If any compute resources fail to provision during startup, what does Snowflake do?

• A. It ignores the failure
• B. It attempts to repair the failed resources
• C. It restarts the entire warehouse
• D. It aborts startup

Answer: (B)

Snowflake is designed to be resilient when bringing compute resources online. When a virtual warehouse starts, it provisions the necessary compute clusters. If any of those provisioning steps fail, Snowflake automatically tries to fix the issue by repairing or re-provisioning the failed resources and then continues the startup process. This retrying behavior helps handle temporary hiccups in resource availability without requiring manual intervention. It’s not about ignoring the failure or aborting the startup immediately; the systemWorks to reallocate or re-create the needed resources so the warehouse can come online as soon as possible. In short, Snowflake will attempt to repair the failed resources and retry provisioning to complete startup.