Databricks Machine Learning Professional Question 103
Select 3You are tasked with running a one-time data preprocessing job that involves training a machine learning model in Databricks. The job will run for a few hours and does not require real-time user interaction. Why would you choose a job cluster over an all-purpose cluster for this task?
- A
Job clusters are automatically terminated after the job completes, reducing unnecessary cloud costs.
- B
Job clusters support higher compute performance compared to all-purpose clusters.
- C
Job clusters are specifically configured and optimized for the workload of a single job.
- D
Job clusters allow multiple users to collaborate in real time while the job is running.
- E
Job clusters are created on demand, ensuring a clean environment free of prior configurations.
Show answer and explanation
Correct answers: A, C, E
Explanation
Job clusters are ideal for one-time or scheduled tasks, such as data preprocessing or training machine learning models, because they are created on demand for a specific job and automatically terminated upon job completion. This reduces costs, ensures an isolated environment, and optimizes the cluster for the specific workload. Unlike all-purpose clusters, job clusters are not designed for real-time collaboration or persistent use.
- A. Correct.
Correct: Job clusters are terminated automatically once the job completes, which helps optimize cloud resource usage and reduce costs.
- B. Incorrect.
Incorrect: Job clusters do not inherently provide higher compute performance compared to all-purpose clusters; performance is determined by the cluster configuration.
- C. Correct.
Correct: Job clusters are designed to handle the workload of a single job and are optimized for that specific job's requirements.
- D. Incorrect.
Incorrect: Job clusters do not support real-time collaboration as they are intended for automated, single-job workloads. All-purpose clusters are better suited for this kind of interactive use case.
- E. Correct.
Correct: Job clusters are created on demand for each job, ensuring a clean and isolated environment free of prior configurations or interference from other jobs.