Memory hook: Copy a baseline, capture change, reconcile, then cut over.
Must remember
Start with data volume, transfer window, source load, supported engines and acceptable downtime. Storage Transfer Service moves supported object/file data; Transfer Appliance addresses appropriate offline bulk transfer constraints. Database Migration Service supports specific database migrations, while Datastream captures supported database changes for downstream pipelines. Product support must match the exact source and target versions.
A common migration sequence is full load, continuous change capture, reconciliation, controlled write cutover and observation. Check row counts, checksums or business totals, schema conversion, character encoding, time zones and sequence behavior. CDC transports changes; it does not automatically make a non-compatible application/schema portable.
Define RPO as tolerable data loss and RTO as restoration time. Replication supports availability but may reproduce bad writes or deletion. Backups and point-in-time recovery provide historical recovery subject to retention and service limits. Test restoration into an isolated environment with realistic permissions, keys and dependencies.
ACID addresses transactional properties; eventual consistency describes convergence behavior. Select consistency and transaction boundaries for the actual business invariant. A distributed pipeline may need deduplication and compensating actions even when each individual database operation is transactional.
Plan regional outages and missing/corrupt data separately. Managed database failover may change endpoints or connections; applications need reconnection and retry logic. Redis/cache recovery should not become the sole recovery strategy for authoritative data. Preserve replayable source records where lawful and economical, and document who can approve cutover or rollback.
Choose under exam pressure
| Requirement | Choice and reason |
|---|---|
| Low-downtime supported relational migration | Baseline plus CDC, validation and controlled cutover. |
| Accidental corruption replicated everywhere | Recover from verified history rather than another damaged replica. |
| Recover stream outputs | Replay retained input with deterministic/idempotent processing. |
Traps
- A replica is not a complete backup strategy.
- A green transfer job does not prove business reconciliation passed.
Active recall
1. RPO versus RTO?
Allowed data loss versus recovery duration.
2. What does CDC capture?
Changes from a supported source after/beside an initial baseline.
3. Why test keys during restore?
Encrypted backup data is unusable without authorized access to the required keys.
4. Why keep a rollback window?
To recover if application or data validation fails after cutover.
5. Why might retries be unsafe?
An earlier attempt may have committed before the response was lost.