Manage job scheduling with Slurm or Kubernetes: Practice Questions — Installation and Deployment (NVIDIA-Certified Professional: AI Operations)
Practice Questions: Manage Job Scheduling with Slurm or Kubernetes These multiple-choice questions are designed to help candidates prepare for the...
Practice Questions: Manage Job Scheduling with Slurm or Kubernetes
These multiple-choice questions are designed to help candidates prepare for the Installation and Deployment section of the NVIDIA-Certified Professional: AI Operations exam, focusing specifically on managing job scheduling with Slurm and Kubernetes.
Which component in Slurm is responsible for allocating resources and scheduling jobs on cluster nodes?
- A) slurmd
- B) slurmctld
- C) srun
- D) sacct
Correct answer: B) slurmctld
Explanation: slurmctld is the central management daemon in Slurm that handles job scheduling and resource allocation across the cluster. slurmd runs on compute nodes, srun is a command to launch jobs, and sacct is used for job accounting.
In Kubernetes, which object is primarily used to manage and schedule AI workload containers across nodes?
- A) Pod
- B) Deployment
- C) Service
- D) Namespace
Correct answer: B) Deployment
Explanation: A Deployment manages the desired state and scaling of Pods, which are the smallest deployable units. It schedules Pods across nodes according to resource availability. Pods are individual containers, Services expose Pods, and Namespaces organize resources.
Which Slurm command is used to submit a batch job script to the scheduler?
- A) squeue
- B) sbatch
- C) scancel
- D) sinfo
Correct answer: B) sbatch
Explanation: The sbatch command submits a batch job script to Slurm for scheduling. squeue views job queues, scancel cancels jobs, and sinfo displays cluster node information.
When configuring Kubernetes for AI workloads, which resource request ensures a Pod is scheduled only on nodes with NVIDIA GPUs?
- A) cpu: "1"
- B) memory: "4Gi"
- C) nvidia.com/gpu: "1"
- D) ephemeral-storage: "10Gi"
Correct answer: C) nvidia.com/gpu: "1"
Explanation: Specifying nvidia.com/gpu as a resource request in the Pod specification ensures Kubernetes schedules the Pod on nodes with available NVIDIA GPUs, which is critical for AI workloads.
In Slurm, what is the purpose of the partition configuration?
- A) To define user access permissions
- B) To group nodes for job scheduling policies
- C) To monitor job performance
- D) To update firmware on nodes
Correct answer: B) To group nodes for job scheduling policies
Explanation: A partition in Slurm groups compute nodes and defines scheduling policies such as job priority and resource limits. It does not directly control user permissions or monitoring.
Which Kubernetes component is responsible for monitoring node health and reporting status to the control plane?
- A) kubelet
- B) kube-proxy
- C) etcd
- D) kube-scheduler
Correct answer: A) kubelet
Explanation: The kubelet runs on each node, monitors the node and Pod status, and reports this information to the Kubernetes control plane. kube-proxy manages networking, etcd stores cluster data, and kube-scheduler schedules Pods.
Which Slurm command would you use to view the current job queue and their states?
- A) scontrol
- B) squeue
- C) sacct
- D) sinfo
Correct answer: B) squeue
Explanation: The squeue command displays the current job queue and the status of jobs waiting or running. scontrol manages jobs and nodes, sacct shows job accounting, and sinfo shows node information.
More in this topic
Ready to test your knowledge?
Put what you've learned into practice with a quick quiz and track your progress.
Test your knowledge →