diff --git a/cluster/calcium/capacity.go b/cluster/calcium/capacity.go
index c686103fc..d7ef747d3 100644
--- a/cluster/calcium/capacity.go
+++ b/cluster/calcium/capacity.go
@@ -21,7 +21,7 @@ func (c *Calcium) CalculateCapacity(ctx context.Context, opts *types.DeployOptio
 		NodeCapacities: map[string]int{},
 	}
 
-	return msg, c.withNodesResourceLocked(ctx, opts.NodeFilter, func(ctx context.Context, nodeMap map[string]*types.Node) error {
+	return msg, c.withNodesPodLocked(ctx, opts.NodeFilter, func(ctx context.Context, nodeMap map[string]*types.Node) error {
 		if opts.DeployStrategy != strategy.Dummy {
 			if _, msg.NodeCapacities, err = c.doAllocResource(ctx, nodeMap, opts); err != nil {
 				logger.Errorf(ctx, "[Calcium.CalculateCapacity] doAllocResource failed: %+v", err)
diff --git a/cluster/calcium/create.go b/cluster/calcium/create.go
index ec5bc3aca..5b66488fb 100644
--- a/cluster/calcium/create.go
+++ b/cluster/calcium/create.go
@@ -97,7 +97,7 @@ func (c *Calcium) doCreateWorkloads(ctx context.Context, opts *types.DeployOptio
 						ch <- &types.CreateWorkloadMessage{Error: logger.Err(ctx, err)}
 					}
 				}()
-				return c.withNodesResourceLocked(ctx, opts.NodeFilter, func(ctx context.Context, nodeMap map[string]*types.Node) (err error) {
+				return c.withNodesPodLocked(ctx, opts.NodeFilter, func(ctx context.Context, nodeMap map[string]*types.Node) (err error) {
 					// calculate plans
 					if plans, deployMap, err = c.doAllocResource(ctx, nodeMap, opts); err != nil {
 						return err
@@ -138,7 +138,7 @@ func (c *Calcium) doCreateWorkloads(ctx context.Context, opts *types.DeployOptio
 					return
 				}
 				for nodename, rollbackIndices := range rollbackMap {
-					if e := c.withNodeResourceLocked(ctx, nodename, func(ctx context.Context, node *types.Node) error {
+					if e := c.withNodePodLocked(ctx, nodename, func(ctx context.Context, node *types.Node) error {
 						for _, plan := range plans {
 							plan.RollbackChangesOnNode(node, rollbackIndices...) // nolint:scopelint
 						}
diff --git a/cluster/calcium/dissociate.go b/cluster/calcium/dissociate.go
index be16bf066..4e3abef07 100644
--- a/cluster/calcium/dissociate.go
+++ b/cluster/calcium/dissociate.go
@@ -25,7 +25,7 @@ func (c *Calcium) DissociateWorkload(ctx context.Context, ids []string) (chan *t
 		defer close(ch)
 
 		for nodename, workloadIDs := range nodeWorkloadGroup {
-			if err := c.withNodeResourceLocked(ctx, nodename, func(ctx context.Context, node *types.Node) error {
+			if err := c.withNodePodLocked(ctx, nodename, func(ctx context.Context, node *types.Node) error {
 				for _, workloadID := range workloadIDs { // nolint:scopelint
 					msg := &types.DissociateWorkloadMessage{WorkloadID: workloadID} // nolint:scopelint
 					if err := c.withWorkloadLocked(ctx, workloadID, func(ctx context.Context, workload *types.Workload) error {
diff --git a/cluster/calcium/lock.go b/cluster/calcium/lock.go
index f1fb7228d..d64b9507f 100644
--- a/cluster/calcium/lock.go
+++ b/cluster/calcium/lock.go
@@ -31,7 +31,6 @@ func (c *Calcium) doLock(ctx context.Context, name string, timeout time.Duration
 	}()
 	rCtx, err = lock.Lock(ctx)
 	return lock, rCtx, errors.WithStack(err)
-
 }
 
 func (c *Calcium) doUnlock(ctx context.Context, lock lock.DistributedLock, msg string) error {
@@ -65,27 +64,6 @@ func (c *Calcium) withWorkloadLocked(ctx context.Context, id string, f func(cont
 	})
 }
 
-func (c *Calcium) withNodeResourceLocked(ctx context.Context, nodename string, f func(context.Context, *types.Node) error) error {
-	return c.withNodeLocked(ctx, nodename, cluster.NodeResourceLock, f)
-}
-
-func (c *Calcium) withNodeOperationLocked(ctx context.Context, nodename string, f func(context.Context, *types.Node) error) error {
-	return c.withNodeLocked(ctx, nodename, cluster.NodeOperationLock, f)
-}
-
-func (c *Calcium) withNodeLocked(ctx context.Context, nodename string, lockKeyPattern string, f func(context.Context, *types.Node) error) error {
-	nf := types.NodeFilter{
-		Includes: []string{nodename},
-		All:      true,
-	}
-	return c.withNodesLocked(ctx, nf, lockKeyPattern, func(ctx context.Context, nodes map[string]*types.Node) error {
-		if n, ok := nodes[nodename]; ok {
-			return f(ctx, n)
-		}
-		return errors.WithStack(types.ErrNodeNotExists)
-	})
-}
-
 func (c *Calcium) withWorkloadsLocked(ctx context.Context, ids []string, f func(context.Context, map[string]*types.Workload) error) error {
 	workloads := map[string]*types.Workload{}
 	locks := map[string]lock.DistributedLock{}
@@ -116,24 +94,54 @@ func (c *Calcium) withWorkloadsLocked(ctx context.Context, ids []string, f func(
 	return f(ctx, workloads)
 }
 
+func (c *Calcium) withNodePodLocked(ctx context.Context, nodename string, f func(context.Context, *types.Node) error) error {
+	nf := types.NodeFilter{
+		Includes: []string{nodename},
+		All:      true,
+	}
+	return c.withNodesPodLocked(ctx, nf, func(ctx context.Context, nodes map[string]*types.Node) error {
+		if n, ok := nodes[nodename]; ok {
+			return f(ctx, n)
+		}
+		return errors.WithStack(types.ErrNodeNotExists)
+	})
+}
+
+func (c *Calcium) withNodeOperationLocked(ctx context.Context, nodename string, f func(context.Context, *types.Node) error) error {
+	nf := types.NodeFilter{
+		Includes: []string{nodename},
+		All:      true,
+	}
+	return c.withNodesOperationLocked(ctx, nf, func(ctx context.Context, nodes map[string]*types.Node) error {
+		if n, ok := nodes[nodename]; ok {
+			return f(ctx, n)
+		}
+		return errors.WithStack(types.ErrNodeNotExists)
+	})
+}
+
 func (c *Calcium) withNodesOperationLocked(ctx context.Context, nf types.NodeFilter, f func(context.Context, map[string]*types.Node) error) error { // nolint
-	return c.withNodesLocked(ctx, nf, cluster.NodeOperationLock, f)
+	genKey := func(node *types.Node) string {
+		return fmt.Sprintf(cluster.NodeOperationLock, node.Podname, node.Name)
+	}
+	return c.withNodesLocked(ctx, nf, genKey, f)
 }
 
-func (c *Calcium) withNodesResourceLocked(ctx context.Context, nf types.NodeFilter, f func(context.Context, map[string]*types.Node) error) error {
-	return c.withNodesLocked(ctx, nf, cluster.NodeResourceLock, f)
+func (c *Calcium) withNodesPodLocked(ctx context.Context, nf types.NodeFilter, f func(context.Context, map[string]*types.Node) error) error {
+	genKey := func(node *types.Node) string {
+		return fmt.Sprintf(cluster.PodLock, node.Podname)
+	}
+	return c.withNodesLocked(ctx, nf, genKey, f)
 }
 
-// withNodesLocked will using NodeFilter `nf` to filter nodes
-// and lock the corresponding nodes for the callback function `f` to use
-func (c *Calcium) withNodesLocked(ctx context.Context, nf types.NodeFilter, lockKeyPattern string, f func(context.Context, map[string]*types.Node) error) error {
-	nodenames := []string{}
+func (c *Calcium) withNodesLocked(ctx context.Context, nf types.NodeFilter, genKey func(*types.Node) string, f func(context.Context, map[string]*types.Node) error) error {
 	nodes := map[string]*types.Node{}
 	locks := map[string]lock.DistributedLock{}
-	defer log.Debugf(ctx, "[withNodesLocked] Nodes %+v unlocked", nf)
+	lockKeys := []string{}
 	defer func() {
-		utils.Reverse(nodenames)
-		c.doUnlockAll(utils.InheritTracingInfo(ctx, context.TODO()), locks, nodenames...)
+		utils.Reverse(lockKeys)
+		c.doUnlockAll(utils.InheritTracingInfo(ctx, context.TODO()), locks, lockKeys...)
+		log.Debugf(ctx, "[withNodesLocked] keys %v unlocked", lockKeys)
 	}()
 
 	ns, err := c.filterNodes(ctx, nf)
@@ -143,19 +151,23 @@ func (c *Calcium) withNodesLocked(ctx context.Context, nf types.NodeFilter, lock
 
 	var lock lock.DistributedLock
 	for _, n := range ns {
-		lock, ctx, err = c.doLock(ctx, fmt.Sprintf(lockKeyPattern, n.Podname, n.Name), c.config.LockTimeout)
-		if err != nil {
-			return err
+		key := genKey(n)
+		if _, ok := locks[key]; !ok {
+			lock, ctx, err = c.doLock(ctx, key, c.config.LockTimeout)
+			if err != nil {
+				return err
+			}
+			log.Debugf(ctx, "[withNodesLocked] key %s locked", key)
+			locks[key] = lock
+			lockKeys = append(lockKeys, key)
 		}
-		log.Debugf(ctx, "[withNodesLocked] Node %s locked", n.Name)
-		locks[n.Name] = lock
+
 		// refresh node
 		node, err := c.GetNode(ctx, n.Name)
 		if err != nil {
 			return err
 		}
 		nodes[n.Name] = node
-		nodenames = append(nodenames, n.Name)
 	}
 	return f(ctx, nodes)
 }
diff --git a/cluster/calcium/lock_test.go b/cluster/calcium/lock_test.go
index 80cdd21cf..82afc3221 100644
--- a/cluster/calcium/lock_test.go
+++ b/cluster/calcium/lock_test.go
@@ -116,7 +116,7 @@ func TestWithWorkloadLocked(t *testing.T) {
 	assert.NoError(t, err)
 }
 
-func TestWithNodesResourceLocked(t *testing.T) {
+func TestWithNodesPodLocked(t *testing.T) {
 	c := NewTestCluster()
 	ctx := context.Background()
 	store := &storemocks.Store{}
@@ -128,17 +128,18 @@ func TestWithNodesResourceLocked(t *testing.T) {
 			Labels: map[string]string{
 				"eru": "1",
 			},
+			Podname: "test",
 		},
 		Available: true,
 	}
 	// failed by list nodes
 	store.On("GetNodesByPod", mock.Anything, mock.Anything, mock.Anything, mock.Anything).Return([]*types.Node{}, types.ErrNoETCD).Once()
-	err := c.withNodesResourceLocked(ctx, types.NodeFilter{Podname: "test", All: false}, func(ctx context.Context, nodes map[string]*types.Node) error { return nil })
+	err := c.withNodesPodLocked(ctx, types.NodeFilter{Podname: "test", All: false}, func(ctx context.Context, nodes map[string]*types.Node) error { return nil })
 	assert.Error(t, err)
 	store.On("GetNodesByPod", mock.Anything, mock.Anything, mock.Anything, mock.Anything).Return([]*types.Node{}, nil).Once()
 	// failed by filter
 	var ns map[string]*types.Node
-	err = c.withNodesResourceLocked(ctx, types.NodeFilter{Podname: "test", Labels: map[string]string{"eru": "2"}, All: false}, func(ctx context.Context, nodes map[string]*types.Node) error {
+	err = c.withNodesPodLocked(ctx, types.NodeFilter{Podname: "test", Labels: map[string]string{"eru": "2"}, All: false}, func(ctx context.Context, nodes map[string]*types.Node) error {
 		ns = nodes
 		return nil
 	})
@@ -147,7 +148,7 @@ func TestWithNodesResourceLocked(t *testing.T) {
 	store.On("GetNodesByPod", mock.Anything, mock.Anything, mock.Anything, mock.Anything).Return([]*types.Node{}, nil)
 	// failed by getnode
 	store.On("GetNode", mock.Anything, mock.Anything).Return(nil, types.ErrNoETCD).Once()
-	err = c.withNodesResourceLocked(ctx, types.NodeFilter{Podname: "test", Includes: []string{"test"}, All: false}, func(ctx context.Context, nodes map[string]*types.Node) error { return nil })
+	err = c.withNodesPodLocked(ctx, types.NodeFilter{Podname: "test", Includes: []string{"test"}, All: false}, func(ctx context.Context, nodes map[string]*types.Node) error { return nil })
 	assert.Error(t, err)
 	store.On("GetNode", mock.Anything, mock.Anything).Return(node1, nil).Once()
 	// failed by lock
@@ -156,23 +157,23 @@ func TestWithNodesResourceLocked(t *testing.T) {
 	lock.On("Unlock", mock.Anything).Return(nil)
 	// failed to get lock
 	lock.On("Lock", mock.Anything).Return(context.TODO(), types.ErrNoETCD).Once()
-	err = c.withNodesResourceLocked(ctx, types.NodeFilter{Podname: "test", Includes: []string{"test"}, All: false}, func(ctx context.Context, nodes map[string]*types.Node) error { return nil })
+	err = c.withNodesPodLocked(ctx, types.NodeFilter{Podname: "test", Includes: []string{"test"}, All: false}, func(ctx context.Context, nodes map[string]*types.Node) error { return nil })
 	assert.Error(t, err)
 	lock.On("Lock", mock.Anything).Return(context.TODO(), nil)
 	// failed by get locked node
 	store.On("GetNode", mock.Anything, mock.Anything).Return(nil, types.ErrNoETCD).Once()
-	err = c.withNodesResourceLocked(ctx, types.NodeFilter{Podname: "test", Includes: []string{"test"}, All: false}, func(ctx context.Context, nodes map[string]*types.Node) error { return nil })
+	err = c.withNodesPodLocked(ctx, types.NodeFilter{Podname: "test", Includes: []string{"test"}, All: false}, func(ctx context.Context, nodes map[string]*types.Node) error { return nil })
 	assert.Error(t, err)
 	store.On("GetNode", mock.Anything, mock.Anything).Return(node1, nil)
 	// success
-	err = c.withNodesResourceLocked(ctx, types.NodeFilter{Podname: "test", Includes: []string{"test"}, All: false}, func(ctx context.Context, nodes map[string]*types.Node) error {
+	err = c.withNodesPodLocked(ctx, types.NodeFilter{Podname: "test", Includes: []string{"test"}, All: false}, func(ctx context.Context, nodes map[string]*types.Node) error {
 		assert.Len(t, nodes, 1)
 		return nil
 	})
 	assert.NoError(t, err)
 }
 
-func TestWithNodeResourceLocked(t *testing.T) {
+func TestWithNodePodLocked(t *testing.T) {
 	c := NewTestCluster()
 	ctx := context.Background()
 	store := &storemocks.Store{}
@@ -184,6 +185,7 @@ func TestWithNodeResourceLocked(t *testing.T) {
 			Labels: map[string]string{
 				"eru": "1",
 			},
+			Podname: "test",
 		},
 		Available: true,
 	}
@@ -194,11 +196,11 @@ func TestWithNodeResourceLocked(t *testing.T) {
 	lock.On("Lock", mock.Anything).Return(context.TODO(), nil)
 	// failed by get locked node
 	store.On("GetNode", mock.Anything, mock.Anything).Return(nil, types.ErrNoETCD).Once()
-	err := c.withNodeResourceLocked(ctx, "test", func(ctx context.Context, node *types.Node) error { return nil })
+	err := c.withNodePodLocked(ctx, "test", func(ctx context.Context, node *types.Node) error { return nil })
 	assert.Error(t, err)
 	store.On("GetNode", mock.Anything, mock.Anything).Return(node1, nil)
 	// success
-	err = c.withNodeResourceLocked(ctx, "test", func(ctx context.Context, node *types.Node) error {
+	err = c.withNodePodLocked(ctx, "test", func(ctx context.Context, node *types.Node) error {
 		assert.Equal(t, node.Name, node1.Name)
 		return nil
 	})
@@ -293,7 +295,7 @@ func TestWithNodeOperationLocked(t *testing.T) {
 	})
 	assert.NoError(t, err)
 
-	err = c.withNodeResourceLocked(ctx, "test", func(ctx context.Context, node *types.Node) error {
+	err = c.withNodePodLocked(ctx, "test", func(ctx context.Context, node *types.Node) error {
 		return c.withNodeOperationLocked(ctx, node.Name, func(ctx context.Context, node *types.Node) error {
 			assert.Equal(t, node.Name, node1.Name)
 			return nil
diff --git a/cluster/calcium/node.go b/cluster/calcium/node.go
index f45b65246..3dd096a9c 100644
--- a/cluster/calcium/node.go
+++ b/cluster/calcium/node.go
@@ -30,7 +30,7 @@ func (c *Calcium) RemoveNode(ctx context.Context, nodename string) error {
 	if nodename == "" {
 		return logger.Err(ctx, errors.WithStack(types.ErrEmptyNodeName))
 	}
-	return c.withNodeResourceLocked(ctx, nodename, func(ctx context.Context, node *types.Node) error {
+	return c.withNodePodLocked(ctx, nodename, func(ctx context.Context, node *types.Node) error {
 		ws, err := c.ListNodeWorkloads(ctx, node.Name, nil)
 		if err != nil {
 			return logger.Err(ctx, err)
@@ -107,7 +107,7 @@ func (c *Calcium) SetNode(ctx context.Context, opts *types.SetNodeOptions) (*typ
 		return nil, logger.Err(ctx, err)
 	}
 	var n *types.Node
-	return n, c.withNodeResourceLocked(ctx, opts.Nodename, func(ctx context.Context, node *types.Node) error {
+	return n, c.withNodePodLocked(ctx, opts.Nodename, func(ctx context.Context, node *types.Node) error {
 		logger.Infof(ctx, "set node")
 		opts.Normalize(node)
 		n = node
diff --git a/cluster/calcium/pod.go b/cluster/calcium/pod.go
index 46e98a333..677deff2e 100644
--- a/cluster/calcium/pod.go
+++ b/cluster/calcium/pod.go
@@ -30,7 +30,7 @@ func (c *Calcium) RemovePod(ctx context.Context, podname string) error {
 		Podname: podname,
 		All:     true,
 	}
-	return c.withNodesResourceLocked(ctx, nf, func(ctx context.Context, nodes map[string]*types.Node) error {
+	return c.withNodesPodLocked(ctx, nf, func(ctx context.Context, nodes map[string]*types.Node) error {
 		// TODO dissociate workload to node
 		// TODO should remove node first
 		return logger.Err(ctx, errors.WithStack(c.store.RemovePod(ctx, podname)))
diff --git a/cluster/calcium/realloc.go b/cluster/calcium/realloc.go
index f8fb8510d..963ef5b3b 100644
--- a/cluster/calcium/realloc.go
+++ b/cluster/calcium/realloc.go
@@ -21,7 +21,7 @@ func (c *Calcium) ReallocResource(ctx context.Context, opts *types.ReallocOption
 	if err != nil {
 		return
 	}
-	return c.withNodeResourceLocked(ctx, workload.Nodename, func(ctx context.Context, node *types.Node) error {
+	return c.withNodePodLocked(ctx, workload.Nodename, func(ctx context.Context, node *types.Node) error {
 
 		return c.withWorkloadLocked(ctx, opts.ID, func(ctx context.Context, workload *types.Workload) error {
 			rrs, err := resources.MakeRequests(
diff --git a/cluster/calcium/remove.go b/cluster/calcium/remove.go
index d40a5f118..2349c8817 100644
--- a/cluster/calcium/remove.go
+++ b/cluster/calcium/remove.go
@@ -33,7 +33,7 @@ func (c *Calcium) RemoveWorkload(ctx context.Context, ids []string, force bool)
 			utils.SentryGo(func(nodename string, workloadIDs []string) func() {
 				return func() {
 					defer wg.Done()
-					if err := c.withNodeResourceLocked(ctx, nodename, func(ctx context.Context, node *types.Node) error {
+					if err := c.withNodePodLocked(ctx, nodename, func(ctx context.Context, node *types.Node) error {
 						for _, workloadID := range workloadIDs {
 							ret := &types.RemoveWorkloadMessage{WorkloadID: workloadID, Success: true, Hook: []*bytes.Buffer{}}
 							if err := c.withWorkloadLocked(ctx, workloadID, func(ctx context.Context, workload *types.Workload) error {
diff --git a/cluster/calcium/resource.go b/cluster/calcium/resource.go
index a2ec67f28..3bb0f055e 100644
--- a/cluster/calcium/resource.go
+++ b/cluster/calcium/resource.go
@@ -64,7 +64,7 @@ func (c *Calcium) NodeResource(ctx context.Context, nodename string, fix bool) (
 
 func (c *Calcium) doGetNodeResource(ctx context.Context, nodename string, fix bool) (*types.NodeResource, error) {
 	var nr *types.NodeResource
-	return nr, c.withNodeResourceLocked(ctx, nodename, func(ctx context.Context, node *types.Node) error {
+	return nr, c.withNodePodLocked(ctx, nodename, func(ctx context.Context, node *types.Node) error {
 		workloads, err := c.ListNodeWorkloads(ctx, node.Name, nil)
 		if err != nil {
 			return err
diff --git a/cluster/cluster.go b/cluster/cluster.go
index e25aec5cc..8ace051e8 100644
--- a/cluster/cluster.go
+++ b/cluster/cluster.go
@@ -30,8 +30,8 @@ const (
 	WorkloadRestart = "restart"
 	// WorkloadLock for lock workload
 	WorkloadLock = "clock_%s"
-	// NodeResourceLock for lock node resource
-	NodeResourceLock = "cnode_%s_%s"
+	// PodLock for lock pod
+	PodLock = "plock_%s"
 	// NodeOperationLock for lock node operation
 	NodeOperationLock = "cnode_op_%s_%s"
 )
diff --git a/go.mod b/go.mod
index d12b43312..85d93c9de 100644
--- a/go.mod
+++ b/go.mod
@@ -34,7 +34,7 @@ require (
 	go.etcd.io/etcd/tests/v3 v3.5.0
 	go.uber.org/automaxprocs v1.3.0
 	golang.org/x/crypto v0.0.0-20210921155107-089bfa567519
-	golang.org/x/exp v0.0.0-20220323204016-c86f0da35e87
+	golang.org/x/exp v0.0.0-20220328175248-053ad81199eb
 	golang.org/x/net v0.0.0-20211015210444-4f30a5c0130f
 	golang.org/x/sync v0.0.0-20210220032951-036812b2e83c
 	google.golang.org/grpc v1.40.0
diff --git a/go.sum b/go.sum
index a4a047247..ef5f8d531 100644
--- a/go.sum
+++ b/go.sum
@@ -665,6 +665,8 @@ golang.org/x/exp v0.0.0-20191030013958-a1ab85dbe136/go.mod h1:JXzH8nQsPlswgeRAPE
 golang.org/x/exp v0.0.0-20200908183739-ae8ad444f925/go.mod h1:1phAWC201xIgDyaFpmDeZkgf70Q4Pd/CNqfRtVPtxNw=
 golang.org/x/exp v0.0.0-20220323204016-c86f0da35e87 h1:pFVxvJFSIGjuRLaw0mTTDfxn/AMeSYzY6Y/Hr7adkVU=
 golang.org/x/exp v0.0.0-20220323204016-c86f0da35e87/go.mod h1:lgLbSvA5ygNOMpwM/9anMpWVlVJ7Z+cHWq/eFuinpGE=
+golang.org/x/exp v0.0.0-20220328175248-053ad81199eb h1:pC9Okm6BVmxEw76PUu0XUbOTQ92JX11hfvqTjAV3qxM=
+golang.org/x/exp v0.0.0-20220328175248-053ad81199eb/go.mod h1:lgLbSvA5ygNOMpwM/9anMpWVlVJ7Z+cHWq/eFuinpGE=
 golang.org/x/image v0.0.0-20190227222117-0694c2d4d067/go.mod h1:kZ7UVZpmo3dzQBMxlp+ypCbDeSB+sBbTgSJuh5dn5js=
 golang.org/x/image v0.0.0-20190802002840-cff245a6509b/go.mod h1:FeLwcggjj3mMvU+oOTbSwawSJRM1uh48EjtB4UJZlP0=
 golang.org/x/lint v0.0.0-20181026193005-c67002cb31c3/go.mod h1:UVdnD1Gm6xHRNCYTkRU2/jEulfH38KcIWyp/GAMgvoE=