Recovery
HANA recovery restores the database from backups after a failure, corruption or for a system copy, using data and log backups to reach a consistent, chosen point in time. A rehearsed recovery procedure is what makes backups worthwhile.
Recovery must be practised.
- To recover, you restore a data backup and then replay log backups up to the desired point (the latest state, or a specific point in…
- Beyond backup-based recovery, HANA System Replication continuously replicates to a secondary system for high availability and…
- Full recovery after data loss or corruption.
- Watch out: Never rehearsing recovery.
The recovery process
To recover, you restore a data backup and then replay log backups up to the desired point (the latest state, or a specific point in time before an incident). HANA reconstructs the in-memory database from the persisted data and logs and returns to a consistent state.
Recovery scenarios
- Full recovery after data loss or corruption.
- Point-in-time recovery to just before a bad change.
- System copy/refresh, recovering PRD data into a non-production system.
System replication for HA/DR
Beyond backup-based recovery, HANA System Replication continuously replicates to a secondary system for high availability and disaster recovery, enabling fast failover with minimal data loss, complementing backups (which protect against corruption and provide point-in-time recovery).
Rehearse
Recovery must be practised. Regular restore/recovery drills validate the backups, confirm the RPO/RTO are achievable, and ensure the team can execute under pressure, so the first real recovery is not the first attempt.
Common pitfalls
- Never rehearsing recovery.
- Missing log backups for point-in-time recovery.
- Confusing replication (HA/DR) with backup, you need both.