Your team uses Cloud Dataproc for Spark ML training jobs. You want to reduce costs for non-critical, fault-tolerant training jobs. Which Dataproc feature should you use for worker nodes?
Trap 1: Use custom machine types with more memory.
Custom machine types may cost more; they don't inherently reduce costs.
Trap 2: Use SSDs instead of HDDs for persistent disks.
SSDs are faster but more expensive, not a cost-reduction measure.
Trap 3: Use committed use discounts for 1-year or 3-year terms.
Committed use discounts reduce cost but require upfront commitment; preemptible instances are cheaper for fault-tolerant jobs without commitment.
- A
Use preemptible instances for worker nodes.
Preemptible instances cost ~60-80% less and are suitable for fault-tolerant batch jobs.
- B
Use custom machine types with more memory.
Why wrong: Custom machine types may cost more; they don't inherently reduce costs.
- C
Use SSDs instead of HDDs for persistent disks.
Why wrong: SSDs are faster but more expensive, not a cost-reduction measure.
- D
Use committed use discounts for 1-year or 3-year terms.
Why wrong: Committed use discounts reduce cost but require upfront commitment; preemptible instances are cheaper for fault-tolerant jobs without commitment.