Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

K8SPXC-1366: Add suspendedDeadlineSeconds #1970

Merged
merged 3 commits into from
Feb 11, 2025
Merged

K8SPXC-1366: Add suspendedDeadlineSeconds #1970

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
025ae1a
K8SPXC-1366: Add suspendedDeadlineSeconds
egegunes Feb 7, 2025
8202127
K8SPXC-1366: Add tests for deadlines
egegunes Feb 10, 2025
3b6bb73
Merge branch 'main' into K8SPXC-1366
hors Feb 11, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions config/crd/bases/pxc.percona.com_perconaxtradbclusterbackups.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -151,6 +151,9 @@ spec:
type: integer
storageName:
type: string
suspendedDeadlineSeconds:
format: int64
type: integer
type: object
status:
properties:
Expand Down
3 changes: 3 additions & 0 deletions config/crd/bases/pxc.percona.com_perconaxtradbclusters.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1083,6 +1083,9 @@ spec:
type: object
type: object
type: object
suspendedDeadlineSeconds:
format: int64
type: integer
type: object
crVersion:
type: string
Expand Down
1 change: 1 addition & 0 deletions deploy/backup/backup.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,6 +9,7 @@ spec:
storageName: fs-pvc
# activeDeadlineSeconds: 3600
# startingDeadlineSeconds: 300
# suspendedDeadlineSeconds: 1200
# containerOptions:
# env:
# - name: VERIFY_TLS
Expand Down
6 changes: 6 additions & 0 deletions deploy/bundle.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -150,6 +150,9 @@ spec:
type: integer
storageName:
type: string
suspendedDeadlineSeconds:
format: int64
type: integer
type: object
status:
properties:
Expand Down Expand Up @@ -1997,6 +2000,9 @@ spec:
type: object
type: object
type: object
suspendedDeadlineSeconds:
format: int64
type: integer
type: object
crVersion:
type: string
Expand Down
1 change: 1 addition & 0 deletions deploy/cr.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -606,6 +606,7 @@ spec:
# backoffLimit: 6
# activeDeadlineSeconds: 3600
# startingDeadlineSeconds: 300
# suspendedDeadlineSeconds: 1200
# serviceAccountName: percona-xtradb-cluster-operator
# imagePullSecrets:
# - name: private-registry-credentials
Expand Down
6 changes: 6 additions & 0 deletions deploy/crd.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -150,6 +150,9 @@ spec:
type: integer
storageName:
type: string
suspendedDeadlineSeconds:
format: int64
type: integer
type: object
status:
properties:
Expand Down Expand Up @@ -1997,6 +2000,9 @@ spec:
type: object
type: object
type: object
suspendedDeadlineSeconds:
format: int64
type: integer
type: object
crVersion:
type: string
Expand Down
6 changes: 6 additions & 0 deletions deploy/cw-bundle.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -150,6 +150,9 @@ spec:
type: integer
storageName:
type: string
suspendedDeadlineSeconds:
format: int64
type: integer
type: object
status:
properties:
Expand Down Expand Up @@ -1997,6 +2000,9 @@ spec:
type: object
type: object
type: object
suspendedDeadlineSeconds:
format: int64
type: integer
type: object
crVersion:
type: string
Expand Down
11 changes: 6 additions & 5 deletions pkg/apis/pxc/v1/pxc_backup_types.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -47,11 +47,12 @@ type PerconaXtraDBClusterBackup struct {
}

type PXCBackupSpec struct {
PXCCluster string `json:"pxcCluster"`
StorageName string `json:"storageName,omitempty"`
ContainerOptions *BackupContainerOptions `json:"containerOptions,omitempty"`
StartingDeadlineSeconds *int64 `json:"startingDeadlineSeconds,omitempty"`
ActiveDeadlineSeconds *int64 `json:"activeDeadlineSeconds,omitempty"`
PXCCluster string `json:"pxcCluster"`
StorageName string `json:"storageName,omitempty"`
ContainerOptions *BackupContainerOptions `json:"containerOptions,omitempty"`
ActiveDeadlineSeconds *int64 `json:"activeDeadlineSeconds,omitempty"`
StartingDeadlineSeconds *int64 `json:"startingDeadlineSeconds,omitempty"`
SuspendedDeadlineSeconds *int64 `json:"suspendedDeadlineSeconds,omitempty"`
}

type PXCBackupStatus struct {
Expand Down
25 changes: 13 additions & 12 deletions pkg/apis/pxc/v1/pxc_types.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -160,18 +160,19 @@ const (
)

type PXCScheduledBackup struct {
AllowParallel *bool `json:"allowParallel,omitempty"`
Image string `json:"image,omitempty"`
ImagePullSecrets []corev1.LocalObjectReference `json:"imagePullSecrets,omitempty"`
ImagePullPolicy corev1.PullPolicy `json:"imagePullPolicy,omitempty"`
Schedule []PXCScheduledBackupSchedule `json:"schedule,omitempty"`
Storages map[string]*BackupStorageSpec `json:"storages,omitempty"`
ServiceAccountName string `json:"serviceAccountName,omitempty"`
Annotations map[string]string `json:"annotations,omitempty"`
PITR PITRSpec `json:"pitr,omitempty"`
BackoffLimit *int32 `json:"backoffLimit,omitempty"`
ActiveDeadlineSeconds *int64 `json:"activeDeadlineSeconds,omitempty"`
StartingDeadlineSeconds *int64 `json:"startingDeadlineSeconds,omitempty"`
AllowParallel *bool `json:"allowParallel,omitempty"`
Image string `json:"image,omitempty"`
ImagePullSecrets []corev1.LocalObjectReference `json:"imagePullSecrets,omitempty"`
ImagePullPolicy corev1.PullPolicy `json:"imagePullPolicy,omitempty"`
Schedule []PXCScheduledBackupSchedule `json:"schedule,omitempty"`
Storages map[string]*BackupStorageSpec `json:"storages,omitempty"`
ServiceAccountName string `json:"serviceAccountName,omitempty"`
Annotations map[string]string `json:"annotations,omitempty"`
PITR PITRSpec `json:"pitr,omitempty"`
BackoffLimit *int32 `json:"backoffLimit,omitempty"`
ActiveDeadlineSeconds *int64 `json:"activeDeadlineSeconds,omitempty"`
StartingDeadlineSeconds *int64 `json:"startingDeadlineSeconds,omitempty"`
SuspendedDeadlineSeconds *int64 `json:"suspendedDeadlineSeconds,omitempty"`
}

func (b *PXCScheduledBackup) GetAllowParallel() bool {
Expand Down
14 changes: 12 additions & 2 deletions pkg/apis/pxc/v1/zz_generated.deepcopy.go
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

99 changes: 54 additions & 45 deletions pkg/controller/pxcbackup/controller.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -151,14 +151,6 @@ func (r *ReconcilePerconaXtraDBClusterBackup) Reconcile(ctx context.Context, req
return rr, nil
}

if err := r.checkStartingDeadline(ctx, cr); err != nil {
if err := r.setFailedStatus(ctx, cr, err); err != nil {
return rr, errors.Wrap(err, "update status")
}

return reconcile.Result{}, nil
}

cluster, err := r.getCluster(ctx, cr)
if err != nil {
return reconcile.Result{}, errors.Wrap(err, "get cluster")
Expand Down Expand Up @@ -187,6 +179,21 @@ func (r *ReconcilePerconaXtraDBClusterBackup) Reconcile(ctx context.Context, req
return reconcile.Result{}, err
}

if err := r.checkDeadlines(ctx, cluster, cr); err != nil {
if err := r.setFailedStatus(ctx, cr, err); err != nil {
return rr, errors.Wrap(err, "update status")
}

if errors.Is(err, errSuspendedDeadlineExceeded) {
log.Info("cleaning up suspended backup job")
if err := r.cleanUpSuspendedJob(ctx, cluster, cr); err != nil {
return reconcile.Result{}, errors.Wrap(err, "clean up suspended job")
}
}

return reconcile.Result{}, nil
}

if err := r.reconcileBackupJob(ctx, cr, cluster); err != nil {
return rr, errors.Wrap(err, "reconcile backup job")
}
Expand Down Expand Up @@ -661,29 +668,6 @@ func (r *ReconcilePerconaXtraDBClusterBackup) updateJobStatus(
return nil
}

func (r *ReconcilePerconaXtraDBClusterBackup) checkStartingDeadline(ctx context.Context, cr *api.PerconaXtraDBClusterBackup) error {
log := logf.FromContext(ctx)

since := time.Since(cr.CreationTimestamp.Time).Seconds()

if cr.Spec.StartingDeadlineSeconds == nil {
return nil
}

if since < float64(*cr.Spec.StartingDeadlineSeconds) {
return nil
}

if cr.Status.State == api.BackupNew {
log.Info("Backup didn't start in startingDeadlineSeconds, failing the backup",
"startingDeadlineSeconds", *cr.Spec.StartingDeadlineSeconds,
"passedSeconds", since)
return errors.New("starting deadline seconds exceeded")
}

return nil
}

func (r *ReconcilePerconaXtraDBClusterBackup) updateStatus(ctx context.Context, cr *api.PerconaXtraDBClusterBackup) error {
return retry.RetryOnConflict(retry.DefaultRetry, func() error {
localCr := new(api.PerconaXtraDBClusterBackup)
Expand Down Expand Up @@ -726,13 +710,8 @@ func (r *ReconcilePerconaXtraDBClusterBackup) suspendJobIfNeeded(

log := logf.FromContext(ctx)

labelKeyBackupType := naming.GetLabelBackupType(cluster)
jobName := naming.BackupJobName(cr.Name, cr.Labels[labelKeyBackupType] == "cron")

err := retry.RetryOnConflict(retry.DefaultRetry, func() error {
job := new(batchv1.Job)

err := r.client.Get(ctx, types.NamespacedName{Namespace: cr.Namespace, Name: jobName}, job)
job, err := r.getBackupJob(ctx, cluster, cr)
if err != nil {
if k8sErrors.IsNotFound(err) {
return nil
Expand All @@ -752,7 +731,7 @@ func (r *ReconcilePerconaXtraDBClusterBackup) suspendJobIfNeeded(
}

log.Info("Suspending backup job",
"job", jobName,
"job", job.Name,
"clusterStatus", cluster.Status.Status,
"readyPXC", cluster.Status.PXC.Ready)

Expand Down Expand Up @@ -785,13 +764,8 @@ func (r *ReconcilePerconaXtraDBClusterBackup) resumeJobIfNeeded(

log := logf.FromContext(ctx)

labelKeyBackupType := naming.GetLabelBackupType(cluster)
jobName := naming.BackupJobName(cr.Name, cr.Labels[labelKeyBackupType] == "cron")

err := retry.RetryOnConflict(retry.DefaultRetry, func() error {
job := new(batchv1.Job)

err := r.client.Get(ctx, types.NamespacedName{Namespace: cr.Namespace, Name: jobName}, job)
job, err := r.getBackupJob(ctx, cluster, cr)
if err != nil {
if k8sErrors.IsNotFound(err) {
return nil
Expand All @@ -811,7 +785,7 @@ func (r *ReconcilePerconaXtraDBClusterBackup) resumeJobIfNeeded(
}

log.Info("Resuming backup job",
"job", jobName,
"job", job.Name,
"clusterStatus", cluster.Status.Status,
"readyPXC", cluster.Status.PXC.Ready)

Expand All @@ -838,3 +812,38 @@ func (r *ReconcilePerconaXtraDBClusterBackup) reconcileBackupJob(

return nil
}

func (r *ReconcilePerconaXtraDBClusterBackup) getBackupJob(
ctx context.Context,
cluster *api.PerconaXtraDBCluster,
cr *api.PerconaXtraDBClusterBackup,
) (*batchv1.Job, error) {
labelKeyBackupType := naming.GetLabelBackupType(cluster)
jobName := naming.BackupJobName(cr.Name, cr.Labels[labelKeyBackupType] == "cron")

job := new(batchv1.Job)

err := r.client.Get(ctx, types.NamespacedName{Namespace: cr.Namespace, Name: jobName}, job)
if err != nil {
return nil, err
}

return job, nil
}

func (r *ReconcilePerconaXtraDBClusterBackup) cleanUpSuspendedJob(
ctx context.Context,
cluster *api.PerconaXtraDBCluster,
cr *api.PerconaXtraDBClusterBackup,
) error {
job, err := r.getBackupJob(ctx, cluster, cr)
if err != nil {
return errors.Wrap(err, "get job")
}

if err := r.client.Delete(ctx, job); err != nil {
return errors.Wrap(err, "delete job")
}

return nil
}
Loading
Loading