github/kubernetes
4
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Rename volumesNeedDevicePath

Browse Source 
To volumesNeedUpdateFromNodeStatus - because both devicePath and uncertain
attach-ability needs to be fixed from node status.
This commit is contained in:
Jan Safranek
2023-07-11 11:15:06 +02:00
parent 7cd60df4aa
commit 1903f5aa2a
5 changed files with 37 additions and 37 deletions
Download Patch File Download Diff File Expand all files Collapse all files

46
pkg/kubelet/volumemanager/reconciler/reconciler_common.go
View File

@@ -104,24 +104,24 @@ func NewReconciler(
volumePluginMgr *volumepkg.VolumePluginMgr, volumePluginMgr *volumepkg.VolumePluginMgr,
kubeletPodsDir string) Reconciler { kubeletPodsDir string) Reconciler {
return &reconciler{ return &reconciler{
kubeClient: kubeClient, kubeClient: kubeClient,
controllerAttachDetachEnabled: controllerAttachDetachEnabled, controllerAttachDetachEnabled: controllerAttachDetachEnabled,
loopSleepDuration: loopSleepDuration, loopSleepDuration: loopSleepDuration,
waitForAttachTimeout: waitForAttachTimeout, waitForAttachTimeout: waitForAttachTimeout,
nodeName: nodeName, nodeName: nodeName,
desiredStateOfWorld: desiredStateOfWorld, desiredStateOfWorld: desiredStateOfWorld,
actualStateOfWorld: actualStateOfWorld, actualStateOfWorld: actualStateOfWorld,
populatorHasAddedPods: populatorHasAddedPods, populatorHasAddedPods: populatorHasAddedPods,
operationExecutor: operationExecutor, operationExecutor: operationExecutor,
mounter: mounter, mounter: mounter,
hostutil: hostutil, hostutil: hostutil,
skippedDuringReconstruction: map[v1.UniqueVolumeName]*globalVolumeInfo{}, skippedDuringReconstruction: map[v1.UniqueVolumeName]*globalVolumeInfo{},
volumePluginMgr: volumePluginMgr, volumePluginMgr: volumePluginMgr,
kubeletPodsDir: kubeletPodsDir, kubeletPodsDir: kubeletPodsDir,
timeOfLastSync: time.Time{}, timeOfLastSync: time.Time{},
volumesFailedReconstruction: make([]podVolume, 0), volumesFailedReconstruction: make([]podVolume, 0),
volumesNeedDevicePath: make([]v1.UniqueVolumeName, 0), volumesNeedUpdateFromNodeStatus: make([]v1.UniqueVolumeName, 0),
volumesNeedReportedInUse: make([]v1.UniqueVolumeName, 0), volumesNeedReportedInUse: make([]v1.UniqueVolumeName, 0),
} }
} }
@@ -141,11 +141,11 @@ type reconciler struct {
skippedDuringReconstruction map[v1.UniqueVolumeName]*globalVolumeInfo skippedDuringReconstruction map[v1.UniqueVolumeName]*globalVolumeInfo
kubeletPodsDir string kubeletPodsDir string
// lock protects timeOfLastSync for updating and checking // lock protects timeOfLastSync for updating and checking
timeOfLastSyncLock sync.Mutex timeOfLastSyncLock sync.Mutex
timeOfLastSync time.Time timeOfLastSync time.Time
volumesFailedReconstruction []podVolume volumesFailedReconstruction []podVolume
volumesNeedDevicePath []v1.UniqueVolumeName volumesNeedUpdateFromNodeStatus []v1.UniqueVolumeName
volumesNeedReportedInUse []v1.UniqueVolumeName volumesNeedReportedInUse []v1.UniqueVolumeName
} }
func (rc *reconciler) Run(stopCh <-chan struct{}) { func (rc *reconciler) Run(stopCh <-chan struct{}) {

4
pkg/kubelet/volumemanager/reconciler/reconciler_new.go
View File

@@ -56,10 +56,10 @@ func (rc *reconciler) reconcileNew() {
rc.cleanOrphanVolumes() rc.cleanOrphanVolumes()
} }
if len(rc.volumesNeedDevicePath) != 0 { if len(rc.volumesNeedUpdateFromNodeStatus) != 0 {
rc.updateReconstructedFromAPIServer() rc.updateReconstructedFromAPIServer()
} }
if len(rc.volumesNeedDevicePath) == 0 { if len(rc.volumesNeedUpdateFromNodeStatus) == 0 {
// ASW is fully populated only after both devicePaths and uncertain volume attach-ability // ASW is fully populated only after both devicePaths and uncertain volume attach-ability
// were reconstructed from the API server. // were reconstructed from the API server.
// This will start reconciliation of node.status.volumesInUse. // This will start reconciliation of node.status.volumesInUse.

6
pkg/kubelet/volumemanager/reconciler/reconciler_new_test.go
View File

@@ -118,14 +118,14 @@ func TestReconcileWithUpdateReconstructedFromAPIServer(t *testing.T) {
assert.False(t, reconciler.StatesHasBeenSynced()) assert.False(t, reconciler.StatesHasBeenSynced())
reconciler.volumesNeedDevicePath = append(reconciler.volumesNeedDevicePath, volumeName1, volumeName2) reconciler.volumesNeedUpdateFromNodeStatus = append(reconciler.volumesNeedUpdateFromNodeStatus, volumeName1, volumeName2)
// Act - run reconcile loop just once. // Act - run reconcile loop just once.
// "volumesNeedDevicePath" is not empty, so no unmount will be triggered. // "volumesNeedUpdateFromNodeStatus" is not empty, so no unmount will be triggered.
reconciler.reconcileNew() reconciler.reconcileNew()
// Assert // Assert
assert.True(t, reconciler.StatesHasBeenSynced()) assert.True(t, reconciler.StatesHasBeenSynced())
assert.Empty(t, reconciler.volumesNeedDevicePath) assert.Empty(t, reconciler.volumesNeedUpdateFromNodeStatus)
attachedVolumes := asw.GetAttachedVolumes() attachedVolumes := asw.GetAttachedVolumes()
assert.Equalf(t, len(attachedVolumes), 2, "two volumes in ASW expected") assert.Equalf(t, len(attachedVolumes), 2, "two volumes in ASW expected")

12
pkg/kubelet/volumemanager/reconciler/reconstruct_new.go
View File

@@ -39,7 +39,7 @@ func (rc *reconciler) readyToUnmount() bool {
// Allow unmount only when ASW device paths were corrected from node.status to prevent // Allow unmount only when ASW device paths were corrected from node.status to prevent
// calling unmount with a wrong devicePath. // calling unmount with a wrong devicePath.
if len(rc.volumesNeedDevicePath) != 0 { if len(rc.volumesNeedUpdateFromNodeStatus) != 0 {
return false return false
} }
return true return true
@@ -97,7 +97,7 @@ func (rc *reconciler) reconstructVolumes() {
// Remember to update DSW with this information. // Remember to update DSW with this information.
rc.volumesNeedReportedInUse = reconstructedVolumeNames rc.volumesNeedReportedInUse = reconstructedVolumeNames
// Remember to update devicePath from node.status.volumesAttached // Remember to update devicePath from node.status.volumesAttached
rc.volumesNeedDevicePath = reconstructedVolumeNames rc.volumesNeedUpdateFromNodeStatus = reconstructedVolumeNames
} }
klog.V(2).InfoS("Volume reconstruction finished") klog.V(2).InfoS("Volume reconstruction finished")
} }
@@ -183,18 +183,18 @@ func (rc *reconciler) updateReconstructedFromAPIServer() {
// Skip reconstructing devicePath from node objects if kubelet is in standalone mode. // Skip reconstructing devicePath from node objects if kubelet is in standalone mode.
// Such kubelet is not expected to mount any attachable volume or Secrets / ConfigMap. // Such kubelet is not expected to mount any attachable volume or Secrets / ConfigMap.
klog.V(2).InfoS("Skipped reconstruction of DevicePaths from node.status in standalone mode") klog.V(2).InfoS("Skipped reconstruction of DevicePaths from node.status in standalone mode")
rc.volumesNeedDevicePath = nil rc.volumesNeedUpdateFromNodeStatus = nil
return return
} }
node, fetchErr := rc.kubeClient.CoreV1().Nodes().Get(context.TODO(), string(rc.nodeName), metav1.GetOptions{}) node, fetchErr := rc.kubeClient.CoreV1().Nodes().Get(context.TODO(), string(rc.nodeName), metav1.GetOptions{})
if fetchErr != nil { if fetchErr != nil {
// This may repeat few times per second until kubelet is able to read its own status for the first time. // This may repeat few times per second until kubelet is able to read its own status for the first time.
klog.V(2).ErrorS(fetchErr, "Failed to get Node status to reconstruct device paths") klog.V(4).ErrorS(fetchErr, "Failed to get Node status to reconstruct device paths")
return return
} }
for _, volumeID := range rc.volumesNeedDevicePath { for _, volumeID := range rc.volumesNeedUpdateFromNodeStatus {
attachable := false attachable := false
for _, attachedVolume := range node.Status.VolumesAttached { for _, attachedVolume := range node.Status.VolumesAttached {
if volumeID != attachedVolume.Name { if volumeID != attachedVolume.Name {
@@ -208,5 +208,5 @@ func (rc *reconciler) updateReconstructedFromAPIServer() {
} }
klog.V(2).InfoS("DevicePaths of reconstructed volumes updated") klog.V(2).InfoS("DevicePaths of reconstructed volumes updated")
rc.volumesNeedDevicePath = nil rc.volumesNeedUpdateFromNodeStatus = nil
} }

6
pkg/kubelet/volumemanager/reconciler/reconstruct_new_test.go
View File

@@ -117,8 +117,8 @@ func TestReconstructVolumes(t *testing.T) {
for i := range tc.expectedVolumesNeedDevicePath { for i := range tc.expectedVolumesNeedDevicePath {
expectedVolumes[i] = v1.UniqueVolumeName(tc.expectedVolumesNeedDevicePath[i]) expectedVolumes[i] = v1.UniqueVolumeName(tc.expectedVolumesNeedDevicePath[i])
} }
if !reflect.DeepEqual(expectedVolumes, rcInstance.volumesNeedDevicePath) { if !reflect.DeepEqual(expectedVolumes, rcInstance.volumesNeedUpdateFromNodeStatus) {
t.Errorf("Expected expectedVolumesNeedDevicePath:\n%v\n got:\n%v", expectedVolumes, rcInstance.volumesNeedDevicePath) t.Errorf("Expected expectedVolumesNeedDevicePath:\n%v\n got:\n%v", expectedVolumes, rcInstance.volumesNeedUpdateFromNodeStatus)
} }
expectedVolumes = make([]v1.UniqueVolumeName, len(tc.expectedVolumesNeedReportedInUse)) expectedVolumes = make([]v1.UniqueVolumeName, len(tc.expectedVolumesNeedReportedInUse))
@@ -333,7 +333,7 @@ func TestReconstructVolumesMount(t *testing.T) {
return true return true
} }
// Mark devices paths as reconciled to allow unmounting of volumes. // Mark devices paths as reconciled to allow unmounting of volumes.
rcInstance.volumesNeedDevicePath = nil rcInstance.volumesNeedUpdateFromNodeStatus = nil
// Act 2 - reconcile once // Act 2 - reconcile once
rcInstance.reconcileNew() rcInstance.reconcileNew()