Merge pull request #85115 from aramase/azure-disk-lock

azure: remove disk locks per vm during attach/detach

staging/src/k8s.io/legacy-cloud-providers/azure/BUILD

@@ -29,6 +29,7 @@ go_library(
         "azure_standard.go",
         "azure_storage.go",
         "azure_storageaccount.go",
+        "azure_utils.go",
         "azure_vmsets.go",
         "azure_vmss.go",
         "azure_vmss_cache.go",
@@ -76,7 +77,6 @@ go_library(
         "//vendor/github.com/Azure/go-autorest/autorest/to:go_default_library",
         "//vendor/github.com/rubiojr/go-vhd/vhd:go_default_library",
         "//vendor/k8s.io/klog:go_default_library",
-        "//vendor/k8s.io/utils/keymutex:go_default_library",
         "//vendor/k8s.io/utils/net:go_default_library",
         "//vendor/sigs.k8s.io/yaml:go_default_library",
     ],
@@ -98,6 +98,7 @@ go_test(
         "azure_storage_test.go",
         "azure_storageaccount_test.go",
         "azure_test.go",
+        "azure_utils_test.go",
         "azure_vmss_cache_test.go",
         "azure_vmss_test.go",
         "azure_wrap_test.go",

staging/src/k8s.io/legacy-cloud-providers/azure/azure.go

@@ -614,6 +614,7 @@ func initDiskControllers(az *Cloud) error {
 		resourceGroup:         az.ResourceGroup,
 		subscriptionID:        az.SubscriptionID,
 		cloud:                 az,
+		vmLockMap:             newLockMap(),
 	}
 
 	az.BlobDiskController = &BlobDiskController{common: common}

staging/src/k8s.io/legacy-cloud-providers/azure/azure_controller_common.go

@@ -33,7 +33,6 @@ import (
 	cloudprovider "k8s.io/cloud-provider"
 	volerr "k8s.io/cloud-provider/volume/errors"
 	"k8s.io/klog"
-	"k8s.io/utils/keymutex"
 )
 
 const (
@@ -58,9 +57,6 @@ var defaultBackOff = kwait.Backoff{
 	Jitter:   0.0,
 }
 
-// acquire lock to attach/detach disk in one node
-var diskOpMutex = keymutex.NewHashed(0)
-
 type controllerCommon struct {
 	subscriptionID        string
 	location              string
@@ -68,7 +64,9 @@ type controllerCommon struct {
 	resourceGroup         string
 	// store disk URI when disk is in attaching or detaching process
 	diskAttachDetachMap sync.Map
-	cloud               *Cloud
+	// vm disk map used to lock per vm update calls
+	vmLockMap *lockMap
+	cloud     *Cloud
 }
 
 // getNodeVMSet gets the VMSet interface based on config.VMType and the real virtual machine type.
@@ -144,8 +142,8 @@ func (c *controllerCommon) AttachDisk(isManagedDisk bool, diskName, diskURI stri
 		return -1, fmt.Errorf("failed to get azure instance id for node %q (%v)", nodeName, err)
 	}
 
-	diskOpMutex.LockKey(instanceid)
+	c.vmLockMap.LockEntry(strings.ToLower(string(nodeName)))
-	defer diskOpMutex.UnlockKey(instanceid)
+	defer c.vmLockMap.UnlockEntry(strings.ToLower(string(nodeName)))
 
 	lun, err := c.GetNextDiskLun(nodeName)
 	if err != nil {
@@ -161,7 +159,7 @@ func (c *controllerCommon) AttachDisk(isManagedDisk bool, diskName, diskURI stri
 
 // DetachDisk detaches a disk from host. The vhd can be identified by diskName or diskURI.
 func (c *controllerCommon) DetachDisk(diskName, diskURI string, nodeName types.NodeName) error {
-	instanceid, err := c.cloud.InstanceID(context.TODO(), nodeName)
+	_, err := c.cloud.InstanceID(context.TODO(), nodeName)
 	if err != nil {
 		if err == cloudprovider.InstanceNotFound {
 			// if host doesn't exist, no need to detach
@@ -181,20 +179,20 @@ func (c *controllerCommon) DetachDisk(diskName, diskURI string, nodeName types.N
 	klog.V(2).Infof("detach %v from node %q", diskURI, nodeName)
 
 	// make the lock here as small as possible
-	diskOpMutex.LockKey(instanceid)
+	c.vmLockMap.LockEntry(strings.ToLower(string(nodeName)))
 	c.diskAttachDetachMap.Store(strings.ToLower(diskURI), "detaching")
 	resp, err := vmset.DetachDisk(diskName, diskURI, nodeName)
 	c.diskAttachDetachMap.Delete(strings.ToLower(diskURI))
-	diskOpMutex.UnlockKey(instanceid)
+	c.vmLockMap.UnlockEntry(strings.ToLower(string(nodeName)))
 
 	if c.cloud.CloudProviderBackoff && shouldRetryHTTPRequest(resp, err) {
 		klog.V(2).Infof("azureDisk - update backing off: detach disk(%s, %s), err: %v", diskName, diskURI, err)
 		retryErr := kwait.ExponentialBackoff(c.cloud.RequestBackoff(), func() (bool, error) {
-			diskOpMutex.LockKey(instanceid)
+			c.vmLockMap.LockEntry(strings.ToLower(string(nodeName)))
 			c.diskAttachDetachMap.Store(strings.ToLower(diskURI), "detaching")
 			resp, err := vmset.DetachDisk(diskName, diskURI, nodeName)
 			c.diskAttachDetachMap.Delete(strings.ToLower(diskURI))
-			diskOpMutex.UnlockKey(instanceid)
+			c.vmLockMap.UnlockEntry(strings.ToLower(string(nodeName)))
 			return c.cloud.processHTTPRetryResponse(nil, "", resp, err)
 		})
 		if retryErr != nil {

staging/src/k8s.io/legacy-cloud-providers/azure/azure_controller_common_test.go

@@ -68,6 +68,7 @@ func TestCommonAttachDisk(t *testing.T) {
 			resourceGroup:         testCloud.ResourceGroup,
 			subscriptionID:        testCloud.SubscriptionID,
 			cloud:                 testCloud,
+			vmLockMap:             newLockMap(),
 		}
 		diskURI := fmt.Sprintf("/subscriptions/%s/resourceGroups/%s/providers/Microsoft.Compute/disks/disk-name",
 			testCloud.SubscriptionID, testCloud.ResourceGroup)
@@ -116,6 +117,7 @@ func TestCommonDetachDisk(t *testing.T) {
 			resourceGroup:         testCloud.ResourceGroup,
 			subscriptionID:        testCloud.SubscriptionID,
 			cloud:                 testCloud,
+			vmLockMap:             newLockMap(),
 		}
 		diskURI := fmt.Sprintf("/subscriptions/%s/resourceGroups/%s/providers/Microsoft.Compute/disks/disk-name",
 			testCloud.SubscriptionID, testCloud.ResourceGroup)
@@ -156,6 +158,7 @@ func TestGetDiskLun(t *testing.T) {
 			resourceGroup:         testCloud.ResourceGroup,
 			subscriptionID:        testCloud.SubscriptionID,
 			cloud:                 testCloud,
+			vmLockMap:             newLockMap(),
 		}
 		setTestVirtualMachines(testCloud, map[string]string{"vm1": "PowerState/Running"}, false)
 
@@ -194,6 +197,7 @@ func TestGetNextDiskLun(t *testing.T) {
 			resourceGroup:         testCloud.ResourceGroup,
 			subscriptionID:        testCloud.SubscriptionID,
 			cloud:                 testCloud,
+			vmLockMap:             newLockMap(),
 		}
 		setTestVirtualMachines(testCloud, map[string]string{"vm1": "PowerState/Running"}, test.isDataDisksFull)
 
@@ -235,6 +239,7 @@ func TestDisksAreAttached(t *testing.T) {
 			resourceGroup:         testCloud.ResourceGroup,
 			subscriptionID:        testCloud.SubscriptionID,
 			cloud:                 testCloud,
+			vmLockMap:             newLockMap(),
 		}
 		setTestVirtualMachines(testCloud, map[string]string{"vm1": "PowerState/Running"}, false)
 

staging/src/k8s.io/legacy-cloud-providers/azure/azure_utils.go

@@ -0,0 +1,71 @@
+// +build !providerless
+
+/*
+Copyright 2019 The Kubernetes Authors.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+*/
+
+package azure
+
+import (
+	"sync"
+)
+
+// lockMap used to lock on entries
+type lockMap struct {
+	sync.Mutex
+	mutexMap map[string]*sync.Mutex
+}
+
+// NewLockMap returns a new lock map
+func newLockMap() *lockMap {
+	return &lockMap{
+		mutexMap: make(map[string]*sync.Mutex),
+	}
+}
+
+// LockEntry acquires a lock associated with the specific entry
+func (lm *lockMap) LockEntry(entry string) {
+	lm.Lock()
+	// check if entry does not exists, then add entry
+	if _, exists := lm.mutexMap[entry]; !exists {
+		lm.addEntry(entry)
+	}
+
+	lm.Unlock()
+	lm.lockEntry(entry)
+}
+
+// UnlockEntry release the lock associated with the specific entry
+func (lm *lockMap) UnlockEntry(entry string) {
+	lm.Lock()
+	defer lm.Unlock()
+
+	if _, exists := lm.mutexMap[entry]; !exists {
+		return
+	}
+	lm.unlockEntry(entry)
+}
+
+func (lm *lockMap) addEntry(entry string) {
+	lm.mutexMap[entry] = &sync.Mutex{}
+}
+
+func (lm *lockMap) lockEntry(entry string) {
+	lm.mutexMap[entry].Lock()
+}
+
+func (lm *lockMap) unlockEntry(entry string) {
+	lm.mutexMap[entry].Unlock()
+}

staging/src/k8s.io/legacy-cloud-providers/azure/azure_utils_test.go

@@ -0,0 +1,85 @@
+// +build !providerless
+
+/*
+Copyright 2019 The Kubernetes Authors.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+*/
+
+package azure
+
+import (
+	"testing"
+	"time"
+)
+
+func TestSimpleLockEntry(t *testing.T) {
+	testLockMap := newLockMap()
+
+	callbackChan1 := make(chan interface{})
+	go testLockMap.lockAndCallback(t, "entry1", callbackChan1)
+	ensureCallbackHappens(t, callbackChan1)
+}
+
+func TestSimpleLockUnlockEntry(t *testing.T) {
+	testLockMap := newLockMap()
+
+	callbackChan1 := make(chan interface{})
+	go testLockMap.lockAndCallback(t, "entry1", callbackChan1)
+	ensureCallbackHappens(t, callbackChan1)
+	testLockMap.UnlockEntry("entry1")
+}
+
+func TestConcurrentLockEntry(t *testing.T) {
+	testLockMap := newLockMap()
+
+	callbackChan1 := make(chan interface{})
+	callbackChan2 := make(chan interface{})
+
+	go testLockMap.lockAndCallback(t, "entry1", callbackChan1)
+	ensureCallbackHappens(t, callbackChan1)
+
+	go testLockMap.lockAndCallback(t, "entry1", callbackChan2)
+	ensureNoCallback(t, callbackChan2)
+
+	testLockMap.UnlockEntry("entry1")
+	ensureCallbackHappens(t, callbackChan2)
+	testLockMap.UnlockEntry("entry1")
+}
+
+func (lm *lockMap) lockAndCallback(t *testing.T, entry string, callbackChan chan<- interface{}) {
+	lm.LockEntry(entry)
+	callbackChan <- true
+}
+
+var callbackTimeout = 2 * time.Second
+
+func ensureCallbackHappens(t *testing.T, callbackChan <-chan interface{}) bool {
+	select {
+	case <-callbackChan:
+		return true
+	case <-time.After(callbackTimeout):
+		t.Fatalf("timed out waiting for callback")
+		return false
+	}
+}
+
+func ensureNoCallback(t *testing.T, callbackChan <-chan interface{}) bool {
+	select {
+	case <-callbackChan:
+		t.Fatalf("unexpected callback")
+		return false
+	case <-time.After(callbackTimeout):
+		return true
+	}
+}