Two options
| Cross-region Autonomous Data Guard | Backup-based DR | |
|---|---|---|
| Standby | Running standby database in the remote region | No running database; backups replicated to the remote region |
| RPO | Seconds to around a minute (async redo) | Typically up to ~1 minute behind for redo-based backups (check current SLO) |
| RTO | Minutes | Longer (hours for large databases, depends on size) |
| Cost | Standby billed (ECPU + storage) | Mostly storage for backups; much cheaper |
| Use | Tier 0/1 | Tier 2, or dev/test |
You can combine: a local Autonomous Data Guard standby (HA) plus a cross-region standby or backup-based peer (DR).
1. Set up
Console → Autonomous Database → Disaster recovery → Add peer database → peer region (for example Hyderabad), type Autonomous Data Guard or Backup-based.
For cross-region standbys with private endpoints, choose the remote VCN subnet and NSGs.
2. Check status
oci db autonomous-database get --autonomous-database-id <primary-ocid> \
--query "data.{role:role, peers:\"peer-db-ids\", dr:\"disaster-recovery-region-type\"}"
3. Switchover (planned)
From the standby in the remote region: Disaster recovery → Switchover. No data loss; roles swap.
4. Failover (disaster)
Cross-region failover is manual (operator decision): from the remote standby, choose Failover. With backup-based DR, failover creates a new database from the replicated backups in the remote region.
5. Connection strings
- The remote database has its own connection strings / wallet (different region hostnames).
- Keep the remote wallet or TLS connection strings ready in the DR app configuration (in Vault).
- With apps in both regions, each app tier points to its local database; failover then pairs with app failover (DNS / Traffic Management).
6. Snapshot standby / reporting
Cross-region standbys can be converted to a snapshot standby for testing (read-write, changes discarded when converting back) or used read-only for reporting in some configurations. Use it to run DR tests without touching production.
7. After failover
- New primary in the DR region; when the original region returns, OCI can reinstate a standby there.
- Update backups, alarms, Database Management and application configs for the new region.
Tip
Backup-based DR is a cheap, real improvement over "we have backups somewhere". Start there for less critical databases, and use Autonomous Data Guard only where the RTO really needs it.