A company runs a critical Amazon EMR data job for 6 hours each day. The job cannot lose any data while it runs. Which EMR cluster configuration will meet these requirements MOST cost-effectively?

The most cost-effective solution for a daily, 6-hour critical EMR job that cannot lose data is a transient cluster with primary and core nodes on On-Demand Instances, and task nodes on Spot Instances. A transient cluster is launched only when needed and terminated upon completion, saving costs compared to a long-running cluster. On-Demand Instances for primary and core nodes ensure job stability and data integrity, as these nodes are critical and cannot be interrupted. Task nodes, which process data and can be gracefully replaced without data loss, are ideal for cost savings using Spot Instances. Incorrect options: Long-running clusters are more expensive as they run continuously even when not processing data. Using Spot Instances for core nodes is risky as their interruption can lead to data loss or job failure, violating the requirement.