What's changed: Created Professional Cloud Database Engineer Chapter 3 (Domain 3 "Migrate": strategy and planning = zero/near-zero downtime (continuous replication)/extended outage/fallback/DDL-DML conversion/reverse replication; tools and replication = Database Migration Service/Datastream (CDC)/Storage Transfer Service/Transfer Appliance/validation).
3.1Migration strategy and planning
Understand migration strategies and planning (zero/near-zero downtime, extended outage, fallback), DDL/DML conversion, reverse replication from Google Cloud to the source, and choosing a migration approach by business requirements.
Many databases arrive on Google Cloud via migration, not greenfield. Plan how much downtime is acceptable and how to fall back on failure, from business requirements.
3.1.1Choosing a migration strategy
Choose by acceptable downtime: zero/near-zero downtime (catch up via continuous replication, then cut over) without stopping the business, or a planned extended outage. In all cases, plan a fallback to return to the source on failure. Map "migrate without stopping = zero/near-zero downtime (continuous replication)" and "return on failure = fallback."
3.1.2Conversion and reverse replication
Heterogeneous (cross-engine) migrations need DDL/DML conversion of schema and code (absorbing type/syntax/stored-procedure differences). As a post-migration safeguard, design reverse replication that streams changes from Google Cloud back to the source, enabling cutback on issues. Choose migration tools by requirements (homogeneous is easier for minimal downtime; heterogeneous needs conversion), and plan validation and production cutover. Map "heterogeneous migration = DDL/DML conversion" and "cutback insurance = reverse replication."
Common: requirement → strategy. E.g., "migrate without stopping the business" = zero/near-zero downtime (continuous replication → cutover); "return to source on failure" = fallback/reverse replication; "migrate to a different DB engine" = DDL/DML conversion; "bulk migrate during planned downtime" = extended-outage strategy.
Watch the mix-ups: (1) Choose zero/near-zero downtime (continuous replication) vs extended outage by requirements. (2) Heterogeneous migration needs DDL/DML conversion (more complex than homogeneous). (3) Plan validation and fallback, not just cutover.
3.1.3Section summary
- Migration strategy = zero/near-zero (continuous replication)/extended outage by acceptable downtime
- Heterogeneous = DDL/DML conversion; cutback via fallback/reverse replication
- Plan validation and production cutover
Sign in to track progress — Log in.
Quick check
(just a quick review)Q1. To migrate a database with minimal business disruption, which strategy best fits?
Q2. To be able to return to the source if migration fails, what should you design?
Q3. What is needed when migrating to a different database engine?
Q4. When a long planned outage is acceptable for a bulk migration, which strategy fits?
Q5. What should a migration plan include so it is not just "cut over and done"?
Q6. What is the central mechanism enabling zero/near-zero downtime migration?