OceanBase backups combine periodic data snapshots with continuous log archiving for point-in-time recovery. Validate restores in staging, confirm log archive continuity, and test PITR to a specific timestamp before relying on backups for compliance.

Production OceanBase operations require tested backups, not just scheduled jobs. PITR depends on both base backups and uninterrupted log archives.

Backup components

  • Data backup: Full or incremental snapshots of tenant data.
  • Log archive: Continuous archived redo for PITR windows.
  • Catalog metadata: Retain OCP or operator records of backup paths and retention.

Restore validation procedure

  1. Restore latest backup to an isolated staging tenant weekly.
  2. Replay archived logs to a target timestamp and run row-count checks on critical tables.
  3. Record restore duration and compare to RTO targets.

PITR workflow (outline)

  1. Identify target timestamp and tenant scope.
  2. Restore base backup to staging or recovery tenant.
  3. Apply archived logs until the target time.
  4. Validate application read queries before redirecting traffic.

Exact commands vary by OceanBase edition and OCP version; use the procedure documented for your deployment.

Common mistakes

  • Log archive gaps during maintenance windows.
  • Never testing restore size and duration at production data scale.
  • PITR rehearsal only at the cluster level when apps are tenant-scoped.

Sources and further reading