NCP-AIO Administration — All Questions With Answers

Question 1hardmultiple choice

Read the full Administration explanation →

Refer to the exhibit. An administrator encounters this error when attempting to run 'nvidia-smi' inside a privileged container. What is the most likely cause for this failure?

Exhibit

Error Log: [nv_ml_error] Failed to initialize NVML: Insufficient Permissions. Status: 10. Check dmesg for details.

Question 2mediummultiple choice

Read the full Administration explanation →

An administrator is managing a cluster where training jobs are failing due to Out-of-Memory (OOM) errors. What is the most effective approach to troubleshoot the memory distribution across multiple GPUs?

Question 3hardmulti select

Read the full Administration explanation →

Which THREE actions are essential when performing a clean upgrade of NVIDIA drivers in a production AI environment to minimize downtime?

Question 4mediummultiple choice

Read the full Administration explanation →

An administrator needs to implement a policy where only specific authorized users can access GPU resources on a shared cluster. What is the recommended approach?

Question 5mediummultiple choice

Read the full Administration explanation →

An administrator needs to optimize GPU utilization across a multi-tenant NVIDIA DGX cluster. Which action ensures equitable resource allocation while preventing a single container from saturating the NVLink fabric?

Question 6mediummulti select

Read the full Administration explanation →

Which TWO steps are required to ensure the NVIDIA Container Toolkit correctly exposes GPUs to a Docker container on a Linux host?

Question 7hardmultiple choice

Read the full Administration explanation →

An administrator is monitoring GPU memory bandwidth utilization. Which tool provides real-time, per-process visibility into GPU memory usage for troubleshooting memory-bound AI workloads?

Question 8easymultiple choice

Read the full Administration explanation →

When deploying NVIDIA AI Enterprise software, which component serves as the core base for managing and monitoring the infrastructure across a heterogeneous cluster?

Question 9mediummultiple choice

Read the full Administration explanation →

An administrator observes high 'ECC error' counts in the system logs for a specific GPU. What is the most appropriate next step for administrative action?

Question 10hardmultiple choice

Read the full Administration explanation →

An administrator is troubleshooting a GPU-accelerated pod that fails to start with the error 'no NVIDIA devices found'. The GPU Operator is installed and the node has GPUs. Which action should the administrator take first to diagnose the issue?

Question 11hardmultiple choice

Read the full Administration explanation →

An administrator is managing a Kubernetes cluster with NVIDIA GPU Operator installed. A new node with NVIDIA GPUs is added, but pods requesting GPUs remain in Pending state. The administrator runs 'kubectl describe node <node-name>' and sees that the node has no 'nvidia.com/gpu' resource. Which component of the GPU Operator is most likely failing to label the node with GPU resources?

Question 12hardmultiple choice

Read the full Administration explanation →

An administrator is troubleshooting a multi-node NVIDIA AI training job that uses NCCL for inter-GPU communication. The job runs on a cluster with 8 GPUs per node connected via NVLink, and nodes connected via InfiniBand. The administrator observes that NCCL is not utilizing the InfiniBand fabric, and communication falls back to Ethernet. Which action should they take to ensure NCCL uses InfiniBand?

Question 13mediummultiple choice

Read the full Administration explanation →

An AI operations engineer needs to isolate GPU-accelerated workloads on a Kubernetes node so that only pods from a specific namespace can consume GPU resources. The cluster uses the NVIDIA GPU Operator and MIG is enabled on the A100 GPUs. Which Kubernetes resource should be configured to enforce this isolation?

Question 14mediummultiple choice

Read the full Administration explanation →

An administrator is deploying a multi-node NVIDIA Base Command Platform cluster. They need to ensure that the cluster's shared storage is properly configured for AI training workloads. Which storage solution is specifically optimized and recommended by NVIDIA for Base Command Platform to provide high-throughput, low-latency access for GPU-accelerated workloads?

Question 15hardmultiple choice

Read the full network assurance explanation →

A platform engineer manages an NVIDIA AI Enterprise cluster with multiple DGX nodes. The security team requires that GPU telemetry and management traffic be isolated from tenant workload traffic. The cluster uses NVIDIA Base Command Manager for provisioning and monitoring. Which configuration should the administrator implement to meet this requirement while preserving full remote management and telemetry collection?

Question 16mediummultiple choice

Read the full Administration explanation →

An administrator is troubleshooting a DGX system where a training job intermittently fails with XID 48 errors in the system logs. The administrator wants to determine whether the error is related to a specific GPU. Which action should the administrator take?

Question 17mediummultiple choice

Read the full Administration explanation →

An administrator is using NVIDIA AI Enterprise with Run:ai for cluster management. They need to implement a policy that allows only members of the 'data-science' group to submit jobs that request more than 4 GPUs, while other users are limited to 4 GPUs. Which Run:ai feature should they use?

Question 18mediummultiple choice

Read the full Administration explanation →

An administrator manages a Kubernetes cluster running NVIDIA GPU Operator. A data science team reports that their PyTorch training pods are stuck in Pending state with the event '0/8 nodes are available: 8 Insufficient nvidia.com/gpu.' The administrator verifies that the nodes have GPUs and that the NVIDIA device plugin pods are running. Which command should the administrator run first to diagnose why the GPU resource is not being advertised?

Question 19mediummultiple choice

Read the full Administration explanation →

An administrator is responsible for a Kubernetes cluster running multiple inference workloads that share GPU nodes. The workloads belong to different teams and have varying performance requirements, but the administrator wants to prevent a single low-priority job from monopolizing a GPU and starving other pods. Which NVIDIA feature should be configured to enforce fair sharing of a single physical GPU among multiple containers?

Question 20hardmultiple choice

Read the full Administration explanation →

An administrator manages an NVIDIA AI Enterprise deployment on Kubernetes. A data science team reports that their training pods are stuck in Pending state with the event '0/8 nodes are available: 8 Insufficient nvidia.com/gpu'. The administrator confirms that GPUs are present and healthy on all nodes, and the NVIDIA GPU Operator is installed. Which action should the administrator take first to resolve the scheduling failure?

Question 21mediummultiple choice

Read the full Administration explanation →

An administrator manages a Kubernetes cluster running AI workloads with NVIDIA GPUs. They need to enforce that only pods with a specific label can access GPU resources, and that GPU usage is tracked per namespace. Which combination of Kubernetes and NVIDIA components should they use?

Question 22mediummultiple choice

Read the full Administration explanation →

An AI operations engineer is preparing a bare-metal DGX H100 node for a new multi-node training cluster. The node must be consistently provisioned with the same NVIDIA driver, CUDA, and container runtime stack as the rest of the fleet, and the team wants to avoid manual per-node drift. Which NVIDIA-provided approach should the engineer use to install and manage the full software stack on this node?

Question 23mediummulti select

Read the full Administration explanation →

An administrator is configuring a Kubernetes cluster with NVIDIA GPU Operator to support multi-instance GPU (MIG) on A100 GPUs. They need to ensure that MIG instances are properly advertised and can be requested by pods. Which two configurations are required to enable MIG support in this environment? (Choose two.)

Question 24mediummultiple choice

Read the full Administration explanation →

An administrator manages a Kubernetes cluster running inference workloads on NVIDIA GPUs. The cluster uses NVIDIA GPU Operator with MIG-enabled A100 GPUs. A data science team reports that their pod is stuck in Pending state with the event '0/8 nodes are available: 8 Insufficient nvidia.com/mig-1g.5gb.' The administrator verifies that each node has seven MIG instances of 1g.5gb available and no other pods are using them. Which action should the administrator take to resolve the pending pod?

Question 25easymultiple choice

Read the full Administration explanation →

An administrator is deploying NVIDIA AI Enterprise on a vSphere cluster with multiple ESXi hosts, each containing several A100 GPUs. The administrator needs to enable GPU sharing among multiple virtual machines (VMs) on the same host to maximize utilization for inference workloads. Which NVIDIA vGPU mode should be configured?

Question 26hardmultiple choice

Read the full Administration explanation →

An administrator is troubleshooting a multi-node NVIDIA GPU cluster where NCCL-based training jobs intermittently hang. They suspect network issues. Which diagnostic step should they perform first to isolate the problem?

Question 27easymultiple choice

Read the full Administration explanation →

An administrator is setting up a new NVIDIA AI Enterprise environment and needs to ensure that the NVIDIA GPU Operator is installed and functioning correctly on a Kubernetes cluster. After installation, which command should the administrator use to verify that all GPU Operator components are healthy and that the GPU is available as a schedulable resource?

Question 28mediummulti select

Read the full Administration explanation →

A platform administrator is configuring NVIDIA MIG (Multi-Instance GPU) on an A100 node that will host multiple small inference tenants. The administrator wants to ensure that each tenant gets an isolated GPU instance with dedicated memory and compute slices, and that the configuration survives a reboot. Which two actions are required? (Choose two.)

Question 29hardmulti select

Read the full Administration explanation →

An administrator is configuring NVIDIA MIG (Multi-Instance GPU) on an A100 GPU to support multiple isolated tenants on a single physical GPU. The administrator wants to ensure that each tenant receives a dedicated slice of memory and compute resources with strict isolation. Which two actions are required to correctly set up MIG for this purpose? (Choose two.)

Question 30easymultiple choice

Read the full network assurance explanation →

An administrator is responsible for managing an NVIDIA DGX SuperPOD used for large-scale AI training. The administrator needs to perform routine health checks and gather telemetry from all DGX nodes, including GPU temperature, power usage, and ECC error counts. Which NVIDIA tool provides a centralized, cluster-wide view of this telemetry and health status?

Question 31hardmulti select

Read the full Administration explanation →

An administrator is configuring NVIDIA MIG (Multi-Instance GPU) on an A100 GPU for a multi-tenant AI inference environment. They need to ensure that each tenant gets isolated GPU resources and that memory bandwidth is partitioned. Which TWO actions must they perform? (Choose two.)

Question 32hardmultiple choice

Read the full Administration explanation →

An administrator is setting up a shared DGX SuperPOD partition for multiple research groups. The requirement is that each group's jobs run with a guaranteed share of GPU resources and that no group can monopolize the partition during peak hours. The administrator also needs to enforce fair-share scheduling and preemption according to group quotas. Which NVIDIA software component should be configured to meet these requirements?

Question 33easymultiple choice

Read the full Administration explanation →

An administrator needs to monitor GPU temperature and power usage across a cluster of NVIDIA GPUs in real time. Which NVIDIA tool provides this capability with a web-based dashboard?

Question 34mediummultiple choice

Read the full Administration explanation →

An administrator is configuring a multi-tenant NVIDIA AI Enterprise environment. Which mechanism is most effective for ensuring hardware-level isolation between concurrent training jobs on a single A100 GPU?

Question 35mediummulti select

Read the full Administration explanation →

An administrator is optimizing a cluster for AI model training using NVIDIA Base Command. Which TWO tasks are critical for ensuring consistent performance across the training nodes?

Question 36hardmultiple choice

Read the full Administration explanation →

Refer to the exhibit. An administrator applies this security policy to a container runtime environment. What is the immediate effect on containerized AI applications within this scope?

Exhibit

{ "policy": "deny", "resources": ["/dev/nvidia*"], "action": "restrict_access" }

Question 37easymultiple choice

Read the full Administration explanation →

When monitoring GPU health in an enterprise cluster, which command provides the most comprehensive snapshot of real-time power, temperature, and memory utilization?

Question 38mediummultiple choice

Read the full Administration explanation →

An administrator notices that GPU utilization is high, but throughput in an AI training job remains low. What is the most likely bottleneck?

Question 39hardmultiple choice

Read the full Administration explanation →

Refer to the exhibit. An administrator observes this in the logs during a multi-GPU training job. What is the performance implication of this setting?

Exhibit

Error Log: [NCCL] NCCL_P2P_DISABLE=1 detected. Using network-based communication instead of NVLink.

Question 40easymultiple choice

Read the full Administration explanation →

Which NVIDIA technology enables a GPU to be shared among multiple virtual machines or containers while maintaining strict hardware isolation?

Question 41mediummultiple choice

Read the full Administration explanation →

An administrator needs to ensure that all GPU drivers are updated across a heterogeneous cluster without causing downtime. What is the best strategy?

Question 42hardmulti select

Read the full Administration explanation →

Which THREE components are required for a container to successfully leverage NVIDIA GPUs on a Kubernetes cluster?

Question 43mediummultiple choice

Read the full network assurance explanation →

An administrator is configuring a new cluster and wants to ensure that telemetry data from GPUs is collected in a centralized manner. Which tool is best suited for this requirement?

Question 44mediummultiple choice

Read the full Administration explanation →

An administrator wants to prevent unauthorized users from accessing sensitive model weights stored in GPU memory. Which security feature should be implemented to ensure hardware-level isolation of the memory space?

Question 45mediummultiple choice

Read the full Administration explanation →

Which administrative practice ensures that a cluster is prepared for the arrival of new NVIDIA GPU hardware with minimal downtime?

Question 46mediummultiple choice

Read the full Administration explanation →

An administrator manages an NVIDIA AI Enterprise cluster running multiple Kubernetes nodes, each with several A100 GPUs. After upgrading the NVIDIA GPU Operator to a newer version, the administrator notices that pods requesting GPUs remain in a Pending state, and the node's allocatable GPU count is reported as zero. Which command should the administrator run first to diagnose the issue?

Question 47mediummultiple choice

Read the full Administration explanation →

An administrator manages an NVIDIA AI Enterprise cluster and needs to enforce GPU resource quotas across multiple Kubernetes namespaces. Which NVIDIA component should be configured to enforce these quotas?

Question 48easymultiple choice

Read the full Administration explanation →

An administrator is deploying NVIDIA AI Enterprise on a Kubernetes cluster and wants to verify that the GPU Operator has successfully installed all required components. Which command should the administrator use to check the status of the GPU Operator pods?

Question 49mediummultiple choice

Read the full Administration explanation →

An administrator manages an NVIDIA AI Enterprise cluster using NVIDIA Run:ai. A data science team complains that their submitted training job has been stuck in a Pending state for over an hour, even though the Run:ai scheduler shows free GPUs in the cluster. The administrator verifies that the job requests 2 GPUs and the node pool has 4 idle GPUs. Which Run:ai administrative configuration is the most likely cause of the job remaining Pending?

Question 50hardmultiple choice

Read the full Administration explanation →

An administrator is configuring NVIDIA Base Command Manager to manage a cluster of DGX nodes. They want to ensure that when a node's GPU temperature exceeds a defined threshold, the node is automatically drained and an alert is sent to the operations team. Which combination of Base Command Manager features should the administrator configure to achieve this?

Question 51mediummultiple choice

Read the full Administration explanation →

An administrator is preparing a bare-metal GPU server for AI workloads and needs to verify that the NVIDIA driver and CUDA toolkit are properly installed. The server has an NVIDIA A100 GPU. Which command should the administrator run to display the GPU model, driver version, and CUDA version?

Question 52mediummulti select

Read the full Administration explanation →

An administrator is responsible for maintaining an NVIDIA AI Enterprise cluster and needs to ensure high availability of GPU resources for critical inference workloads. Which two practices should the administrator implement? (Choose two.)

Question 53easymultiple choice

Read the full Administration explanation →

An administrator is deploying NVIDIA AI Enterprise on a vSphere cluster with multiple ESXi hosts, each containing several NVIDIA A100 GPUs. They need to enable GPU sharing across multiple virtual machines to maximize utilization. Which vSphere feature should they configure?

Question 54mediummultiple choice

Read the full Administration explanation →

An administrator is preparing a multi-node NVIDIA DGX H100 cluster for a distributed training job using NVIDIA Base Command. The cluster nodes have InfiniBand adapters, but the job's inter-node throughput is far below expectations. The administrator runs `ibstat` and sees that the ports are in the INIT state rather than ACTIVE. Which action should the administrator take first?

Question 55mediummultiple choice

Read the full network assurance explanation →

An administrator needs to collect GPU telemetry from an NVIDIA AI Enterprise cluster and store it in a time-series database for long-term analysis. Which component should be deployed to export GPU metrics in Prometheus format?

Question 56hardmultiple choice

Read the full network assurance explanation →

An administrator manages an NVIDIA AI Enterprise deployment on Kubernetes and needs to ensure that GPU telemetry is exported to an existing Prometheus instance. The administrator deploys the NVIDIA DCGM Exporter but sees no GPU metrics in Prometheus. Which configuration should the administrator verify first?

Question 57mediummultiple choice

Read the full Administration explanation →

An administrator supports a shared Kubernetes cluster running NVIDIA GPU Operator. Data scientists report that their inference pods remain in Pending state, yet the GPU Operator pods and node feature discovery pods are healthy, and the GPU nodes show no hardware alarms. The administrator confirms that the cluster has a mixture of MIG-capable A100 nodes and non-MIG T4 nodes. Which immediate administrative action is most appropriate to diagnose the scheduling failure?

Question 58easymultiple choice

Read the full Administration explanation →

An administrator is setting up an NVIDIA AI Enterprise cluster and wants to verify that the NVIDIA GPU Operator has successfully deployed all required components on a worker node. Which command should the administrator use to list the GPU Operator pods running on that node?

Question 59mediummultiple choice

Read the full Administration explanation →

An AI operations engineer is preparing a DGX H100 system for a multi-node training workload. The engineer runs `nvidia-smi topo -m` and notices that GPU4 and GPU5 report a connection type of SYS, while all other GPU pairs show NV18. What is the most likely cause of this topology anomaly?

Question 60easymultiple choice

Read the full Administration explanation →

An administrator is setting up NVIDIA Base Command Manager to provision and manage a new AI cluster. They need to ensure that the cluster can automatically discover and configure new GPU nodes. Which component is responsible for node discovery and initial configuration?

Question 61mediummulti select

Read the full Administration explanation →

An administrator is preparing a cluster for a new large language model training job that will use NVIDIA Magnum IO GPUDirect Storage to stream training data directly from a parallel file system to GPU memory. The administrator must verify that the environment supports GPUDirect Storage before the job starts. (Choose two.)

Question 62easymultiple choice

Read the full network assurance explanation →

An administrator is responsible for monitoring a large-scale AI cluster with hundreds of NVIDIA GPUs. They need to collect telemetry data such as GPU utilization, temperature, and power consumption from all nodes and store it centrally for analysis and alerting. Which NVIDIA tool should they use to collect and export GPU metrics to a monitoring system like Prometheus?

Question 63mediummulti select

Read the full Administration explanation →

An AI operations team is using NVIDIA DCGM (Data Center GPU Manager) to monitor a cluster of A100 GPUs. They want to set up proactive health checks to detect and mitigate GPU issues before they cause job failures. Which two DCGM features should they configure? (Choose two.)

Question 64mediummultiple choice

Read the full Administration explanation →

An administrator is troubleshooting a multi-node NVIDIA GPU training job that intermittently hangs during the all-reduce phase. The job uses NCCL over InfiniBand. Logs show that some ranks time out while others complete. The administrator suspects a network fabric issue. Which action should the administrator take first to isolate whether the problem is in the InfiniBand fabric or in the NCCL configuration?

Question 65hardmultiple choice

Read the full network assurance explanation →

An administrator is responsible for a large NVIDIA DGX SuperPOD used for multi-node training. They need to ensure that GPU telemetry and health metrics are collected centrally and can trigger alerts when GPUs exceed temperature thresholds. Which component of NVIDIA Base Command Manager (BCM) should they configure to achieve this?

Question 66mediummultiple choice

Read the full Administration explanation →

An AI operations team is deploying NVIDIA Base Command Manager to manage a cluster of DGX nodes. They want to ensure that only authorized users can submit jobs and that all job submissions are audited. Which combination of Base Command Manager features should the administrator configure to meet these requirements?

Question 67mediummultiple choice

Read the full Administration explanation →

A company runs multiple AI workloads on a shared Kubernetes cluster with NVIDIA GPUs. The administrator needs to enforce that only pods with a specific label can consume GPU resources, while other pods are denied. Which Kubernetes admission control mechanism should be used to implement this policy?

Question 68hardmultiple choice

Read the full Administration explanation →

A site reliability engineer is troubleshooting a DGX A100 node that intermittently drops out of the cluster during large NCCL all-reduce jobs. `nvidia-smi` shows all eight A100 GPUs healthy, but DCGM reports XID errors 74 and 79 on one GPU during the failures. The engineer needs to determine the most likely cause and the correct administrative action. Which combination best describes the cause and the appropriate first step?

Question 69easymultiple choice

Read the full Administration explanation →

An administrator is responsible for maintaining a fleet of NVIDIA-certified servers running AI workloads. They need to quickly identify which servers have GPUs that are overheating and may throttle performance. Which NVIDIA tool should the administrator use to monitor GPU temperature across the fleet in real time?

Question 70mediummultiple choice

Read the full Administration explanation →

An administrator is managing an NVIDIA DGX SuperPOD used for large-scale AI training. The cluster uses a Slurm workload manager. The administrator needs to ensure that jobs are scheduled only on nodes with healthy GPUs and that failed GPUs are automatically drained from the pool. Which integration should be configured to achieve this?

Question 71mediummultiple choice

Read the full Administration explanation →

An administrator is using NVIDIA Base Command Manager to provision a new GPU cluster. They need to ensure that the compute nodes are configured with the correct GPU driver and CUDA toolkit versions. Which Base Command Manager feature should they use to automate this?

Question 72hardmultiple choice

Read the full Administration explanation →

An administrator is responsible for an NVIDIA AI Enterprise deployment on Kubernetes. The security team requires that all GPU-accelerated pods run with the least privilege necessary and that GPU device nodes are not exposed to pods that do not request them. Which combination of configurations should the administrator implement to meet these requirements?

Question 73easymultiple choice

Read the full network assurance explanation →

An AI operations team needs to monitor GPU health and utilization across a fleet of DGX nodes from a single dashboard. They want per-GPU metrics such as power, temperature, utilization, and ECC errors, and they want to retain historical data for capacity planning. Which NVIDIA tool is purpose-built to collect and expose these GPU telemetry metrics for centralized monitoring?

Question 74hardmulti select

Read the full Administration explanation →

An administrator is configuring NVIDIA GPUDirect Storage (GDS) on a cluster to accelerate data loading for AI training jobs. The cluster uses Mellanox InfiniBand adapters and NVMe storage. Which two actions are required to enable GDS and ensure optimal performance? (Choose two.)

Question 75mediummultiple choice

Read the full Administration explanation →

An administrator is managing a Kubernetes cluster with NVIDIA GPU Operator. A new policy requires that all GPU workloads run with specific environment variables set, such as NVIDIA_VISIBLE_DEVICES and NVIDIA_DRIVER_CAPABILITIES. The administrator wants to enforce these variables automatically for any pod that requests a GPU, without modifying each pod specification manually. Which approach should the administrator use?

Question 76hardmultiple choice

Read the full Administration explanation →

An administrator is troubleshooting a performance degradation in a multi-node NVIDIA NCCL-based training job. The job spans four DGX nodes connected via InfiniBand. The administrator suspects that NCCL is not using the optimal network path. Which action should the administrator take to verify and enforce the use of GPUDirect RDMA for inter-node communication?

Question 77hardmulti select

Read the full Administration explanation →

An administrator is configuring an NVIDIA AI Enterprise cluster to run multi-tenant inference workloads on Kubernetes. The administrator must ensure that GPU resources are isolated and that tenants cannot access each other's GPU memory. Which two actions should the administrator take? (Choose two.)

Question 1hardmultiple choice

Read the full Administration explanation →

Refer to the exhibit. An administrator encounters this error when attempting to run 'nvidia-smi' inside a privileged container. What is the most likely cause for this failure?

Exhibit

Error Log: [nv_ml_error] Failed to initialize NVML: Insufficient Permissions. Status: 10. Check dmesg for details.

Question 2mediummultiple choice

Read the full Administration explanation →

An administrator is managing a cluster where training jobs are failing due to Out-of-Memory (OOM) errors. What is the most effective approach to troubleshoot the memory distribution across multiple GPUs?

Question 3hardmulti select

Read the full Administration explanation →

Which THREE actions are essential when performing a clean upgrade of NVIDIA drivers in a production AI environment to minimize downtime?

Question 4mediummultiple choice

Read the full Administration explanation →

An administrator needs to implement a policy where only specific authorized users can access GPU resources on a shared cluster. What is the recommended approach?

Question 5mediummultiple choice

Read the full Administration explanation →

An administrator needs to optimize GPU utilization across a multi-tenant NVIDIA DGX cluster. Which action ensures equitable resource allocation while preventing a single container from saturating the NVLink fabric?

Question 6mediummulti select

Read the full Administration explanation →

Which TWO steps are required to ensure the NVIDIA Container Toolkit correctly exposes GPUs to a Docker container on a Linux host?

Question 7hardmultiple choice

Read the full Administration explanation →

An administrator is monitoring GPU memory bandwidth utilization. Which tool provides real-time, per-process visibility into GPU memory usage for troubleshooting memory-bound AI workloads?

Question 8easymultiple choice

Read the full Administration explanation →

When deploying NVIDIA AI Enterprise software, which component serves as the core base for managing and monitoring the infrastructure across a heterogeneous cluster?

Question 9mediummultiple choice

Read the full Administration explanation →

An administrator observes high 'ECC error' counts in the system logs for a specific GPU. What is the most appropriate next step for administrative action?

Question 10hardmultiple choice

Read the full Administration explanation →

An administrator is troubleshooting a GPU-accelerated pod that fails to start with the error 'no NVIDIA devices found'. The GPU Operator is installed and the node has GPUs. Which action should the administrator take first to diagnose the issue?

Question 11hardmultiple choice

Read the full Administration explanation →

An administrator is managing a Kubernetes cluster with NVIDIA GPU Operator installed. A new node with NVIDIA GPUs is added, but pods requesting GPUs remain in Pending state. The administrator runs 'kubectl describe node <node-name>' and sees that the node has no 'nvidia.com/gpu' resource. Which component of the GPU Operator is most likely failing to label the node with GPU resources?

Question 12hardmultiple choice

Read the full Administration explanation →

An administrator is troubleshooting a multi-node NVIDIA AI training job that uses NCCL for inter-GPU communication. The job runs on a cluster with 8 GPUs per node connected via NVLink, and nodes connected via InfiniBand. The administrator observes that NCCL is not utilizing the InfiniBand fabric, and communication falls back to Ethernet. Which action should they take to ensure NCCL uses InfiniBand?

Question 13mediummultiple choice

Read the full Administration explanation →

An AI operations engineer needs to isolate GPU-accelerated workloads on a Kubernetes node so that only pods from a specific namespace can consume GPU resources. The cluster uses the NVIDIA GPU Operator and MIG is enabled on the A100 GPUs. Which Kubernetes resource should be configured to enforce this isolation?

Question 14mediummultiple choice

Read the full Administration explanation →

An administrator is deploying a multi-node NVIDIA Base Command Platform cluster. They need to ensure that the cluster's shared storage is properly configured for AI training workloads. Which storage solution is specifically optimized and recommended by NVIDIA for Base Command Platform to provide high-throughput, low-latency access for GPU-accelerated workloads?

Question 15hardmultiple choice

Read the full network assurance explanation →

A platform engineer manages an NVIDIA AI Enterprise cluster with multiple DGX nodes. The security team requires that GPU telemetry and management traffic be isolated from tenant workload traffic. The cluster uses NVIDIA Base Command Manager for provisioning and monitoring. Which configuration should the administrator implement to meet this requirement while preserving full remote management and telemetry collection?

Question 16mediummultiple choice

Read the full Administration explanation →

An administrator is troubleshooting a DGX system where a training job intermittently fails with XID 48 errors in the system logs. The administrator wants to determine whether the error is related to a specific GPU. Which action should the administrator take?

Question 17mediummultiple choice

Read the full Administration explanation →

An administrator is using NVIDIA AI Enterprise with Run:ai for cluster management. They need to implement a policy that allows only members of the 'data-science' group to submit jobs that request more than 4 GPUs, while other users are limited to 4 GPUs. Which Run:ai feature should they use?

Question 18mediummultiple choice

Read the full Administration explanation →

An administrator manages a Kubernetes cluster running NVIDIA GPU Operator. A data science team reports that their PyTorch training pods are stuck in Pending state with the event '0/8 nodes are available: 8 Insufficient nvidia.com/gpu.' The administrator verifies that the nodes have GPUs and that the NVIDIA device plugin pods are running. Which command should the administrator run first to diagnose why the GPU resource is not being advertised?

Question 19mediummultiple choice

Read the full Administration explanation →

An administrator is responsible for a Kubernetes cluster running multiple inference workloads that share GPU nodes. The workloads belong to different teams and have varying performance requirements, but the administrator wants to prevent a single low-priority job from monopolizing a GPU and starving other pods. Which NVIDIA feature should be configured to enforce fair sharing of a single physical GPU among multiple containers?

Question 20hardmultiple choice

Read the full Administration explanation →

An administrator manages an NVIDIA AI Enterprise deployment on Kubernetes. A data science team reports that their training pods are stuck in Pending state with the event '0/8 nodes are available: 8 Insufficient nvidia.com/gpu'. The administrator confirms that GPUs are present and healthy on all nodes, and the NVIDIA GPU Operator is installed. Which action should the administrator take first to resolve the scheduling failure?

Question 21mediummultiple choice

Read the full Administration explanation →

An administrator manages a Kubernetes cluster running AI workloads with NVIDIA GPUs. They need to enforce that only pods with a specific label can access GPU resources, and that GPU usage is tracked per namespace. Which combination of Kubernetes and NVIDIA components should they use?

Question 22mediummultiple choice

Read the full Administration explanation →

An AI operations engineer is preparing a bare-metal DGX H100 node for a new multi-node training cluster. The node must be consistently provisioned with the same NVIDIA driver, CUDA, and container runtime stack as the rest of the fleet, and the team wants to avoid manual per-node drift. Which NVIDIA-provided approach should the engineer use to install and manage the full software stack on this node?

Question 23mediummulti select

Read the full Administration explanation →

An administrator is configuring a Kubernetes cluster with NVIDIA GPU Operator to support multi-instance GPU (MIG) on A100 GPUs. They need to ensure that MIG instances are properly advertised and can be requested by pods. Which two configurations are required to enable MIG support in this environment? (Choose two.)

Question 24mediummultiple choice

Read the full Administration explanation →

An administrator manages a Kubernetes cluster running inference workloads on NVIDIA GPUs. The cluster uses NVIDIA GPU Operator with MIG-enabled A100 GPUs. A data science team reports that their pod is stuck in Pending state with the event '0/8 nodes are available: 8 Insufficient nvidia.com/mig-1g.5gb.' The administrator verifies that each node has seven MIG instances of 1g.5gb available and no other pods are using them. Which action should the administrator take to resolve the pending pod?

Question 25easymultiple choice

Read the full Administration explanation →

An administrator is deploying NVIDIA AI Enterprise on a vSphere cluster with multiple ESXi hosts, each containing several A100 GPUs. The administrator needs to enable GPU sharing among multiple virtual machines (VMs) on the same host to maximize utilization for inference workloads. Which NVIDIA vGPU mode should be configured?

Question 26hardmultiple choice

Read the full Administration explanation →

An administrator is troubleshooting a multi-node NVIDIA GPU cluster where NCCL-based training jobs intermittently hang. They suspect network issues. Which diagnostic step should they perform first to isolate the problem?

Question 27easymultiple choice

Read the full Administration explanation →

An administrator is setting up a new NVIDIA AI Enterprise environment and needs to ensure that the NVIDIA GPU Operator is installed and functioning correctly on a Kubernetes cluster. After installation, which command should the administrator use to verify that all GPU Operator components are healthy and that the GPU is available as a schedulable resource?

Question 28mediummulti select

Read the full Administration explanation →

A platform administrator is configuring NVIDIA MIG (Multi-Instance GPU) on an A100 node that will host multiple small inference tenants. The administrator wants to ensure that each tenant gets an isolated GPU instance with dedicated memory and compute slices, and that the configuration survives a reboot. Which two actions are required? (Choose two.)

Question 29hardmulti select

Read the full Administration explanation →

An administrator is configuring NVIDIA MIG (Multi-Instance GPU) on an A100 GPU to support multiple isolated tenants on a single physical GPU. The administrator wants to ensure that each tenant receives a dedicated slice of memory and compute resources with strict isolation. Which two actions are required to correctly set up MIG for this purpose? (Choose two.)

Question 30easymultiple choice

Read the full network assurance explanation →

An administrator is responsible for managing an NVIDIA DGX SuperPOD used for large-scale AI training. The administrator needs to perform routine health checks and gather telemetry from all DGX nodes, including GPU temperature, power usage, and ECC error counts. Which NVIDIA tool provides a centralized, cluster-wide view of this telemetry and health status?

Question 31hardmulti select

Read the full Administration explanation →

An administrator is configuring NVIDIA MIG (Multi-Instance GPU) on an A100 GPU for a multi-tenant AI inference environment. They need to ensure that each tenant gets isolated GPU resources and that memory bandwidth is partitioned. Which TWO actions must they perform? (Choose two.)

Question 32hardmultiple choice

Read the full Administration explanation →

An administrator is setting up a shared DGX SuperPOD partition for multiple research groups. The requirement is that each group's jobs run with a guaranteed share of GPU resources and that no group can monopolize the partition during peak hours. The administrator also needs to enforce fair-share scheduling and preemption according to group quotas. Which NVIDIA software component should be configured to meet these requirements?

Question 33easymultiple choice

Read the full Administration explanation →

An administrator needs to monitor GPU temperature and power usage across a cluster of NVIDIA GPUs in real time. Which NVIDIA tool provides this capability with a web-based dashboard?

Question 34mediummultiple choice

Read the full Administration explanation →

An administrator is configuring a multi-tenant NVIDIA AI Enterprise environment. Which mechanism is most effective for ensuring hardware-level isolation between concurrent training jobs on a single A100 GPU?

Question 35mediummulti select

Read the full Administration explanation →

An administrator is optimizing a cluster for AI model training using NVIDIA Base Command. Which TWO tasks are critical for ensuring consistent performance across the training nodes?

Question 36hardmultiple choice

Read the full Administration explanation →

Refer to the exhibit. An administrator applies this security policy to a container runtime environment. What is the immediate effect on containerized AI applications within this scope?

Exhibit

{ "policy": "deny", "resources": ["/dev/nvidia*"], "action": "restrict_access" }

Question 37easymultiple choice

Read the full Administration explanation →

When monitoring GPU health in an enterprise cluster, which command provides the most comprehensive snapshot of real-time power, temperature, and memory utilization?

Question 38mediummultiple choice

Read the full Administration explanation →

An administrator notices that GPU utilization is high, but throughput in an AI training job remains low. What is the most likely bottleneck?

Question 39hardmultiple choice

Read the full Administration explanation →

Refer to the exhibit. An administrator observes this in the logs during a multi-GPU training job. What is the performance implication of this setting?

Exhibit

Error Log: [NCCL] NCCL_P2P_DISABLE=1 detected. Using network-based communication instead of NVLink.

Question 40easymultiple choice

Read the full Administration explanation →

Which NVIDIA technology enables a GPU to be shared among multiple virtual machines or containers while maintaining strict hardware isolation?

Question 41mediummultiple choice

Read the full Administration explanation →

An administrator needs to ensure that all GPU drivers are updated across a heterogeneous cluster without causing downtime. What is the best strategy?

Question 42hardmulti select

Read the full Administration explanation →

Which THREE components are required for a container to successfully leverage NVIDIA GPUs on a Kubernetes cluster?

Question 43mediummultiple choice

Read the full network assurance explanation →

An administrator is configuring a new cluster and wants to ensure that telemetry data from GPUs is collected in a centralized manner. Which tool is best suited for this requirement?

Question 44mediummultiple choice

Read the full Administration explanation →

An administrator wants to prevent unauthorized users from accessing sensitive model weights stored in GPU memory. Which security feature should be implemented to ensure hardware-level isolation of the memory space?

Question 45mediummultiple choice

Read the full Administration explanation →

Which administrative practice ensures that a cluster is prepared for the arrival of new NVIDIA GPU hardware with minimal downtime?

Question 46mediummultiple choice

Read the full Administration explanation →

An administrator manages an NVIDIA AI Enterprise cluster running multiple Kubernetes nodes, each with several A100 GPUs. After upgrading the NVIDIA GPU Operator to a newer version, the administrator notices that pods requesting GPUs remain in a Pending state, and the node's allocatable GPU count is reported as zero. Which command should the administrator run first to diagnose the issue?

Question 47mediummultiple choice

Read the full Administration explanation →

An administrator manages an NVIDIA AI Enterprise cluster and needs to enforce GPU resource quotas across multiple Kubernetes namespaces. Which NVIDIA component should be configured to enforce these quotas?

Question 48easymultiple choice

Read the full Administration explanation →

An administrator is deploying NVIDIA AI Enterprise on a Kubernetes cluster and wants to verify that the GPU Operator has successfully installed all required components. Which command should the administrator use to check the status of the GPU Operator pods?

Question 49mediummultiple choice

Read the full Administration explanation →

An administrator manages an NVIDIA AI Enterprise cluster using NVIDIA Run:ai. A data science team complains that their submitted training job has been stuck in a Pending state for over an hour, even though the Run:ai scheduler shows free GPUs in the cluster. The administrator verifies that the job requests 2 GPUs and the node pool has 4 idle GPUs. Which Run:ai administrative configuration is the most likely cause of the job remaining Pending?

Question 50hardmultiple choice

Read the full Administration explanation →

An administrator is configuring NVIDIA Base Command Manager to manage a cluster of DGX nodes. They want to ensure that when a node's GPU temperature exceeds a defined threshold, the node is automatically drained and an alert is sent to the operations team. Which combination of Base Command Manager features should the administrator configure to achieve this?

Question 51mediummultiple choice

Read the full Administration explanation →

An administrator is preparing a bare-metal GPU server for AI workloads and needs to verify that the NVIDIA driver and CUDA toolkit are properly installed. The server has an NVIDIA A100 GPU. Which command should the administrator run to display the GPU model, driver version, and CUDA version?

Question 52mediummulti select

Read the full Administration explanation →

An administrator is responsible for maintaining an NVIDIA AI Enterprise cluster and needs to ensure high availability of GPU resources for critical inference workloads. Which two practices should the administrator implement? (Choose two.)

Question 53easymultiple choice

Read the full Administration explanation →

An administrator is deploying NVIDIA AI Enterprise on a vSphere cluster with multiple ESXi hosts, each containing several NVIDIA A100 GPUs. They need to enable GPU sharing across multiple virtual machines to maximize utilization. Which vSphere feature should they configure?

Question 54mediummultiple choice

Read the full Administration explanation →

An administrator is preparing a multi-node NVIDIA DGX H100 cluster for a distributed training job using NVIDIA Base Command. The cluster nodes have InfiniBand adapters, but the job's inter-node throughput is far below expectations. The administrator runs `ibstat` and sees that the ports are in the INIT state rather than ACTIVE. Which action should the administrator take first?

Question 55mediummultiple choice

Read the full network assurance explanation →

An administrator needs to collect GPU telemetry from an NVIDIA AI Enterprise cluster and store it in a time-series database for long-term analysis. Which component should be deployed to export GPU metrics in Prometheus format?

Question 56hardmultiple choice

Read the full network assurance explanation →

An administrator manages an NVIDIA AI Enterprise deployment on Kubernetes and needs to ensure that GPU telemetry is exported to an existing Prometheus instance. The administrator deploys the NVIDIA DCGM Exporter but sees no GPU metrics in Prometheus. Which configuration should the administrator verify first?

Question 57mediummultiple choice

Read the full Administration explanation →

An administrator supports a shared Kubernetes cluster running NVIDIA GPU Operator. Data scientists report that their inference pods remain in Pending state, yet the GPU Operator pods and node feature discovery pods are healthy, and the GPU nodes show no hardware alarms. The administrator confirms that the cluster has a mixture of MIG-capable A100 nodes and non-MIG T4 nodes. Which immediate administrative action is most appropriate to diagnose the scheduling failure?

Question 58easymultiple choice

Read the full Administration explanation →

An administrator is setting up an NVIDIA AI Enterprise cluster and wants to verify that the NVIDIA GPU Operator has successfully deployed all required components on a worker node. Which command should the administrator use to list the GPU Operator pods running on that node?

Question 59mediummultiple choice

Read the full Administration explanation →

An AI operations engineer is preparing a DGX H100 system for a multi-node training workload. The engineer runs `nvidia-smi topo -m` and notices that GPU4 and GPU5 report a connection type of SYS, while all other GPU pairs show NV18. What is the most likely cause of this topology anomaly?

Question 60easymultiple choice

Read the full Administration explanation →

An administrator is setting up NVIDIA Base Command Manager to provision and manage a new AI cluster. They need to ensure that the cluster can automatically discover and configure new GPU nodes. Which component is responsible for node discovery and initial configuration?

Question 61mediummulti select

Read the full Administration explanation →

An administrator is preparing a cluster for a new large language model training job that will use NVIDIA Magnum IO GPUDirect Storage to stream training data directly from a parallel file system to GPU memory. The administrator must verify that the environment supports GPUDirect Storage before the job starts. (Choose two.)

Question 62easymultiple choice

Read the full network assurance explanation →

An administrator is responsible for monitoring a large-scale AI cluster with hundreds of NVIDIA GPUs. They need to collect telemetry data such as GPU utilization, temperature, and power consumption from all nodes and store it centrally for analysis and alerting. Which NVIDIA tool should they use to collect and export GPU metrics to a monitoring system like Prometheus?

Question 63mediummulti select

Read the full Administration explanation →

An AI operations team is using NVIDIA DCGM (Data Center GPU Manager) to monitor a cluster of A100 GPUs. They want to set up proactive health checks to detect and mitigate GPU issues before they cause job failures. Which two DCGM features should they configure? (Choose two.)

Question 64mediummultiple choice

Read the full Administration explanation →

An administrator is troubleshooting a multi-node NVIDIA GPU training job that intermittently hangs during the all-reduce phase. The job uses NCCL over InfiniBand. Logs show that some ranks time out while others complete. The administrator suspects a network fabric issue. Which action should the administrator take first to isolate whether the problem is in the InfiniBand fabric or in the NCCL configuration?

Question 65hardmultiple choice

Read the full network assurance explanation →

An administrator is responsible for a large NVIDIA DGX SuperPOD used for multi-node training. They need to ensure that GPU telemetry and health metrics are collected centrally and can trigger alerts when GPUs exceed temperature thresholds. Which component of NVIDIA Base Command Manager (BCM) should they configure to achieve this?

Question 66mediummultiple choice

Read the full Administration explanation →

An AI operations team is deploying NVIDIA Base Command Manager to manage a cluster of DGX nodes. They want to ensure that only authorized users can submit jobs and that all job submissions are audited. Which combination of Base Command Manager features should the administrator configure to meet these requirements?

Question 67mediummultiple choice

Read the full Administration explanation →

A company runs multiple AI workloads on a shared Kubernetes cluster with NVIDIA GPUs. The administrator needs to enforce that only pods with a specific label can consume GPU resources, while other pods are denied. Which Kubernetes admission control mechanism should be used to implement this policy?

Question 68hardmultiple choice

Read the full Administration explanation →

A site reliability engineer is troubleshooting a DGX A100 node that intermittently drops out of the cluster during large NCCL all-reduce jobs. `nvidia-smi` shows all eight A100 GPUs healthy, but DCGM reports XID errors 74 and 79 on one GPU during the failures. The engineer needs to determine the most likely cause and the correct administrative action. Which combination best describes the cause and the appropriate first step?

Question 69easymultiple choice

Read the full Administration explanation →

An administrator is responsible for maintaining a fleet of NVIDIA-certified servers running AI workloads. They need to quickly identify which servers have GPUs that are overheating and may throttle performance. Which NVIDIA tool should the administrator use to monitor GPU temperature across the fleet in real time?

Question 70mediummultiple choice

Read the full Administration explanation →

An administrator is managing an NVIDIA DGX SuperPOD used for large-scale AI training. The cluster uses a Slurm workload manager. The administrator needs to ensure that jobs are scheduled only on nodes with healthy GPUs and that failed GPUs are automatically drained from the pool. Which integration should be configured to achieve this?

Question 71mediummultiple choice

Read the full Administration explanation →

An administrator is using NVIDIA Base Command Manager to provision a new GPU cluster. They need to ensure that the compute nodes are configured with the correct GPU driver and CUDA toolkit versions. Which Base Command Manager feature should they use to automate this?

Question 72hardmultiple choice

Read the full Administration explanation →

An administrator is responsible for an NVIDIA AI Enterprise deployment on Kubernetes. The security team requires that all GPU-accelerated pods run with the least privilege necessary and that GPU device nodes are not exposed to pods that do not request them. Which combination of configurations should the administrator implement to meet these requirements?

Question 73easymultiple choice

Read the full network assurance explanation →

An AI operations team needs to monitor GPU health and utilization across a fleet of DGX nodes from a single dashboard. They want per-GPU metrics such as power, temperature, utilization, and ECC errors, and they want to retain historical data for capacity planning. Which NVIDIA tool is purpose-built to collect and expose these GPU telemetry metrics for centralized monitoring?

Question 74hardmulti select

Read the full Administration explanation →

An administrator is configuring NVIDIA GPUDirect Storage (GDS) on a cluster to accelerate data loading for AI training jobs. The cluster uses Mellanox InfiniBand adapters and NVMe storage. Which two actions are required to enable GDS and ensure optimal performance? (Choose two.)

Question 75mediummultiple choice

Read the full Administration explanation →

An administrator is managing a Kubernetes cluster with NVIDIA GPU Operator. A new policy requires that all GPU workloads run with specific environment variables set, such as NVIDIA_VISIBLE_DEVICES and NVIDIA_DRIVER_CAPABILITIES. The administrator wants to enforce these variables automatically for any pod that requests a GPU, without modifying each pod specification manually. Which approach should the administrator use?

Question 76hardmultiple choice

Read the full Administration explanation →

An administrator is troubleshooting a performance degradation in a multi-node NVIDIA NCCL-based training job. The job spans four DGX nodes connected via InfiniBand. The administrator suspects that NCCL is not using the optimal network path. Which action should the administrator take to verify and enforce the use of GPUDirect RDMA for inter-node communication?

Question 77hardmulti select

Read the full Administration explanation →

An administrator is configuring an NVIDIA AI Enterprise cluster to run multi-tenant inference workloads on Kubernetes. The administrator must ensure that GPU resources are isolated and that tenants cannot access each other's GPU memory. Which two actions should the administrator take? (Choose two.)