NCP-AIO Practice Questions

A.  

The control plane consists of the kube-apiserver, etcd, kube-scheduler, and kube-controller-manager, while worker nodes run kubelet and kube-proxy.

B.  

Worker nodes manage the kube-apiserver and etcd, while the control plane handles all container runtimes.

C.  

The control plane is responsible for running all application containers, while worker nodes manage network traffic through etcd.

D.  

The control plane includes the kubelet and kube-proxy, and worker nodes are responsible for running etcd and the scheduler.

A.  

Ubuntu Server, Docker-CE, NVIDIA Container Toolkit, CSP CLI, NGC CLI, NVIDIA Driver

B.  

Cent OS, Docker-CE, NVIDIA Container Toolkit, CSP CLI, NGC CLI

C.  

Ubuntu Server, Docker-CE, NVIDIA Container Toolkit, CSP CLI, NGC CLI, NVIDIA Driver, Rapids

D.  

Ubuntu Server, Docker-CE, NVIDIA Container Toolkit, CSP CLI, NGC CLI

A.  

Disable NVLink to prevent conflicts between GPUs during data transfer.

B.  

Reduce the size of datasets being processed by splitting them into smaller chunks.

C.  

Increase the swap space on the host system to handle larger datasets.

D.  

Ensure that GPUDirect Storage is configured to allow direct data transfer from storage to GPU memory.

A.  

Drain the compute node using scontrol update.

B.  

Set the node state to down in Slurm before completing maintenance.

C.  

Set the node state to down in Slurm before completing maintenance.

D.  

Disable job scheduling on all compute nodes in Slurm before completing maintenance.

A.  

Increase the number of GPUs requested in the job script to avoid using unconfigured GPUs.

B.  

Check if MIG (Multi-Instance GPU) mode has been enabled incorrectly and reconfigure Slurm accordingly.

C.  

Verify that non-MIG GPUs are automatically configured in Slurm when detected, and adjust configurations if needed.

D.  

Ensure that GPU resource limits have been correctly defined in Slurm’s configuration file for each job type.

A.  

Use BCM's built-in load balancer to distribute workloads evenly between on-premises and cloud resources without any pre-configuration.

B.  

Manually provision additional cloud nodes in AWS when the on-premises cluster reaches its limit.

C.  

Set up a standby deployment in AWS and manually switch workloads to the cloud during peak times.

D.  

Use BCM's Cluster Extension feature to automatically provision AWS resources when local resources are exhausted.

A.  

Load balancing across all nodes in the cluster.

B.  

Manual intervention by the system administrator to restart services.

C.  

The failover mechanism that automatically transfers workloads to a standby node.

D.  

Data replication between nodes to ensure data integrity.

A.  

smap -j

B.  

scontrol show job

C.  

sstat -j

D.  

sinfo -j

A.  

L1 Researcher

B.  

Department Administrator

C.  

Application Administrator

D.  

Research Manager