Skip to content

Backup & Restore

The operator provides two custom resources for data protection: HyperbytedbBackup for creating backups and HyperbytedbRestore for restoring from them. Both target S3-compatible storage.

HyperbytedbBackup

One-Shot Backup

Create a single immediate backup:

apiVersion: hyperbytedb.hyperbyte.cloud/v1alpha1
kind: HyperbytedbBackup
metadata:
  name: hyperbytedb-manual-backup
  namespace: default
spec:
  clusterName: hyperbytedb-cluster
  destination:
    s3:
      bucket: hyperbytedb-backups
      prefix: "manual/"
      region: us-east-1
      credentialsSecretName: s3-backup-credentials
  retentionDays: 30
  backupType: full

Scheduled Backup

Use a cron expression to run backups on a schedule:

apiVersion: hyperbytedb.hyperbyte.cloud/v1alpha1
kind: HyperbytedbBackup
metadata:
  name: hyperbytedb-daily-backup
  namespace: default
spec:
  clusterName: hyperbytedb-cluster
  schedule: "0 2 * * *"
  destination:
    s3:
      bucket: hyperbytedb-backups
      prefix: "daily/"
      region: us-east-1
      credentialsSecretName: s3-backup-credentials
  retentionDays: 7
  backupType: full

This runs a full backup every day at 02:00 UTC and retains backups for 7 days.

Selective Database Backup

Back up specific databases instead of the entire cluster:

spec:
  clusterName: hyperbytedb-cluster
  databases:
    - production_metrics
    - application_logs
  destination:
    s3:
      bucket: hyperbytedb-backups
      prefix: "selective/"
      region: us-east-1
      credentialsSecretName: s3-backup-credentials

S3 Credentials Secret

The credentials Secret must contain access_key_id and secret_access_key keys:

apiVersion: v1
kind: Secret
metadata:
  name: s3-backup-credentials
  namespace: default
type: Opaque
stringData:
  access_key_id: "AKIAIOSFODNN7EXAMPLE"
  secret_access_key: "wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY"

Spec Reference

Field Type Default Description
clusterName string Name of the HyperbytedbCluster to back up (required, min length: 1)
schedule string Cron expression for scheduled backups. Omit for a one-shot backup
destination.s3.bucket string S3 bucket name (required)
destination.s3.prefix string Key prefix within the bucket
destination.s3.region string AWS region
destination.s3.endpoint string Custom S3-compatible endpoint (e.g., MinIO)
destination.s3.credentialsSecretName string Secret with S3 credentials
retentionDays int32 7 Days to retain backup artifacts before cleanup (min: 1)
databases list Restrict backup to specific databases. Empty means all databases
backupType string full Backup type: full or incremental

Status

Field Type Description
phase string Pending, Running, Completed, Failed
startTime Time When the backup started
completionTime Time When the backup finished
backupSize string Human-readable size (e.g., 1.2 GiB)
backupPath string S3 path where the backup was stored
lastCleanupTime Time Last time expired backups were cleaned up
conditions list Standard Kubernetes conditions

Monitor backup progress:

kubectl get hyperbytedbbackup hyperbytedb-daily-backup -o wide

NAME                    CLUSTER            SCHEDULE    PHASE       SIZE       AGE
hyperbytedb-daily-backup   hyperbytedb-cluster   0 2 * * *  Completed   1.2 GiB   1d

HyperbytedbRestore

Restore from a Backup

The simplest restore references a HyperbytedbBackup by name. The operator resolves the S3 location from the backup's status:

apiVersion: hyperbytedb.hyperbyte.cloud/v1alpha1
kind: HyperbytedbRestore
metadata:
  name: hyperbytedb-restore-latest
  namespace: default
spec:
  clusterName: hyperbytedb-cluster
  backupName: hyperbytedb-daily-backup

Restore from a Specific S3 Path

Override the source location to restore from a specific backup snapshot:

apiVersion: hyperbytedb.hyperbyte.cloud/v1alpha1
kind: HyperbytedbRestore
metadata:
  name: hyperbytedb-restore-specific
  namespace: default
spec:
  clusterName: hyperbytedb-cluster
  backupName: hyperbytedb-daily-backup
  source:
    s3:
      bucket: hyperbytedb-backups
      prefix: "daily/20260301-020000/"
      region: us-east-1
      credentialsSecretName: s3-backup-credentials

Restore Workflow

When a HyperbytedbRestore resource is created, the operator performs these steps:

  1. ScalingDown -- Scales the target cluster's StatefulSet to 0 replicas
  2. Restoring -- Downloads the backup from S3 and writes it to each PVC
  3. ScalingUp -- Scales the StatefulSet back to the original replica count
  4. Completed -- All replicas are healthy and serving traffic

Warning

A restore overwrites all data in the target cluster. The cluster is unavailable during the restore process. Plan accordingly for production restores.

Spec Reference

Field Type Default Description
clusterName string Name of the HyperbytedbCluster to restore into (required, min length: 1)
backupName string Name of the HyperbytedbBackup to restore from (required, min length: 1)
source object Override the backup source location (optional; resolved from the backup CR if omitted)
source.s3.bucket string S3 bucket name
source.s3.prefix string Key prefix for the specific backup snapshot
source.s3.region string AWS region
source.s3.endpoint string Custom S3-compatible endpoint
source.s3.credentialsSecretName string Secret with S3 credentials
restoreTimestamp Time RFC 3339 timestamp for point-in-time restore (reserved for future use)

Status

Field Type Description
phase string Pending, ScalingDown, Restoring, ScalingUp, Completed, Failed
startTime Time When the restore started
completionTime Time When the restore finished
restoredPVCs int32 Number of PVCs restored
conditions list Standard Kubernetes conditions

Monitor restore progress:

kubectl get hyperbytedbrestore hyperbytedb-restore-latest -o wide

NAME                      CLUSTER            BACKUP                  PHASE       AGE
hyperbytedb-restore-latest   hyperbytedb-cluster   hyperbytedb-daily-backup   Completed   5m

See Also