1Z0-997-25 Question 150
Single answerYou are tasked with migrating a large (5 TB) on-premises Oracle Database (version 19c) to an Oracle Autonomous Database on shared Exadata infrastructure in OCI. The initial bulk load using Oracle Data Pump completed successfully, but the cutover window is limited, and the business requires minimal downtime. Which approach will help you efficiently apply the remaining on-premises database changes to the target to meet the downtime requirement?
- A
Use Oracle Data Pump again in FULL mode and overwrite existing data in the target
- B
Use Oracle GoldenGate to replicate incremental changes after the initial load
- C
Perform full exports every night and import them into the target in parallel
- D
Suspend all on-premises transactions during the final Data Pump import
Show answer and explanation
Correct answer: B
Explanation
Oracle recommends using an online replication approach�such as Oracle GoldenGate�for minimal downtime during database migrations. After using Data Pump for the initial load, GoldenGate can continuously synchronize changes from the source to the target. This ensures that when you switch over, only a small delta remains, thereby reducing the cutover window significantly. For further details, refer to the Oracle Cloud Infrastructure documentation on data migration best practices and Oracle GoldenGate usage for real-time database replication.
- A. Incorrect.
Option 1: Use Oracle Data Pump again in FULL mode and overwrite existing data in the target. Explanation: Running another full export/import after the initial load will require a significant downtime window because it effectively replaces all data. This is not ideal for minimal downtime requirements.
- B. Correct.
Option 2: Use Oracle GoldenGate to replicate incremental changes after the initial load. Explanation: GoldenGate replicates ongoing transactions from the source to the target in near real time, reducing the cutover window to only the final switchover. This approach is well-suited to environments needing minimal downtime.
- C. Incorrect.
Option 3: Perform full exports every night and import them into the target in parallel. Explanation: Repeatedly performing full exports is inefficient for large databases. Even with parallel imports, this still involves a large downtime window and redundant data movement.
- D. Incorrect.
Option 4: Suspend all on-premises transactions during the final Data Pump import. Explanation: Freezing transactions can reduce data discrepancies, but it halts critical operations on the source. This is impractical when minimal downtime is required in a production environment.