Administer Slurm clusters: Practice Questions — Administration (NVIDIA-Certified Professional: AI Operations)

Practice Questions: Administering Slurm Clusters for NVIDIA-Certified Professional: AI Operations

These multiple-choice questions are designed to help you prepare for the Administration section of the NVIDIA-Certified Professional: AI Operations exam, focusing specifically on administering Slurm clusters.

1. Which Slurm command is used to view the current status of nodes in a cluster?

   • A. scontrol show nodes
   • B. sinfo
   • C. squeue
   • D. sbatch

   Correct Answer: B. sinfo

   Explanation: The sinfo command provides a summary of the cluster's nodes and partitions, showing their current state. While scontrol show nodes gives detailed node information, sinfo is the standard command for a quick overview.

2. What is the primary purpose of the Slurm sbatch command?

   • A. Submit a batch job to the scheduler
   • B. Cancel a running job
   • C. Display job queue status
   • D. Modify node configuration

   Correct Answer: A. Submit a batch job to the scheduler

   Explanation: The sbatch command submits a batch script to the Slurm scheduler for execution. It does not cancel jobs, display queues, or modify nodes.

3. In Slurm, which configuration file typically defines the cluster nodes and partitions?

   • A. slurm.conf
   • B. slurmdbd.conf
   • C. munge.key
   • D. cgroup.conf

   Correct Answer: A. slurm.conf

   Explanation: The slurm.conf file contains the main configuration for the Slurm cluster, including node definitions, partitions, scheduling policies, and other settings.

4. Which Slurm command would you use to cancel a job with job ID 12345?

   • A. squeue -j 12345
   • B. scancel 12345
   • C. scontrol cancel 12345
   • D. sbatch --cancel 12345

   Correct Answer: B. scancel 12345

   Explanation: The scancel command is used to terminate jobs by their job ID. The other options are either invalid or used for different purposes.

5. What is the role of the slurmd daemon in a Slurm cluster?

   • A. It manages the job queue on the controller node
   • B. It runs on compute nodes to launch and monitor jobs
   • C. It authenticates users submitting jobs
   • D. It schedules jobs across partitions

   Correct Answer: B. It runs on compute nodes to launch and monitor jobs

   Explanation: The slurmd daemon runs on each compute node and is responsible for launching, monitoring, and reporting job status back to the controller.

6. Which Slurm command provides detailed information about a specific job?

   • A. squeue -j [jobid]
   • B. scontrol show job [jobid]
   • C. sacct -j [jobid]
   • D. sinfo -j [jobid]

   Correct Answer: B. scontrol show job [jobid]

   Explanation: The scontrol show job command gives detailed information about a job’s current state and configuration. squeue shows queue status, and sacct reports historical accounting data.

7. How can an administrator configure Slurm to enforce resource limits such as memory or GPU usage per job?

   • A. By setting limits in the gres.conf file
   • B. By defining resource constraints in job submission scripts and enforcing them via slurm.conf
   • C. By modifying the munge.key file
   • D. By adjusting the slurmd daemon startup parameters

   Correct Answer: B. By defining resource constraints in job submission scripts and enforcing them via slurm.conf

   Explanation: Resource limits such as memory and GPUs are specified in job scripts and enforced by Slurm based on configuration in slurm.conf. The gres.conf file defines generic resources but does not enforce limits.

8. Which of the following best describes the function of the Slurm controller daemon slurmctld?

   • A. It manages job scheduling and resource allocation
   • B. It runs jobs on compute nodes
   • C. It authenticates user credentials
   • D. It monitors network traffic between nodes

   Correct Answer: A. It manages job scheduling and resource allocation

   Explanation: The slurmctld daemon runs on the controller node and is responsible for scheduling jobs, managing resources, and coordinating the cluster.

More in this topic

Configure Multi-Instance GPU (MIG) — Administration (NVIDIA-Certified Professional: AI Operations)Describe datacenter architecture for AI workloads — Administration (NVIDIA-Certified Professional: AI Operations)Administer Slurm clusters: Quick Reference — Administration (NVIDIA-Certified Professional: AI Operations)Administer Run:ai platforms — Administration (NVIDIA-Certified Professional: AI Operations)Administer Slurm clusters: Common Mistakes — Administration (NVIDIA-Certified Professional: AI Operations)Administer Slurm clusters — Administration (NVIDIA-Certified Professional: AI Operations)Administration — NVIDIA-Certified Professional: AI OperationsAdminister Slurm clusters: Worked Example — Administration (NVIDIA-Certified Professional: AI Operations)Administer Kubernetes environments — Administration (NVIDIA-Certified Professional: AI Operations)