velero/pkg/controller/data_download_controller.go

933 lines
36 KiB
Go

/*
Copyright The Velero Contributors.
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 controller
import (
"context"
"fmt"
"strings"
"time"
"github.com/pkg/errors"
"github.com/sirupsen/logrus"
v1 "k8s.io/api/core/v1"
apierrors "k8s.io/apimachinery/pkg/api/errors"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/apimachinery/pkg/types"
"k8s.io/apimachinery/pkg/util/wait"
"k8s.io/client-go/kubernetes"
"k8s.io/utils/clock"
ctrl "sigs.k8s.io/controller-runtime"
"sigs.k8s.io/controller-runtime/pkg/builder"
"sigs.k8s.io/controller-runtime/pkg/client"
"sigs.k8s.io/controller-runtime/pkg/controller/controllerutil"
"sigs.k8s.io/controller-runtime/pkg/event"
"sigs.k8s.io/controller-runtime/pkg/manager"
"sigs.k8s.io/controller-runtime/pkg/predicate"
"sigs.k8s.io/controller-runtime/pkg/reconcile"
"github.com/vmware-tanzu/velero/pkg/apis/velero/shared"
velerov1api "github.com/vmware-tanzu/velero/pkg/apis/velero/v1"
velerov2alpha1api "github.com/vmware-tanzu/velero/pkg/apis/velero/v2alpha1"
"github.com/vmware-tanzu/velero/pkg/constant"
datamover "github.com/vmware-tanzu/velero/pkg/datamover"
"github.com/vmware-tanzu/velero/pkg/datapath"
"github.com/vmware-tanzu/velero/pkg/exposer"
"github.com/vmware-tanzu/velero/pkg/metrics"
"github.com/vmware-tanzu/velero/pkg/nodeagent"
"github.com/vmware-tanzu/velero/pkg/uploader"
"github.com/vmware-tanzu/velero/pkg/util"
"github.com/vmware-tanzu/velero/pkg/util/kube"
)
// DataDownloadReconciler reconciles a DataDownload object
type DataDownloadReconciler struct {
client client.Client
kubeClient kubernetes.Interface
mgr manager.Manager
logger logrus.FieldLogger
Clock clock.WithTickerAndDelayedExecution
restoreExposer exposer.GenericRestoreExposer
nodeName string
dataPathMgr *datapath.Manager
restorePVCConfig nodeagent.RestorePVC
podResources v1.ResourceRequirements
preparingTimeout time.Duration
metrics *metrics.ServerMetrics
}
func NewDataDownloadReconciler(client client.Client, mgr manager.Manager, kubeClient kubernetes.Interface, dataPathMgr *datapath.Manager,
restorePVCConfig nodeagent.RestorePVC, podResources v1.ResourceRequirements, nodeName string, preparingTimeout time.Duration,
logger logrus.FieldLogger, metrics *metrics.ServerMetrics) *DataDownloadReconciler {
return &DataDownloadReconciler{
client: client,
kubeClient: kubeClient,
mgr: mgr,
logger: logger.WithField("controller", "DataDownload"),
Clock: &clock.RealClock{},
nodeName: nodeName,
restoreExposer: exposer.NewGenericRestoreExposer(kubeClient, logger),
restorePVCConfig: restorePVCConfig,
dataPathMgr: dataPathMgr,
podResources: podResources,
preparingTimeout: preparingTimeout,
metrics: metrics,
}
}
// +kubebuilder:rbac:groups=velero.io,resources=datadownloads,verbs=get;list;watch;create;update;patch;delete
// +kubebuilder:rbac:groups=velero.io,resources=datadownloads/status,verbs=get;update;patch
// +kubebuilder:rbac:groups="",resources=pods,verbs=get
// +kubebuilder:rbac:groups="",resources=persistentvolumes,verbs=get
// +kubebuilder:rbac:groups="",resources=persistentvolumerclaims,verbs=get
func (r *DataDownloadReconciler) Reconcile(ctx context.Context, req ctrl.Request) (ctrl.Result, error) {
log := r.logger.WithFields(logrus.Fields{
"controller": "datadownload",
"datadownload": req.NamespacedName,
})
log.Infof("Reconcile %s", req.Name)
dd := &velerov2alpha1api.DataDownload{}
if err := r.client.Get(ctx, types.NamespacedName{Namespace: req.Namespace, Name: req.Name}, dd); err != nil {
if apierrors.IsNotFound(err) {
log.Warn("DataDownload not found, skip")
return ctrl.Result{}, nil
}
log.WithError(err).Error("Unable to get the DataDownload")
return ctrl.Result{}, err
}
if !datamover.IsBuiltInUploader(dd.Spec.DataMover) {
log.WithField("data mover", dd.Spec.DataMover).Info("it is not one built-in data mover which is not supported by Velero")
return ctrl.Result{}, nil
}
if r.restoreExposer == nil {
return r.errorOut(ctx, dd, errors.New("uninitialized generic exposer"), "uninitialized exposer", log)
}
// Add finalizer
// Logic for clear resources when datadownload been deleted
if dd.DeletionTimestamp.IsZero() { // add finalizer for all cr at beginning
if !isDataDownloadInFinalState(dd) && !controllerutil.ContainsFinalizer(dd, DataUploadDownloadFinalizer) {
succeeded, err := r.exclusiveUpdateDataDownload(ctx, dd, func(dd *velerov2alpha1api.DataDownload) {
controllerutil.AddFinalizer(dd, DataUploadDownloadFinalizer)
})
if err != nil {
log.Errorf("failed to add finalizer with error %s for %s/%s", err.Error(), dd.Namespace, dd.Name)
return ctrl.Result{}, err
} else if !succeeded {
log.Warnf("failed to add finalizer for %s/%s and will requeue later", dd.Namespace, dd.Name)
return ctrl.Result{Requeue: true}, nil
}
}
} else if controllerutil.ContainsFinalizer(dd, DataUploadDownloadFinalizer) && !dd.Spec.Cancel && !isDataDownloadInFinalState(dd) {
// when delete cr we need to clear up internal resources created by Velero, here we use the cancel mechanism
// to help clear up resources instead of clear them directly in case of some conflict with Expose action
log.Warnf("Cancel dd under phase %s because it is being deleted", dd.Status.Phase)
if err := UpdateDataDownloadWithRetry(ctx, r.client, req.NamespacedName, log, func(dataDownload *velerov2alpha1api.DataDownload) bool {
if dataDownload.Spec.Cancel {
return false
}
dataDownload.Spec.Cancel = true
dataDownload.Status.Message = "Cancel datadownload because it is being deleted"
return true
}); err != nil {
log.Errorf("failed to set cancel flag with error %s for %s/%s", err.Error(), dd.Namespace, dd.Name)
return ctrl.Result{}, err
}
}
if dd.Status.Phase == "" || dd.Status.Phase == velerov2alpha1api.DataDownloadPhaseNew {
log.Info("Data download starting")
if _, err := r.getTargetPVC(ctx, dd); err != nil {
log.WithField("error", err).Debugf("Cannot find target PVC for DataDownload yet. Retry later.")
return ctrl.Result{Requeue: true}, nil
}
accepted, err := r.acceptDataDownload(ctx, dd)
if err != nil {
return r.errorOut(ctx, dd, err, "error to accept the data download", log)
}
if !accepted {
log.Debug("Data download is not accepted")
return ctrl.Result{}, nil
}
log.Info("Data download is accepted")
if dd.Spec.Cancel {
log.Debugf("Data download is been canceled %s in Phase %s", dd.GetName(), dd.Status.Phase)
r.OnDataDownloadCancelled(ctx, dd.GetNamespace(), dd.GetName())
return ctrl.Result{}, nil
}
exposeParam, err := r.setupExposeParam(dd)
if err != nil {
return r.errorOut(ctx, dd, err, "failed to set exposer parameters", log)
}
// Expose() will trigger to create one pod whose volume is restored by a given volume snapshot,
// but the pod maybe is not in the same node of the current controller, so we need to return it here.
// And then only the controller who is in the same node could do the rest work.
err = r.restoreExposer.Expose(ctx, getDataDownloadOwnerObject(dd), exposeParam)
if err != nil {
if err := r.client.Get(ctx, req.NamespacedName, dd); err != nil {
if !apierrors.IsNotFound(err) {
return ctrl.Result{}, errors.Wrap(err, "getting DataUpload")
}
}
if isDataDownloadInFinalState(dd) {
log.Warnf("expose snapshot with err %v but it may caused by clean up resources in cancel action", err)
r.restoreExposer.CleanUp(ctx, getDataDownloadOwnerObject(dd))
return ctrl.Result{}, nil
} else {
return r.errorOut(ctx, dd, err, "error to expose snapshot", log)
}
}
log.Info("Restore is exposed")
// we need to get CR again for it may canceled by datadownload controller on other
// nodes when doing expose action, if detectd cancel action we need to clear up the internal
// resources created by velero during backup.
if err := r.client.Get(ctx, req.NamespacedName, dd); err != nil {
if apierrors.IsNotFound(err) {
log.Debug("Unable to find datadownload")
return ctrl.Result{}, nil
}
return ctrl.Result{}, errors.Wrap(err, "getting datadownload")
}
// we need to clean up resources as resources created in Expose it may later than cancel action or prepare time
// and need to clean up resources again
if isDataDownloadInFinalState(dd) {
r.restoreExposer.CleanUp(ctx, getDataDownloadOwnerObject(dd))
}
return ctrl.Result{}, nil
} else if dd.Status.Phase == velerov2alpha1api.DataDownloadPhaseAccepted {
if dd.Spec.Cancel {
log.Debugf("Data download is been canceled %s in Phase %s", dd.GetName(), dd.Status.Phase)
r.tryCancelAcceptedDataDownload(ctx, dd, "")
} else if peekErr := r.restoreExposer.PeekExposed(ctx, getDataDownloadOwnerObject(dd)); peekErr != nil {
r.tryCancelAcceptedDataDownload(ctx, dd, fmt.Sprintf("found a datadownload %s/%s with expose error: %s. mark it as cancel", dd.Namespace, dd.Name, peekErr))
log.Errorf("Cancel dd %s/%s because of expose error %s", dd.Namespace, dd.Name, peekErr)
} else if dd.Status.AcceptedTimestamp != nil {
if time.Since(dd.Status.AcceptedTimestamp.Time) >= r.preparingTimeout {
r.onPrepareTimeout(ctx, dd)
}
}
return ctrl.Result{}, nil
} else if dd.Status.Phase == velerov2alpha1api.DataDownloadPhasePrepared {
log.Info("Data download is prepared")
if dd.Spec.Cancel {
log.Debugf("Data download is been canceled %s in Phase %s", dd.GetName(), dd.Status.Phase)
r.OnDataDownloadCancelled(ctx, dd.GetNamespace(), dd.GetName())
return ctrl.Result{}, nil
}
asyncBR := r.dataPathMgr.GetAsyncBR(dd.Name)
if asyncBR != nil {
log.Info("Cancellable data path is already started")
return ctrl.Result{}, nil
}
result, err := r.restoreExposer.GetExposed(ctx, getDataDownloadOwnerObject(dd), r.client, r.nodeName, dd.Spec.OperationTimeout.Duration)
if err != nil {
return r.errorOut(ctx, dd, err, "restore exposer is not ready", log)
} else if result == nil {
log.Debug("Get empty restore exposer")
return ctrl.Result{}, nil
}
log.Info("Restore PVC is ready and creating data path routine")
// Need to first create file system BR and get data path instance then update data upload status
callbacks := datapath.Callbacks{
OnCompleted: r.OnDataDownloadCompleted,
OnFailed: r.OnDataDownloadFailed,
OnCancelled: r.OnDataDownloadCancelled,
OnProgress: r.OnDataDownloadProgress,
}
asyncBR, err = r.dataPathMgr.CreateMicroServiceBRWatcher(ctx, r.client, r.kubeClient, r.mgr, datapath.TaskTypeRestore,
dd.Name, dd.Namespace, result.ByPod.HostingPod.Name, result.ByPod.HostingContainer, dd.Name, callbacks, false, log)
if err != nil {
if err == datapath.ConcurrentLimitExceed {
log.Info("Data path instance is concurrent limited requeue later")
return ctrl.Result{Requeue: true, RequeueAfter: time.Second * 5}, nil
} else {
return r.errorOut(ctx, dd, err, "error to create data path", log)
}
}
if err := r.initCancelableDataPath(ctx, asyncBR, result, log); err != nil {
log.WithError(err).Errorf("Failed to init cancelable data path for %s", dd.Name)
r.closeDataPath(ctx, dd.Name)
return r.errorOut(ctx, dd, err, "error initializing data path", log)
}
// Update status to InProgress
original := dd.DeepCopy()
dd.Status.Phase = velerov2alpha1api.DataDownloadPhaseInProgress
dd.Status.StartTimestamp = &metav1.Time{Time: r.Clock.Now()}
if err := r.client.Patch(ctx, dd, client.MergeFrom(original)); err != nil {
log.WithError(err).Warnf("Failed to update datadownload %s to InProgress, will close data path and retry", dd.Name)
r.closeDataPath(ctx, dd.Name)
return ctrl.Result{Requeue: true, RequeueAfter: time.Second * 5}, nil
}
log.Info("Data download is marked as in progress")
if err := r.startCancelableDataPath(asyncBR, dd, result, log); err != nil {
log.WithError(err).Errorf("Failed to start cancelable data path for %s", dd.Name)
r.closeDataPath(ctx, dd.Name)
return r.errorOut(ctx, dd, err, "error starting data path", log)
}
return ctrl.Result{}, nil
} else if dd.Status.Phase == velerov2alpha1api.DataDownloadPhaseInProgress {
log.Info("Data download is in progress")
if dd.Spec.Cancel {
log.Info("Data download is being canceled")
asyncBR := r.dataPathMgr.GetAsyncBR(dd.Name)
if asyncBR == nil {
if r.nodeName == dd.Status.Node {
r.OnDataDownloadCancelled(ctx, dd.GetNamespace(), dd.GetName())
} else {
log.Info("Data path is not started in this node and will not canceled by current node")
}
return ctrl.Result{}, nil
}
// Update status to Canceling.
original := dd.DeepCopy()
dd.Status.Phase = velerov2alpha1api.DataDownloadPhaseCanceling
if err := r.client.Patch(ctx, dd, client.MergeFrom(original)); err != nil {
log.WithError(err).Error("error updating data download status")
return ctrl.Result{}, err
}
asyncBR.Cancel()
return ctrl.Result{}, nil
}
return ctrl.Result{}, nil
} else {
// put the finalizer remove action here for all cr will goes to the final status, we could check finalizer and do remove action in final status
// instead of intermediate state
// remove finalizer no matter whether the cr is being deleted or not for it is no longer needed when internal resources are all cleaned up
// also in final status cr won't block the direct delete of the velero namespace
if isDataDownloadInFinalState(dd) && controllerutil.ContainsFinalizer(dd, DataUploadDownloadFinalizer) {
original := dd.DeepCopy()
controllerutil.RemoveFinalizer(dd, DataUploadDownloadFinalizer)
if err := r.client.Patch(ctx, dd, client.MergeFrom(original)); err != nil {
log.WithError(err).Error("error to remove finalizer")
}
}
return ctrl.Result{}, nil
}
}
func (r *DataDownloadReconciler) initCancelableDataPath(ctx context.Context, asyncBR datapath.AsyncBR, res *exposer.ExposeResult, log logrus.FieldLogger) error {
log.Info("Init cancelable dataDownload")
if err := asyncBR.Init(ctx, nil); err != nil {
return errors.Wrap(err, "error initializing asyncBR")
}
log.Infof("async restore init for pod %s, volume %s", res.ByPod.HostingPod.Name, res.ByPod.VolumeName)
return nil
}
func (r *DataDownloadReconciler) startCancelableDataPath(asyncBR datapath.AsyncBR, dd *velerov2alpha1api.DataDownload, res *exposer.ExposeResult, log logrus.FieldLogger) error {
log.Info("Start cancelable dataDownload")
if err := asyncBR.StartRestore(dd.Spec.SnapshotID, datapath.AccessPoint{
ByPath: res.ByPod.VolumeName,
}, dd.Spec.DataMoverConfig); err != nil {
return errors.Wrapf(err, "error starting async restore for pod %s, volume %s", res.ByPod.HostingPod.Name, res.ByPod.VolumeName)
}
log.Infof("Async restore started for pod %s, volume %s", res.ByPod.HostingPod.Name, res.ByPod.VolumeName)
return nil
}
func (r *DataDownloadReconciler) OnDataDownloadCompleted(ctx context.Context, namespace string, ddName string, result datapath.Result) {
defer r.dataPathMgr.RemoveAsyncBR(ddName)
log := r.logger.WithField("datadownload", ddName)
log.Info("Async fs restore data path completed")
var dd velerov2alpha1api.DataDownload
if err := r.client.Get(ctx, types.NamespacedName{Name: ddName, Namespace: namespace}, &dd); err != nil {
log.WithError(err).Warn("Failed to get datadownload on completion")
return
}
objRef := getDataDownloadOwnerObject(&dd)
err := r.restoreExposer.RebindVolume(ctx, objRef, dd.Spec.TargetVolume.PVC, dd.Spec.TargetVolume.Namespace, dd.Spec.OperationTimeout.Duration)
if err != nil {
log.WithError(err).Error("Failed to rebind PV to target PVC on completion")
return
}
log.Info("Cleaning up exposed environment")
r.restoreExposer.CleanUp(ctx, objRef)
original := dd.DeepCopy()
dd.Status.Phase = velerov2alpha1api.DataDownloadPhaseCompleted
dd.Status.CompletionTimestamp = &metav1.Time{Time: r.Clock.Now()}
if err := r.client.Patch(ctx, &dd, client.MergeFrom(original)); err != nil {
log.WithError(err).Error("error updating data download status")
} else {
log.Infof("Data download is marked as %s", dd.Status.Phase)
r.metrics.RegisterDataDownloadSuccess(r.nodeName)
}
}
func (r *DataDownloadReconciler) OnDataDownloadFailed(ctx context.Context, namespace string, ddName string, err error) {
defer r.dataPathMgr.RemoveAsyncBR(ddName)
log := r.logger.WithField("datadownload", ddName)
log.WithError(err).Error("Async fs restore data path failed")
var dd velerov2alpha1api.DataDownload
if getErr := r.client.Get(ctx, types.NamespacedName{Name: ddName, Namespace: namespace}, &dd); getErr != nil {
log.WithError(getErr).Warn("Failed to get data download on failure")
} else {
_, _ = r.errorOut(ctx, &dd, err, "data path restore failed", log)
}
}
func (r *DataDownloadReconciler) OnDataDownloadCancelled(ctx context.Context, namespace string, ddName string) {
defer r.dataPathMgr.RemoveAsyncBR(ddName)
log := r.logger.WithField("datadownload", ddName)
log.Warn("Async fs backup data path canceled")
var dd velerov2alpha1api.DataDownload
if getErr := r.client.Get(ctx, types.NamespacedName{Name: ddName, Namespace: namespace}, &dd); getErr != nil {
log.WithError(getErr).Warn("Failed to get datadownload on cancel")
} else {
// cleans up any objects generated during the snapshot expose
r.restoreExposer.CleanUp(ctx, getDataDownloadOwnerObject(&dd))
original := dd.DeepCopy()
dd.Status.Phase = velerov2alpha1api.DataDownloadPhaseCanceled
if dd.Status.StartTimestamp.IsZero() {
dd.Status.StartTimestamp = &metav1.Time{Time: r.Clock.Now()}
}
dd.Status.CompletionTimestamp = &metav1.Time{Time: r.Clock.Now()}
if err := r.client.Patch(ctx, &dd, client.MergeFrom(original)); err != nil {
log.WithError(err).Error("error updating data download status")
} else {
r.metrics.RegisterDataDownloadCancel(r.nodeName)
}
}
}
func (r *DataDownloadReconciler) tryCancelAcceptedDataDownload(ctx context.Context, dd *velerov2alpha1api.DataDownload, message string) {
log := r.logger.WithField("datadownload", dd.Name)
log.Warn("Accepted data download is canceled")
succeeded, err := r.exclusiveUpdateDataDownload(ctx, dd, func(dataDownload *velerov2alpha1api.DataDownload) {
dataDownload.Status.Phase = velerov2alpha1api.DataDownloadPhaseCanceled
if dataDownload.Status.StartTimestamp.IsZero() {
dataDownload.Status.StartTimestamp = &metav1.Time{Time: r.Clock.Now()}
}
dataDownload.Status.CompletionTimestamp = &metav1.Time{Time: r.Clock.Now()}
if message != "" {
dataDownload.Status.Message = message
}
})
if err != nil {
log.WithError(err).Error("error updating datadownload status")
return
} else if !succeeded {
log.Warn("conflict in updating datadownload status and will try it again later")
return
}
// success update
r.metrics.RegisterDataDownloadCancel(r.nodeName)
r.restoreExposer.CleanUp(ctx, getDataDownloadOwnerObject(dd))
}
func (r *DataDownloadReconciler) OnDataDownloadProgress(ctx context.Context, namespace string, ddName string, progress *uploader.Progress) {
log := r.logger.WithField("datadownload", ddName)
var dd velerov2alpha1api.DataDownload
if err := r.client.Get(ctx, types.NamespacedName{Name: ddName, Namespace: namespace}, &dd); err != nil {
log.WithError(err).Warn("Failed to get data download on progress")
return
}
original := dd.DeepCopy()
dd.Status.Progress = shared.DataMoveOperationProgress{TotalBytes: progress.TotalBytes, BytesDone: progress.BytesDone}
if err := r.client.Patch(ctx, &dd, client.MergeFrom(original)); err != nil {
log.WithError(err).Error("Failed to update restore snapshot progress")
}
}
// SetupWithManager registers the DataDownload controller.
// The fresh new DataDownload CR first created will trigger to create one pod (long time, maybe failure or unknown status) by one of the datadownload controllers
// then the request will get out of the Reconcile queue immediately by not blocking others' CR handling, in order to finish the rest data download process we need to
// re-enqueue the previous related request once the related pod is in running status to keep going on the rest logic. and below logic will avoid handling the unwanted
// pod status and also avoid block others CR handling
func (r *DataDownloadReconciler) SetupWithManager(mgr ctrl.Manager) error {
gp := kube.NewGenericEventPredicate(func(object client.Object) bool {
dd := object.(*velerov2alpha1api.DataDownload)
return (dd.Status.Phase == velerov2alpha1api.DataDownloadPhaseAccepted)
})
s := kube.NewPeriodicalEnqueueSource(r.logger.WithField("controller", constant.ControllerDataDownload), r.client, &velerov2alpha1api.DataDownloadList{}, preparingMonitorFrequency, kube.PeriodicalEnqueueSourceOption{
Predicates: []predicate.Predicate{gp},
})
return ctrl.NewControllerManagedBy(mgr).
For(&velerov2alpha1api.DataDownload{}).
WatchesRawSource(s).
Watches(&v1.Pod{}, kube.EnqueueRequestsFromMapUpdateFunc(r.findSnapshotRestoreForPod),
builder.WithPredicates(predicate.Funcs{
UpdateFunc: func(ue event.UpdateEvent) bool {
newObj := ue.ObjectNew.(*v1.Pod)
if _, ok := newObj.Labels[velerov1api.DataDownloadLabel]; !ok {
return false
}
if newObj.Spec.NodeName == "" {
return false
}
return true
},
CreateFunc: func(event.CreateEvent) bool {
return false
},
DeleteFunc: func(de event.DeleteEvent) bool {
return false
},
GenericFunc: func(ge event.GenericEvent) bool {
return false
},
})).
Complete(r)
}
func (r *DataDownloadReconciler) findSnapshotRestoreForPod(ctx context.Context, podObj client.Object) []reconcile.Request {
pod := podObj.(*v1.Pod)
dd, err := findDataDownloadByPod(r.client, *pod)
log := r.logger.WithField("pod", pod.Name)
if err != nil {
log.WithError(err).Error("unable to get DataDownload")
return []reconcile.Request{}
} else if dd == nil {
log.Error("get empty DataDownload")
return []reconcile.Request{}
}
log = log.WithFields(logrus.Fields{
"Dataddownload": dd.Name,
})
if dd.Status.Phase != velerov2alpha1api.DataDownloadPhaseAccepted {
return []reconcile.Request{}
}
if pod.Status.Phase == v1.PodRunning {
log.Info("Preparing data download")
// we don't expect anyone else update the CR during the Prepare process
updated, err := r.exclusiveUpdateDataDownload(context.Background(), dd, r.prepareDataDownload)
if err != nil || !updated {
log.WithField("updated", updated).WithError(err).Warn("failed to update datadownload, prepare will halt for this datadownload")
return []reconcile.Request{}
}
} else if unrecoverable, reason := kube.IsPodUnrecoverable(pod, log); unrecoverable {
err := UpdateDataDownloadWithRetry(context.Background(), r.client, types.NamespacedName{Namespace: dd.Namespace, Name: dd.Name}, r.logger.WithField("datadownlad", dd.Name),
func(dataDownload *velerov2alpha1api.DataDownload) bool {
if dataDownload.Spec.Cancel {
return false
}
dataDownload.Spec.Cancel = true
dataDownload.Status.Message = fmt.Sprintf("Cancel datadownload because the exposing pod %s/%s is in abnormal status for reason %s", pod.Namespace, pod.Name, reason)
return true
})
if err != nil {
log.WithError(err).Warn("failed to cancel datadownload, and it will wait for prepare timeout")
return []reconcile.Request{}
}
log.Infof("Exposed pod is in abnormal status(reason %s) and datadownload is marked as cancel", reason)
} else {
return []reconcile.Request{}
}
request := reconcile.Request{
NamespacedName: types.NamespacedName{
Namespace: dd.Namespace,
Name: dd.Name,
},
}
return []reconcile.Request{request}
}
func (r *DataDownloadReconciler) prepareDataDownload(ssb *velerov2alpha1api.DataDownload) {
ssb.Status.Phase = velerov2alpha1api.DataDownloadPhasePrepared
ssb.Status.Node = r.nodeName
}
func (r *DataDownloadReconciler) errorOut(ctx context.Context, dd *velerov2alpha1api.DataDownload, err error, msg string, log logrus.FieldLogger) (ctrl.Result, error) {
if r.restoreExposer != nil {
r.restoreExposer.CleanUp(ctx, getDataDownloadOwnerObject(dd))
}
return ctrl.Result{}, r.updateStatusToFailed(ctx, dd, err, msg, log)
}
func (r *DataDownloadReconciler) updateStatusToFailed(ctx context.Context, dd *velerov2alpha1api.DataDownload, err error, msg string, log logrus.FieldLogger) error {
log.Infof("update data download status to %v", dd.Status.Phase)
original := dd.DeepCopy()
dd.Status.Phase = velerov2alpha1api.DataDownloadPhaseFailed
dd.Status.Message = errors.WithMessage(err, msg).Error()
dd.Status.CompletionTimestamp = &metav1.Time{Time: r.Clock.Now()}
if patchErr := r.client.Patch(ctx, dd, client.MergeFrom(original)); patchErr != nil {
log.WithError(patchErr).Error("error updating DataDownload status")
} else {
r.metrics.RegisterDataDownloadFailure(r.nodeName)
}
return err
}
func (r *DataDownloadReconciler) acceptDataDownload(ctx context.Context, dd *velerov2alpha1api.DataDownload) (bool, error) {
r.logger.Infof("Accepting data download %s", dd.Name)
// For all data download controller in each node-agent will try to update download CR, and only one controller will success,
// and the success one could handle later logic
updated := dd.DeepCopy()
updateFunc := func(datadownload *velerov2alpha1api.DataDownload) {
datadownload.Status.Phase = velerov2alpha1api.DataDownloadPhaseAccepted
datadownload.Status.AcceptedByNode = r.nodeName
datadownload.Status.AcceptedTimestamp = &metav1.Time{Time: r.Clock.Now()}
}
succeeded, err := r.exclusiveUpdateDataDownload(ctx, updated, updateFunc)
if err != nil {
return false, err
}
if succeeded {
updateFunc(dd) // If update success, it's need to update du values in memory
r.logger.WithField("DataDownload", dd.Name).Infof("This datadownload has been accepted by %s", r.nodeName)
return true, nil
}
r.logger.WithField("DataDownload", dd.Name).Info("This datadownload has been accepted by others")
return false, nil
}
func (r *DataDownloadReconciler) onPrepareTimeout(ctx context.Context, dd *velerov2alpha1api.DataDownload) {
log := r.logger.WithField("DataDownload", dd.Name)
log.Info("Timeout happened for preparing datadownload")
succeeded, err := r.exclusiveUpdateDataDownload(ctx, dd, func(dd *velerov2alpha1api.DataDownload) {
dd.Status.Phase = velerov2alpha1api.DataDownloadPhaseFailed
dd.Status.Message = "timeout on preparing data download"
})
if err != nil {
log.WithError(err).Warn("Failed to update datadownload")
return
}
if !succeeded {
log.Warn("Dataupload has been updated by others")
return
}
diags := strings.Split(r.restoreExposer.DiagnoseExpose(ctx, getDataDownloadOwnerObject(dd)), "\n")
for _, diag := range diags {
log.Warnf("[Diagnose DD expose]%s", diag)
}
r.restoreExposer.CleanUp(ctx, getDataDownloadOwnerObject(dd))
log.Info("Dataupload has been cleaned up")
r.metrics.RegisterDataDownloadFailure(r.nodeName)
}
func (r *DataDownloadReconciler) exclusiveUpdateDataDownload(ctx context.Context, dd *velerov2alpha1api.DataDownload,
updateFunc func(*velerov2alpha1api.DataDownload)) (bool, error) {
updateFunc(dd)
err := r.client.Update(ctx, dd)
if err == nil {
return true, nil
}
// it won't rollback dd in memory when error
if apierrors.IsConflict(err) {
return false, nil
} else {
return false, err
}
}
func (r *DataDownloadReconciler) getTargetPVC(ctx context.Context, dd *velerov2alpha1api.DataDownload) (*v1.PersistentVolumeClaim, error) {
return r.kubeClient.CoreV1().PersistentVolumeClaims(dd.Spec.TargetVolume.Namespace).Get(ctx, dd.Spec.TargetVolume.PVC, metav1.GetOptions{})
}
func (r *DataDownloadReconciler) closeDataPath(ctx context.Context, ddName string) {
asyncBR := r.dataPathMgr.GetAsyncBR(ddName)
if asyncBR != nil {
asyncBR.Close(ctx)
}
r.dataPathMgr.RemoveAsyncBR(ddName)
}
func (r *DataDownloadReconciler) setupExposeParam(dd *velerov2alpha1api.DataDownload) (exposer.GenericRestoreExposeParam, error) {
log := r.logger.WithField("datadownload", dd.Name)
nodeOS := string(dd.Spec.NodeOS)
if nodeOS == "" {
log.Info("nodeOS is empty in DD, fallback to linux")
nodeOS = kube.NodeOSLinux
}
if err := kube.HasNodeWithOS(context.Background(), nodeOS, r.kubeClient.CoreV1()); err != nil {
return exposer.GenericRestoreExposeParam{}, errors.Wrapf(err, "no appropriate node to run datadownload %s/%s", dd.Namespace, dd.Name)
}
hostingPodLabels := map[string]string{velerov1api.DataDownloadLabel: dd.Name}
for _, k := range util.ThirdPartyLabels {
if v, err := nodeagent.GetLabelValue(context.Background(), r.kubeClient, dd.Namespace, k, nodeOS); err != nil {
if err != nodeagent.ErrNodeAgentLabelNotFound {
log.WithError(err).Warnf("Failed to check node-agent label, skip adding host pod label %s", k)
}
} else {
hostingPodLabels[k] = v
}
}
return exposer.GenericRestoreExposeParam{
TargetPVCName: dd.Spec.TargetVolume.PVC,
TargetNamespace: dd.Spec.TargetVolume.Namespace,
HostingPodLabels: hostingPodLabels,
Resources: r.podResources,
OperationTimeout: dd.Spec.OperationTimeout.Duration,
ExposeTimeout: r.preparingTimeout,
NodeOS: nodeOS,
RestorePVCConfig: r.restorePVCConfig,
}, nil
}
func getDataDownloadOwnerObject(dd *velerov2alpha1api.DataDownload) v1.ObjectReference {
return v1.ObjectReference{
Kind: dd.Kind,
Namespace: dd.Namespace,
Name: dd.Name,
UID: dd.UID,
APIVersion: dd.APIVersion,
}
}
func findDataDownloadByPod(client client.Client, pod v1.Pod) (*velerov2alpha1api.DataDownload, error) {
if label, exist := pod.Labels[velerov1api.DataDownloadLabel]; exist {
dd := &velerov2alpha1api.DataDownload{}
err := client.Get(context.Background(), types.NamespacedName{
Namespace: pod.Namespace,
Name: label,
}, dd)
if err != nil {
return nil, errors.Wrapf(err, "error to find DataDownload by pod %s/%s", pod.Namespace, pod.Name)
}
return dd, nil
}
return nil, nil
}
func isDataDownloadInFinalState(dd *velerov2alpha1api.DataDownload) bool {
return dd.Status.Phase == velerov2alpha1api.DataDownloadPhaseFailed ||
dd.Status.Phase == velerov2alpha1api.DataDownloadPhaseCanceled ||
dd.Status.Phase == velerov2alpha1api.DataDownloadPhaseCompleted
}
func UpdateDataDownloadWithRetry(ctx context.Context, client client.Client, namespacedName types.NamespacedName, log *logrus.Entry, updateFunc func(*velerov2alpha1api.DataDownload) bool) error {
return wait.PollUntilContextCancel(ctx, time.Second, true, func(ctx context.Context) (bool, error) {
dd := &velerov2alpha1api.DataDownload{}
if err := client.Get(ctx, namespacedName, dd); err != nil {
return false, errors.Wrap(err, "getting DataDownload")
}
if updateFunc(dd) {
err := client.Update(ctx, dd)
if err != nil {
if apierrors.IsConflict(err) {
log.Warnf("failed to update datadownload for %s/%s and will retry it", dd.Namespace, dd.Name)
return false, nil
} else {
return false, errors.Wrapf(err, "error updating datadownload %s/%s", dd.Namespace, dd.Name)
}
}
}
return true, nil
})
}
var funcResumeCancellableDataRestore = (*DataDownloadReconciler).resumeCancellableDataPath
func (r *DataDownloadReconciler) AttemptDataDownloadResume(ctx context.Context, logger *logrus.Entry, ns string) error {
dataDownloads := &velerov2alpha1api.DataDownloadList{}
if err := r.client.List(ctx, dataDownloads, &client.ListOptions{Namespace: ns}); err != nil {
r.logger.WithError(errors.WithStack(err)).Error("failed to list datadownloads")
return errors.Wrapf(err, "error to list datadownloads")
}
for i := range dataDownloads.Items {
dd := &dataDownloads.Items[i]
if dd.Status.Phase == velerov2alpha1api.DataDownloadPhasePrepared {
// keep doing nothing let controller re-download the data
// the Prepared CR could be still handled by datadownload controller after node-agent restart
logger.WithField("datadownload", dd.GetName()).Debug("find a datadownload with status prepared")
} else if dd.Status.Phase == velerov2alpha1api.DataDownloadPhaseInProgress {
if dd.Status.Node != r.nodeName {
logger.WithField("dd", dd.Name).WithField("current node", r.nodeName).Infof("DD should be resumed by another node %s", dd.Status.Node)
continue
}
err := funcResumeCancellableDataRestore(r, ctx, dd, logger)
if err == nil {
logger.WithField("dd", dd.Name).WithField("current node", r.nodeName).Info("Completed to resume in progress DD")
continue
}
logger.WithField("datadownload", dd.GetName()).WithError(err).Warn("Failed to resume data path for dd, have to cancel it")
resumeErr := err
err = UpdateDataDownloadWithRetry(ctx, r.client, types.NamespacedName{Namespace: dd.Namespace, Name: dd.Name}, logger.WithField("datadownload", dd.Name),
func(dataDownload *velerov2alpha1api.DataDownload) bool {
if dataDownload.Spec.Cancel {
return false
}
dataDownload.Spec.Cancel = true
dataDownload.Status.Message = fmt.Sprintf("Resume InProgress datadownload failed with error %v, mark it as cancel", resumeErr)
return true
})
if err != nil {
logger.WithError(errors.WithStack(err)).WithError(errors.WithStack(err)).Error("Failed to trigger dataupload cancel")
}
} else if dd.Status.Phase == velerov2alpha1api.DataDownloadPhaseAccepted {
r.logger.WithField("datadownload", dd.GetName()).Warn("Cancel dd under Accepted phase")
err := UpdateDataDownloadWithRetry(ctx, r.client, types.NamespacedName{Namespace: dd.Namespace, Name: dd.Name},
r.logger.WithField("datadownload", dd.Name), func(dataDownload *velerov2alpha1api.DataDownload) bool {
if dataDownload.Spec.Cancel {
return false
}
dataDownload.Spec.Cancel = true
dataDownload.Status.Message = "Datadownload is in Accepted status during the node-agent starting, mark it as cancel"
return true
})
if err != nil {
r.logger.WithField("datadownload", dd.GetName()).WithError(err).Errorf("Failed to trigger dataupload cancel")
}
}
}
return nil
}
func (r *DataDownloadReconciler) resumeCancellableDataPath(ctx context.Context, dd *velerov2alpha1api.DataDownload, log logrus.FieldLogger) error {
log.Info("Resume cancelable dataDownload")
res, err := r.restoreExposer.GetExposed(ctx, getDataDownloadOwnerObject(dd), r.client, r.nodeName, dd.Spec.OperationTimeout.Duration)
if err != nil {
return errors.Wrapf(err, "error to get exposed volume for dd %s", dd.Name)
}
if res == nil {
return errors.Errorf("expose info missed for dd %s", dd.Name)
}
callbacks := datapath.Callbacks{
OnCompleted: r.OnDataDownloadCompleted,
OnFailed: r.OnDataDownloadFailed,
OnCancelled: r.OnDataDownloadCancelled,
OnProgress: r.OnDataDownloadProgress,
}
asyncBR, err := r.dataPathMgr.CreateMicroServiceBRWatcher(ctx, r.client, r.kubeClient, r.mgr, datapath.TaskTypeBackup, dd.Name, dd.Namespace, res.ByPod.HostingPod.Name, res.ByPod.HostingContainer, dd.Name, callbacks, true, log)
if err != nil {
return errors.Wrapf(err, "error to create asyncBR watcher for dd %s", dd.Name)
}
resumeComplete := false
defer func() {
if !resumeComplete {
r.closeDataPath(ctx, dd.Name)
}
}()
if err := asyncBR.Init(ctx, nil); err != nil {
return errors.Wrapf(err, "error to init asyncBR watcher for dd %s", dd.Name)
}
if err := asyncBR.StartRestore(dd.Spec.SnapshotID, datapath.AccessPoint{
ByPath: res.ByPod.VolumeName,
}, nil); err != nil {
return errors.Wrapf(err, "error to resume asyncBR watcher for dd %s", dd.Name)
}
resumeComplete = true
log.Infof("asyncBR is resumed for dd %s", dd.Name)
return nil
}