ADA-C01 Question 333
Single answerExplain the impact on the local disc cache when a warehouse is suspended or resumedA data engineering team runs a reporting workload every morning on a dedicated Snowflake virtual warehouse. To reduce cost, the warehouse is configured to AUTO_SUSPEND after 5 minutes of inactivity and AUTO_RESUME when queries arrive. The team notices that the first dashboard refresh after the warehouse resumes is slower than subsequent refreshes, even though the underlying tables have not changed. Which statement best explains this behavior?
- A
When the warehouse suspends, its local disk cache is cleared, so after resume the warehouse must rebuild cache from remote storage.
- B
When the warehouse resumes, Snowflake automatically restores the previous local disk cache from cloud services metadata.
- C
The result cache is deleted when the warehouse suspends, so the first query after resume must recompute all results even if the SQL text is identical.
- D
Suspending and resuming a warehouse preserves the local disk cache, but invalidates the metadata cache for micro-partitions.
Show answer and explanation
Correct answer: A
Explanation
This scenario tests understanding of Snowflake caching layers in an operational context. A virtual warehouse maintains a local disk cache while it is running, which can improve performance for repeated reads of the same data. However, that cache exists only on the active compute resources. When a warehouse is suspended, those resources are released and the local disk cache is lost. When the warehouse resumes, it starts with an empty local cache, so initial queries may run more slowly until data is read again and the cache is repopulated. This is distinct from the persisted query result cache, which is managed separately from warehouse lifecycle and can still be reused if normal result cache conditions are met. Snowflake documentation and performance best practices distinguish between warehouse local caching and persisted result caching, and administrators should account for this when balancing cost savings from auto-suspend against warm-cache performance for repetitive workloads.
- A. Correct.
Correct. Snowflake's local disk cache is tied to the compute resources backing the virtual warehouse. When the warehouse is suspended, those compute resources are shut down and the local cache is lost. After resume, queries may need to fetch data again from remote storage, so the first execution can be slower until the cache is warmed.
- B. Incorrect.
Incorrect. Snowflake does not persist and reload the warehouse's local disk cache across suspend/resume events. This option reflects a common misconception that cache state is checkpointed and restored, but local disk cache is ephemeral and associated with the running warehouse nodes.
- C. Incorrect.
Incorrect. This confuses the local disk cache with the persisted query result cache. The result cache is not tied to a warehouse's running state, and suspend/resume does not by itself delete it. If the query is eligible for result cache reuse, it can still be used after resume.
- D. Incorrect.
Incorrect. The opposite is true for the local disk cache: it is not preserved across suspend/resume. Also, this option introduces an inaccurate claim about metadata cache invalidation as the primary effect. The key behavior relevant to warehouse suspend/resume is loss of the local disk cache on the compute layer.