feat(backup): support recurring and ondemand full backup

ref: longhorn/longhorn 7070

Signed-off-by: Jack Lin <jack.lin@suse.com>

api/model.go

@@ -163,12 +163,15 @@ type Backup struct {
 	Created                string               `json:"created"`
 	Size                   string               `json:"size"`
 	Labels                 map[string]string    `json:"labels"`
+	BackupMode             longhorn.BackupMode  `json:"backupMode"`
 	Messages               map[string]string    `json:"messages"`
 	VolumeName             string               `json:"volumeName"`
 	VolumeSize             string               `json:"volumeSize"`
 	VolumeCreated          string               `json:"volumeCreated"`
 	VolumeBackingImageName string               `json:"volumeBackingImageName"`
 	CompressionMethod      string               `json:"compressionMethod"`
+	NewlyUploadedDataSize  string               `json:"newlyUploadDataSize"`
+	ReUploadedDataSize     string               `json:"reUploadedDataSize"`
 }
 
 type BackupBackingImage struct {
@@ -286,8 +289,9 @@ type DetachInput struct {
 }
 
 type SnapshotInput struct {
-	Name   string            `json:"name"`
-	Labels map[string]string `json:"labels"`
+	Name       string            `json:"name"`
+	Labels     map[string]string `json:"labels"`
+	BackupMode string            `json:"backupMode"`
 }
 
 type SnapshotCRInput struct {
@@ -548,6 +552,7 @@ type InstanceManager struct {
 type RecurringJob struct {
 	client.Resource
 	longhorn.RecurringJobSpec
+	longhorn.RecurringJobStatus
 }
 
 type Orphan struct {
@@ -828,6 +833,11 @@ func recurringJobSchema(job *client.Schema) {
 	labels.Type = "map[string]"
 	labels.Nullable = true
 	job.ResourceFields["labels"] = labels
+
+	parameters := job.ResourceFields["parameters"]
+	parameters.Type = "map[string]"
+	parameters.Nullable = true
+	job.ResourceFields["parameters"] = parameters
 }
 
 func kubernetesStatusSchema(status *client.Schema) {
@@ -1891,12 +1901,15 @@ func toBackupResource(b *longhorn.Backup) *Backup {
 		Created:                b.Status.BackupCreatedAt,
 		Size:                   b.Status.Size,
 		Labels:                 b.Status.Labels,
+		BackupMode:             b.Spec.BackupMode,
 		Messages:               b.Status.Messages,
 		VolumeName:             b.Status.VolumeName,
 		VolumeSize:             b.Status.VolumeSize,
 		VolumeCreated:          b.Status.VolumeCreated,
 		VolumeBackingImageName: b.Status.VolumeBackingImageName,
 		CompressionMethod:      string(b.Status.CompressionMethod),
+		NewlyUploadedDataSize:  b.Status.NewlyUploadedDataSize,
+		ReUploadedDataSize:     b.Status.ReUploadedDataSize,
 	}
 	// Set the volume name from backup CR's label if it's empty.
 	// This field is empty probably because the backup state is not Ready
@@ -2239,6 +2252,10 @@ func toRecurringJobResource(recurringJob *longhorn.RecurringJob, apiContext *api
 			Retain:      recurringJob.Spec.Retain,
 			Concurrency: recurringJob.Spec.Concurrency,
 			Labels:      recurringJob.Spec.Labels,
+			Parameters:  recurringJob.Spec.Parameters,
+		},
+		RecurringJobStatus: longhorn.RecurringJobStatus{
+			ExecutionCount: recurringJob.Status.ExecutionCount,
 		},
 	}
 }

api/recurringjob.go

@@ -63,6 +63,7 @@ func (s *Server) RecurringJobCreate(rw http.ResponseWriter, req *http.Request) e
 		Retain:      input.Retain,
 		Concurrency: input.Concurrency,
 		Labels:      input.Labels,
+		Parameters:  input.Parameters,
 	})
 	if err != nil {
 		return errors.Wrapf(err, "failed to create recurring job %v", input.Name)
@@ -90,6 +91,7 @@ func (s *Server) RecurringJobUpdate(rw http.ResponseWriter, req *http.Request) e
 			Retain:      input.Retain,
 			Concurrency: input.Concurrency,
 			Labels:      input.Labels,
+			Parameters:  input.Parameters,
 		})
 	})
 	if err != nil {

api/snapshot.go

@@ -186,6 +186,10 @@ func (s *Server) SnapshotBackup(w http.ResponseWriter, req *http.Request) (err e
 		return err
 	}
 
+	if err := util.ValidateBackupMode(input.BackupMode); err != nil {
+		return err
+	}
+
 	// Cannot directly compare the structs since KubernetesStatus contains a slice which cannot be compared.
 	if !reflect.DeepEqual(vol.Status.KubernetesStatus, longhorn.KubernetesStatus{}) {
 		kubeStatus, err := json.Marshal(vol.Status.KubernetesStatus)
@@ -195,7 +199,7 @@ func (s *Server) SnapshotBackup(w http.ResponseWriter, req *http.Request) (err e
 		labels[types.KubernetesStatusLabel] = string(kubeStatus)
 	}
 
-	if err := s.m.BackupSnapshot(bsutil.GenerateName("backup"), volName, input.Name, labels); err != nil {
+	if err := s.m.BackupSnapshot(bsutil.GenerateName("backup"), volName, input.Name, labels, input.BackupMode); err != nil {
 		return err
 	}
 

app/recurring_job.go

@@ -48,14 +48,16 @@ const (
 )
 
 type Job struct {
-	logger       logrus.FieldLogger
-	lhClient     lhclientset.Interface
-	namespace    string
-	volumeName   string
-	snapshotName string
-	retain       int
-	task         longhorn.RecurringJobType
-	labels       map[string]string
+	logger         logrus.FieldLogger
+	lhClient       lhclientset.Interface
+	namespace      string
+	volumeName     string
+	snapshotName   string
+	retain         int
+	task           longhorn.RecurringJobType
+	labels         map[string]string
+	parameters     map[string]string
+	executionCount int
 
 	eventRecorder record.EventRecorder
 
@@ -111,6 +113,13 @@ func recurringJob(c *cli.Context) (err error) {
 	var jobConcurrent int = recurringJob.Spec.Concurrency
 	jobTask := recurringJob.Spec.Task
 
+	recurringJob.Status.ExecutionCount += 1
+	if _, err = lhClient.LonghornV1beta2().RecurringJobs(namespace).UpdateStatus(context.TODO(), recurringJob, metav1.UpdateOptions{}); err != nil {
+		return errors.Wrap(err, "failed to update job execution count")
+	}
+
+	jobExecutionCount := recurringJob.Status.ExecutionCount
+
 	jobLabelMap := map[string]string{}
 	if recurringJob.Spec.Labels != nil {
 		jobLabelMap = recurringJob.Spec.Labels
@@ -121,6 +130,15 @@ func recurringJob(c *cli.Context) (err error) {
 		return errors.Wrap(err, "failed to get JSON encoding for labels")
 	}
 
+	jobParameterMap := map[string]string{}
+	if recurringJob.Spec.Parameters != nil {
+		jobParameterMap = recurringJob.Spec.Parameters
+	}
+	parameterJSON, err := json.Marshal(jobParameterMap)
+	if err != nil {
+		return errors.Wrap(err, "failed to get JSON encoding for parameters")
+	}
+
 	allowDetachedSetting := types.SettingNameAllowRecurringJobWhileVolumeDetached
 	allowDetached, err := getSettingAsBoolean(allowDetachedSetting, namespace, lhClient)
 	if err != nil {
@@ -153,7 +171,7 @@ func recurringJob(c *cli.Context) (err error) {
 	for _, volumeName := range filteredVolumes {
 		startJobVolumeName := volumeName
 		ewg.Go(func() error {
-			return startVolumeJob(startJobVolumeName, logger, concurrentLimiter, managerURL, jobName, jobTask, jobRetain, jobConcurrent, jobGroups, jobLabelMap, labelJSON)
+			return startVolumeJob(startJobVolumeName, logger, concurrentLimiter, managerURL, jobName, jobTask, jobRetain, jobConcurrent, jobGroups, jobLabelMap, labelJSON, jobParameterMap, parameterJSON, jobExecutionCount)
 		})
 	}
 
@@ -162,21 +180,23 @@ func recurringJob(c *cli.Context) (err error) {
 
 func startVolumeJob(
 	volumeName string, logger *logrus.Logger, concurrentLimiter chan struct{}, managerURL string,
-	jobName string, jobTask longhorn.RecurringJobType, jobRetain int, jobConcurrent int, jobGroups []string, jobLabelMap map[string]string, labelJSON []byte) error {
+	jobName string, jobTask longhorn.RecurringJobType, jobRetain int, jobConcurrent int, jobGroups []string, jobLabelMap map[string]string, labelJSON []byte, jobParameterMap map[string]string, parameterJSON []byte, jobExecutionCount int) error {
 
 	concurrentLimiter <- struct{}{}
 	defer func() {
 		<-concurrentLimiter
 	}()
 
 	log := logger.WithFields(logrus.Fields{
-		"job":        jobName,
-		"volume":     volumeName,
-		"task":       jobTask,
-		"retain":     jobRetain,
-		"concurrent": jobConcurrent,
-		"groups":     strings.Join(jobGroups, ","),
-		"labels":     string(labelJSON),
+		"job":            jobName,
+		"volume":         volumeName,
+		"task":           jobTask,
+		"retain":         jobRetain,
+		"concurrent":     jobConcurrent,
+		"groups":         strings.Join(jobGroups, ","),
+		"labels":         string(labelJSON),
+		"parameters":     string(parameterJSON),
+		"executionCount": jobExecutionCount,
 	})
 	log.Info("Creating job")
 
@@ -187,8 +207,10 @@ func startVolumeJob(
 		volumeName,
 		snapshotName,
 		jobLabelMap,
+		jobParameterMap,
 		jobRetain,
-		jobTask)
+		jobTask,
+		jobExecutionCount)
 	if err != nil {
 		log.WithError(err).Error("Failed to create new job for volume")
 		return err
@@ -213,7 +235,7 @@ func sliceStringSafely(s string, begin, end int) string {
 	return s[begin:end]
 }
 
-func newJob(logger logrus.FieldLogger, managerURL, volumeName, snapshotName string, labels map[string]string, retain int, task longhorn.RecurringJobType) (*Job, error) {
+func newJob(logger logrus.FieldLogger, managerURL, volumeName, snapshotName string, labels map[string]string, parameters map[string]string, retain int, task longhorn.RecurringJobType, executionCount int) (*Job, error) {
 	namespace := os.Getenv(types.EnvPodNamespace)
 	if namespace == "" {
 		return nil, fmt.Errorf("failed detect pod namespace, environment variable %v is missing", types.EnvPodNamespace)
@@ -238,12 +260,13 @@ func newJob(logger logrus.FieldLogger, managerURL, volumeName, snapshotName stri
 	}
 
 	logger = logger.WithFields(logrus.Fields{
-		"namespace":    namespace,
-		"volumeName":   volumeName,
-		"snapshotName": snapshotName,
-		"labels":       labels,
-		"retain":       retain,
-		"task":         task,
+		"namespace":      namespace,
+		"volumeName":     volumeName,
+		"snapshotName":   snapshotName,
+		"labels":         labels,
+		"retain":         retain,
+		"task":           task,
+		"executionCount": executionCount,
 	})
 
 	scheme := runtime.NewScheme()
@@ -257,15 +280,17 @@ func newJob(logger logrus.FieldLogger, managerURL, volumeName, snapshotName stri
 	}
 
 	return &Job{
-		logger:       logger,
-		lhClient:     lhClient,
-		namespace:    namespace,
-		volumeName:   volumeName,
-		snapshotName: snapshotName,
-		labels:       labels,
-		retain:       retain,
-		task:         task,
-		api:          apiClient,
+		logger:         logger,
+		lhClient:       lhClient,
+		namespace:      namespace,
+		volumeName:     volumeName,
+		snapshotName:   snapshotName,
+		labels:         labels,
+		parameters:     parameters,
+		executionCount: executionCount,
+		retain:         retain,
+		task:           task,
+		api:            apiClient,
 
 		eventRecorder: eventBroadcaster.NewRecorder(scheme, corev1.EventSource{Component: "longhorn-recurring-job"}),
 	}, nil
@@ -600,9 +625,24 @@ func (job *Job) doRecurringBackup() (err error) {
 		return err
 	}
 
+	backupMode := longhorn.BackupModeIncremental
+	if intervalStr, exists := job.parameters[types.RecurringJobBackupParameterFullBackupInterval]; exists {
+		interval, err := strconv.Atoi(intervalStr)
+		if err != nil {
+			return errors.Wrapf(err, "interval %v is not number", intervalStr)
+		}
+
+		if interval != 0 {
+			if job.executionCount%interval == 0 {
+				backupMode = longhorn.BackupModeFull
+			}
+		}
+	}
+
 	if _, err := job.api.Volume.ActionSnapshotBackup(volume, &longhornclient.SnapshotInput{
-		Labels: job.labels,
-		Name:   job.snapshotName,
+		Labels:     job.labels,
+		Name:       job.snapshotName,
+		BackupMode: string(backupMode),
 	}); err != nil {
 		return err
 	}

client/generated_recurring_job.go

@@ -11,6 +11,8 @@ type RecurringJob struct {
 
 	Cron string `json:"cron,omitempty" yaml:"cron,omitempty"`
 
+	ExecutionCount int64 `json:"executionCount,omitempty" yaml:"execution_count,omitempty"`
+
 	Groups []string `json:"groups,omitempty" yaml:"groups,omitempty"`
 
 	Labels map[string]string `json:"labels,omitempty" yaml:"labels,omitempty"`

client/generated_snapshot_input.go

@@ -10,6 +10,8 @@ type SnapshotInput struct {
 	Labels map[string]string `json:"labels,omitempty" yaml:"labels,omitempty"`
 
 	Name string `json:"name,omitempty" yaml:"name,omitempty"`
+
+	BackupMode string `json:"backupMode,omitempty" yaml:"backupMode,omitempty"`
 }
 
 type SnapshotInputCollection struct {

controller/backup_controller.go

@@ -441,6 +441,8 @@ func (bc *BackupController) reconcile(backupName string) (err error) {
 	backup.Status.VolumeBackingImageName = backupInfo.VolumeBackingImageName
 	backup.Status.CompressionMethod = longhorn.BackupCompressionMethod(backupInfo.CompressionMethod)
 	backup.Status.LastSyncedAt = syncTime
+	backup.Status.NewlyUploadedDataSize = backupInfo.NewlyUploadedDataSize
+	backup.Status.ReUploadedDataSize = backupInfo.ReUploadedDataSize
 	return nil
 }
 

datastore/longhorn.go

@@ -4193,6 +4193,41 @@ func ValidateRecurringJob(job longhorn.RecurringJobSpec) error {
 			return err
 		}
 	}
+	if job.Parameters != nil {
+		if err := ValidateRecurringJobParameters(job.Task, job.Labels); err != nil {
+			return err
+		}
+	}
+	return nil
+}
+
+func ValidateRecurringJobParameters(task longhorn.RecurringJobType, parameters map[string]string) (err error) {
+	switch task {
+	case longhorn.RecurringJobTypeBackup, longhorn.RecurringJobTypeBackupForceCreate:
+		for key, value := range parameters {
+			if err := validateRecurringJobBackupParameter(key, value); err != nil {
+				return errors.Wrapf(err, "failed to validate recurring job backup task parameters")
+			}
+		}
+	// we don't support any parameters for other tasks currently
+	default:
+		return nil
+	}
+
+	return nil
+}
+
+func validateRecurringJobBackupParameter(key, value string) error {
+	switch key {
+	case types.RecurringJobBackupParameterFullBackupInterval:
+		_, err := strconv.Atoi(value)
+		if err != nil {
+			return errors.Wrapf(err, "%v:%v is not number", key, value)
+		}
+	default:
+		return fmt.Errorf("%v:%v is not a valid parameter", key, value)
+	}
+
 	return nil
 }