COF-C03 Question 315
Single answerCloningA data engineering team needs a fast way to create a test environment from the PROD_DB database before a major release. They want the test copy to reflect the exact current state of production at the time the copy is created, while minimizing storage costs and avoiding disruption to production workloads. A developer suggests using zero-copy cloning. Which approach best meets these requirements?
- A
Create a clone of PROD_DB. The clone is created quickly using existing micro-partitions, and additional storage is only consumed as data changes in either the source or the clone.
- B
Export PROD_DB to external stage files and reload the data into a new database. This ensures the copy is metadata-linked to production and minimizes storage costs.
- C
Create a database from a database share based on PROD_DB. This creates an independent writable copy for testing with no additional steps required.
- D
Use Time Travel to query PROD_DB directly for testing. This creates an isolated writable test environment without copying any objects.
Show answer and explanation
Correct answer: A
Explanation
The best answer is to create a zero-copy clone of PROD_DB. In Snowflake, cloning is designed for exactly this type of scenario: rapidly provisioning development, testing, and QA environments from production with minimal delay and low incremental storage cost. A clone is a point-in-time copy of the source object and initially shares the same underlying micro-partitions. As changes occur to the source or clone, Snowflake maintains independence by writing new micro-partitions as needed. This makes cloning significantly more efficient than exporting and reloading data. It is also more appropriate than secure data sharing, which produces read-only shared data, and more suitable than Time Travel alone, which supports historical access but not isolated writable environments. This aligns with Snowflake documentation and common best practices for non-production environment provisioning.
- A. Correct.
Correct. Zero-copy cloning creates a point-in-time copy of a database, schema, or table almost instantly by reusing the existing micro-partitions rather than physically copying all data. The clone reflects the state of the source at clone time and is independent for subsequent changes. Storage usage remains low initially, with additional storage consumed only for changed or newly added micro-partitions after the clone is created. This is the standard Snowflake approach for quickly creating dev/test environments with minimal impact on production.
- B. Incorrect.
Incorrect. Unloading and reloading data is much slower, operationally heavier, and usually consumes more storage and compute than cloning. It also does not create a metadata-linked copy of the production objects. This approach may be used for data movement, but it does not provide the fast, low-cost, point-in-time environment that zero-copy cloning provides.
- C. Incorrect.
Incorrect. A database created from a share is not the same as a clone. Shared databases are read-only for consumers and are intended for secure data sharing, not for creating a writable test environment. To make changes, the team would need additional steps such as copying objects into a local database, so this does not best satisfy the requirement.
- D. Incorrect.
Incorrect. Time Travel allows querying historical data versions from a previous point in time, but it does not by itself create an isolated writable environment. The team specifically needs a separate test environment that developers can modify safely without affecting production. Querying production data with Time Travel does not meet that need.