
 
-	// Lock for local instance operations to prevent races
-	im.operationMu.Lock()
-	defer im.operationMu.Unlock()
+	// Lock this specific instance only
+	lock := im.lockInstance(name)
+	lock.Lock()
+	defer lock.Unlock()
 
 	if !inst.IsRunning() {
 		return inst, fmt.Errorf("instance with name %s is already stopped", name)
@@ -479,9 +483,10 @@ func (im *instanceManager) RestartInstance(name string) (*instance.Instance, err
 		return inst, nil
 	}
 
-	// Lock for the entire restart operation to ensure atomicity
-	im.operationMu.Lock()
-	defer im.operationMu.Unlock()
+	// Lock this specific instance for the entire restart operation to ensure atomicity
+	lock := im.lockInstance(name)
+	lock.Lock()
+	defer lock.Unlock()
 
 	// Stop the instance
 	if inst.IsRunning() {