McAfee-Secured Website

Exam Code: NCP-AII

Exam Name: NCP-AI Infrastructure

Certification Provider: NVIDIA

Questions & Answers

100% NVIDIA NCP-AII Guaranteed Success With Testking Training Materials

Get Real NCP-AII Exam Questions & Expert Verified Answers as Experienced on the Real Exam!

Questions & Answers

170 Questions & Answers with Testing Engine

"NCP-AI Infrastructure Exam", also known as NCP-AII exam, is a NVIDIA certification exam.

Pass your tests with the always up-to-date NCP-AII Exam Engine. Your NCP-AII training materials keep you at the head of the pack!

Satisfaction Guaranteed

Money Back Guarantee

Testking provides hassle-free money back guarantee with our products. That is because we have 100% trust in the abilities of our professional and experience product team, and our record is a proof of that.

99.6% PASS RATE
Was: $137.49
Now: $124.99

NVIDIA NCP-AII Practice Test 5

Question 40

During a GPU datacenter maintenance window, an engineer discovers: A rack has a functioning GPU data fabric but its remote operators cannot reach the BMC after a workload network outage. Which repeatable operational safeguard best prevents recurrence of this failure?

  1. Require hardened BMC initialization with credential rotation and network isolation in the rack acceptance checklist.
  2. Collect reproducible TPM and boot attestation evidence for every approved hardware and firmware update.
  3. Validate part compatibility and power fault resilience during serialized hardware inventory acceptance.
  4. Map and periodically exercise physical management transport and power dependencies.

Explanation

Correct Answer

D:

This preventive practice addresses the underlying failure (An out-of-band label is meaningless if management traffic depends on the same path as the failed data plane.) and supports continued assurance through: Disconnecting both host data uplinks still permits an authorized administrator to inspect hardware alerts through the OOB management network.

Incorrect Answers

A:

This prevention control targets another operational risk: A newly received DGX-class server has default BMC credentials and its management interface is reachable from the general campus network. The currently reported mechanism remains: An out-of-band label is meaningless if management traffic depends on the same path as the failed data plane.

B:

This prevention control targets another operational risk: An AI appliance is deployed with measured boot enabled but platform ownership and TPM-backed trust measurements were never recorded. The currently reported mechanism remains: An out-of-band label is meaningless if management traffic depends on the same path as the failed data plane.

C:

This prevention control targets another operational risk: A recently delivered accelerator system has inconsistent PSU models across sibling nodes, and unexpected power alarms occur only during high utilization. The currently reported mechanism remains: An out-of-band label is meaningless if management traffic depends on the same path as the failed data plane.

Question 41

A GPU cluster commissioning engineer investigates this situation: An AI appliance is deployed with measured boot enabled but platform ownership and TPM-backed trust measurements were never recorded. Which corrective deployment or operational action best addresses the stated problem?

  1. Configure unique BMC credentials, isolated out-of-band addressing, role-limited access, and an audited first-login process.
  2. Enroll and verify the TPM trust state and approved secure-boot measurement baseline before releasing the host into the training cluster.
  3. Replace shared privileged access with named role-scoped identities and centralized authentication and audit logging where supported.
  4. Require a documented rack-to-power-to-network dependency map and signed preflight before first energization.

Explanation

Correct Answer

B:

For the stated NVIDIA AI infrastructure condition (An AI appliance is deployed with measured boot enabled but platform ownership and TPM-backed trust measurements were never recorded.), this action addresses the concrete failure: Enroll and verify the TPM trust state and approved secure-boot measurement baseline before releasing the host into the training cluster. The underlying mechanism is: Enabling security features does not prove platform identity or integrity without an expected reference and verification process. Verification would require: The attestation log matches the expected firmware and boot chain, while an intentionally altered baseline is rejected or escalated.

Incorrect Answers

A:

This option could be appropriate for a different NVIDIA infrastructure problem (1.3: Perform initial configuration of BMC, out-of-band management, and TPM): A newly received DGX-class server has default BMC credentials and its management interface is reachable from the general campus network. The present issue is more specifically caused by: Enabling security features does not prove platform identity or integrity without an expected reference and verification process.

C:

This option could be appropriate for a different NVIDIA infrastructure problem (1.3: Perform initial configuration of BMC, out-of-band management, and TPM): A contractor uses an all-powerful shared BMC account during a GPU rack installation, and later no one can determine who performed a remote power cycle. The present issue is more specifically caused by: Enabling security features does not prove platform identity or integrity without an expected reference and verification process.

D:

This option could be appropriate for a different NVIDIA infrastructure problem (1.1: Describe sequence of events for deployment and validation): A new GPU pod is powered on before rack inventory and PDU mapping are approved, and engineers cannot identify which chassis will lose power during maintenance. The present issue is more specifically caused by: Enabling security features does not prove platform identity or integrity without an expected reference and verification process.

Question 42

A GPU cluster commissioning engineer investigates this situation: An AI appliance is deployed with measured boot enabled but platform ownership and TPM-backed trust measurements were never recorded. Which measured result most directly proves that this specific problem is resolved?

  1. A test technician can execute only permitted BMC tasks and each administrative action is attributable to a distinct authorized principal.
  2. The attestation log matches the expected firmware and boot chain, while an intentionally altered baseline is rejected or escalated.
  3. A sampled controller query returns the correct serial number, IP address, rack position, and recent alarm history consistently.
  4. Measured iteration time stays within the training objective while asynchronous replication satisfies inter-region data movement requirements.

Explanation

Correct Answer

B:

This is the measurable acceptance result for the described case: The attestation log matches the expected firmware and boot chain, while an intentionally altered baseline is rejected or escalated. It validates the corrective operation (Enroll and verify the TPM trust state and approved secure-boot measurement baseline before releasing the host into the training cluster.) rather than a related but different system check.

Incorrect Answers

A:

This observation would help verify a different incident: A contractor uses an all-powerful shared BMC account during a GPU rack installation, and later no one can determine who performed a remote power cycle. It does not directly establish the required result for the current case: The attestation log matches the expected firmware and boot chain, while an intentionally altered baseline is rejected or escalated.

C:

This observation would help verify a different incident: Two BMC controllers ship with duplicate static management addresses and the cluster monitoring system alternately attributes temperature alarms to the wrong chassis. It does not directly establish the required result for the current case: The attestation log matches the expected firmware and boot chain, while an intentionally altered baseline is rejected or escalated.

D:

This observation would help verify a different incident: A campus-to-data-center AI service relies on long-distance model synchronization, but the proposed network topology ignores the latency effects of cross-region collective operations. It does not directly establish the required result for the current case: The attestation log matches the expected firmware and boot chain, while an intentionally altered baseline is rejected or escalated.

Question 43

A GPU cluster commissioning engineer investigates this situation: An AI appliance is deployed with measured boot enabled but platform ownership and TPM-backed trust measurements were never recorded. Which root cause or infrastructure principle best explains the observed failure?

  1. Enabling security features does not prove platform identity or integrity without an expected reference and verification process.
  2. Default management identities and publicly reachable BMC interfaces expand the compromise surface before operating systems are deployed.
  3. Unattributed shared management access undermines accountability over sensitive physical server functions.
  4. AI server deployment depends on validating physical infrastructure dependencies before introducing operating system or cluster workloads.

Explanation

Correct Answer

A:

This technical mechanism explains the reported condition: Enabling security features does not prove platform identity or integrity without an expected reference and verification process. The observed situation is: An AI appliance is deployed with measured boot enabled but platform ownership and TPM-backed trust measurements were never recorded. A requirement-aligned correction is: Enroll and verify the TPM trust state and approved secure-boot measurement baseline before releasing the host into the training cluster.

Incorrect Answers

B:

This underlying cause is associated with another GPU platform incident: A newly received DGX-class server has default BMC credentials and its management interface is reachable from the general campus network. The observed case instead concerns: Enabling security features does not prove platform identity or integrity without an expected reference and verification process.

C:

This underlying cause is associated with another GPU platform incident: A contractor uses an all-powerful shared BMC account during a GPU rack installation, and later no one can determine who performed a remote power cycle. The observed case instead concerns: Enabling security features does not prove platform identity or integrity without an expected reference and verification process.

D:

This underlying cause is associated with another GPU platform incident: A new GPU pod is powered on before rack inventory and PDU mapping are approved, and engineers cannot identify which chassis will lose power during maintenance. The observed case instead concerns: Enabling security features does not prove platform identity or integrity without an expected reference and verification process.

Question 44

A GPU cluster commissioning engineer investigates this situation: An AI appliance is deployed with measured boot enabled but platform ownership and TPM-backed trust measurements were never recorded. Which repeatable operational safeguard best prevents recurrence of this failure?

  1. Validate BMC addressing and device identity as an automated part of first-stage provisioning.
  2. Map and periodically exercise physical management transport and power dependencies.
  3. Collect reproducible TPM and boot attestation evidence for every approved hardware and firmware update.
  4. Define explicit pass/fail gates from physical installation through cluster-level application verification.

Explanation

Correct Answer

C:

This preventive practice addresses the underlying failure (Enabling security features does not prove platform identity or integrity without an expected reference and verification process.) and supports continued assurance through: The attestation log matches the expected firmware and boot chain, while an intentionally altered baseline is rejected or escalated.

Incorrect Answers

A:

This prevention control targets another operational risk: Two BMC controllers ship with duplicate static management addresses and the cluster monitoring system alternately attributes temperature alarms to the wrong chassis. The currently reported mechanism remains: Enabling security features does not prove platform identity or integrity without an expected reference and verification process.

B:

This prevention control targets another operational risk: A rack has a functioning GPU data fabric but its remote operators cannot reach the BMC after a workload network outage. The currently reported mechanism remains: Enabling security features does not prove platform identity or integrity without an expected reference and verification process.

D:

This prevention control targets another operational risk: A data center declares a 32-node training cluster ready because every power LED is green, although no node-to-node connectivity or GPU enumeration was tested. The currently reported mechanism remains: Enabling security features does not prove platform identity or integrity without an expected reference and verification process.

Question 45

During NVIDIA AI factory acceptance, a technician observes: Two BMC controllers ship with duplicate static management addresses and the cluster monitoring system alternately attributes temperature alarms to the wrong chassis. Which corrective deployment or operational action best addresses the stated problem?

  1. Replace shared privileged access with named role-scoped identities and centralized authentication and audit logging where supported.
  2. Provide genuinely independent out-of-band transport and test remote console and power operations without the production fabric.
  3. Validate multi-node GPU communication with NCCL under the intended fabric and test peer reachability and interface selection.
  4. Assign authoritative unique OOB addresses from inventory and bind each controller identity to the physical asset serial number.

Explanation

Correct Answer

D:

For the stated NVIDIA AI infrastructure condition (Two BMC controllers ship with duplicate static management addresses and the cluster monitoring system alternately attributes temperature alarms to the wrong chassis.), this action addresses the concrete failure: Assign authoritative unique OOB addresses from inventory and bind each controller identity to the physical asset serial number. The underlying mechanism is: Address collisions in the management network make remote recovery and diagnostic attribution unreliable. Verification would require: A sampled controller query returns the correct serial number, IP address, rack position, and recent alarm history consistently.

Incorrect Answers

A:

This option could be appropriate for a different NVIDIA infrastructure problem (1.3: Perform initial configuration of BMC, out-of-band management, and TPM): A contractor uses an all-powerful shared BMC account during a GPU rack installation, and later no one can determine who performed a remote power cycle. The present issue is more specifically caused by: Address collisions in the management network make remote recovery and diagnostic attribution unreliable.

B:

This option could be appropriate for a different NVIDIA infrastructure problem (1.3: Perform initial configuration of BMC, out-of-band management, and TPM): A rack has a functioning GPU data fabric but its remote operators cannot reach the BMC after a workload network outage. The present issue is more specifically caused by: Address collisions in the management network make remote recovery and diagnostic attribution unreliable.

C:

This option could be appropriate for a different NVIDIA infrastructure problem (1.10: Validate hardware operation for workloads): A single-node CUDA test passes, yet a distributed job hangs because the NVIDIA collective communication transport cannot communicate across hosts. The present issue is more specifically caused by: Address collisions in the management network make remote recovery and diagnostic attribution unreliable.

Question 46

During NVIDIA AI factory acceptance, a technician observes: Two BMC controllers ship with duplicate static management addresses and the cluster monitoring system alternately attributes temperature alarms to the wrong chassis. Which measured result most directly proves that this specific problem is resolved?

  1. Disconnecting both host data uplinks still permits an authorized administrator to inspect hardware alerts through the OOB management network.
  2. The attestation log matches the expected firmware and boot chain, while an intentionally altered baseline is rejected or escalated.
  3. A sampled controller query returns the correct serial number, IP address, rack position, and recent alarm history consistently.
  4. A reproducible commissioning test clears hardware alarms before driver installation and shows every accepted node matches the release baseline.

Explanation

Correct Answer

C:

This is the measurable acceptance result for the described case: A sampled controller query returns the correct serial number, IP address, rack position, and recent alarm history consistently. It validates the corrective operation (Assign authoritative unique OOB addresses from inventory and bind each controller identity to the physical asset serial number.) rather than a related but different system check.

Incorrect Answers

A:

This observation would help verify a different incident: A rack has a functioning GPU data fabric but its remote operators cannot reach the BMC after a workload network outage. It does not directly establish the required result for the current case: A sampled controller query returns the correct serial number, IP address, rack position, and recent alarm history consistently.

B:

This observation would help verify a different incident: An AI appliance is deployed with measured boot enabled but platform ownership and TPM-backed trust measurements were never recorded. It does not directly establish the required result for the current case: A sampled controller query returns the correct serial number, IP address, rack position, and recent alarm history consistently.

D:

This observation would help verify a different incident: A hosting team installs accelerator drivers before confirming BIOS and BMC firmware levels, and several nodes fail post-installation hardware checks. It does not directly establish the required result for the current case: A sampled controller query returns the correct serial number, IP address, rack position, and recent alarm history consistently.

Question 47

During NVIDIA AI factory acceptance, a technician observes: Two BMC controllers ship with duplicate static management addresses and the cluster monitoring system alternately attributes temperature alarms to the wrong chassis. Which root cause or infrastructure principle best explains the observed failure?

  1. Unattributed shared management access undermines accountability over sensitive physical server functions.
  2. Address collisions in the management network make remote recovery and diagnostic attribution unreliable.
  3. Enabling security features does not prove platform identity or integrity without an expected reference and verification process.
  4. A physical link-light state does not prove full-width and error-free transport capacity.

Explanation

Correct Answer

B:

This technical mechanism explains the reported condition: Address collisions in the management network make remote recovery and diagnostic attribution unreliable. The observed situation is: Two BMC controllers ship with duplicate static management addresses and the cluster monitoring system alternately attributes temperature alarms to the wrong chassis. A requirement-aligned correction is: Assign authoritative unique OOB addresses from inventory and bind each controller identity to the physical asset serial number.

Incorrect Answers

A:

This underlying cause is associated with another GPU platform incident: A contractor uses an all-powerful shared BMC account during a GPU rack installation, and later no one can determine who performed a remote power cycle. The observed case instead concerns: Address collisions in the management network make remote recovery and diagnostic attribution unreliable.

C:

This underlying cause is associated with another GPU platform incident: An AI appliance is deployed with measured boot enabled but platform ownership and TPM-backed trust measurements were never recorded. The observed case instead concerns: Address collisions in the management network make remote recovery and diagnostic attribution unreliable.

D:

This underlying cause is associated with another GPU platform incident: A technician sees a healthy network interface LED on a ConnectX adapter but a cluster workload receives only a fraction of the planned throughput. The observed case instead concerns: Address collisions in the management network make remote recovery and diagnostic attribution unreliable.

Question 48

During NVIDIA AI factory acceptance, a technician observes: Two BMC controllers ship with duplicate static management addresses and the cluster monitoring system alternately attributes temperature alarms to the wrong chassis. Which repeatable operational safeguard best prevents recurrence of this failure?

  1. Validate BMC addressing and device identity as an automated part of first-stage provisioning.
  2. Collect reproducible TPM and boot attestation evidence for every approved hardware and firmware update.
  3. Require hardened BMC initialization with credential rotation and network isolation in the rack acceptance checklist.
  4. Gate each deployment phase on recorded hardware, firmware, networking, and host software acceptance evidence.

Explanation

Correct Answer

A:

This preventive practice addresses the underlying failure (Address collisions in the management network make remote recovery and diagnostic attribution unreliable.) and supports continued assurance through: A sampled controller query returns the correct serial number, IP address, rack position, and recent alarm history consistently.

Incorrect Answers

B:

This prevention control targets another operational risk: An AI appliance is deployed with measured boot enabled but platform ownership and TPM-backed trust measurements were never recorded. The currently reported mechanism remains: Address collisions in the management network make remote recovery and diagnostic attribution unreliable.

C:

This prevention control targets another operational risk: A newly received DGX-class server has default BMC credentials and its management interface is reachable from the general campus network. The currently reported mechanism remains: Address collisions in the management network make remote recovery and diagnostic attribution unreliable.

D:

This prevention control targets another operational risk: A hosting team installs accelerator drivers before confirming BIOS and BMC firmware levels, and several nodes fail post-installation hardware checks. The currently reported mechanism remains: Address collisions in the management network make remote recovery and diagnostic attribution unreliable.

Question 49

An infrastructure administrator receives the following incident: A contractor uses an all-powerful shared BMC account during a GPU rack installation, and later no one can determine who performed a remote power cycle. Which corrective deployment or operational action best addresses the stated problem?

  1. Configure unique BMC credentials, isolated out-of-band addressing, role-limited access, and an audited first-login process.
  2. Replace shared privileged access with named role-scoped identities and centralized authentication and audit logging where supported.
  3. Assign authoritative unique OOB addresses from inventory and bind each controller identity to the physical asset serial number.
  4. Compare physical slot and cable seating with PCIe enumeration and NVIDIA SMI device identifiers before running training.

Explanation

Correct Answer

B:

For the stated NVIDIA AI infrastructure condition (A contractor uses an all-powerful shared BMC account during a GPU rack installation, and later no one can determine who performed a remote power cycle.), this action addresses the concrete failure: Replace shared privileged access with named role-scoped identities and centralized authentication and audit logging where supported. The underlying mechanism is: Unattributed shared management access undermines accountability over sensitive physical server functions. Verification would require: A test technician can execute only permitted BMC tasks and each administrative action is attributable to a distinct authorized principal.

Incorrect Answers

A:

This option could be appropriate for a different NVIDIA infrastructure problem (1.3: Perform initial configuration of BMC, out-of-band management, and TPM): A newly received DGX-class server has default BMC credentials and its management interface is reachable from the general campus network. The present issue is more specifically caused by: Unattributed shared management access undermines accountability over sensitive physical server functions.

C:

This option could be appropriate for a different NVIDIA infrastructure problem (1.3: Perform initial configuration of BMC, out-of-band management, and TPM): Two BMC controllers ship with duplicate static management addresses and the cluster monitoring system alternately attributes temperature alarms to the wrong chassis. The present issue is more specifically caused by: Unattributed shared management access undermines accountability over sensitive physical server functions.

D:

This option could be appropriate for a different NVIDIA infrastructure problem (1.6: Install GPU-based servers and use NVIDIA SMI): A server inventory says eight GPUs should be installed, but one accelerator intermittently disappears only after the chassis is moved between racks. The present issue is more specifically caused by: Unattributed shared management access undermines accountability over sensitive physical server functions.

Question 50

An infrastructure administrator receives the following incident: A contractor uses an all-powerful shared BMC account during a GPU rack installation, and later no one can determine who performed a remote power cycle. Which measured result most directly proves that this specific problem is resolved?

  1. The attestation log matches the expected firmware and boot chain, while an intentionally altered baseline is rejected or escalated.
  2. A test technician can execute only permitted BMC tasks and each administrative action is attributable to a distinct authorized principal.
  3. Disconnecting both host data uplinks still permits an authorized administrator to inspect hardware alerts through the OOB management network.
  4. Restoring a validated compatibility setting returns expected PCIe link width without replacing a healthy accelerator.

Explanation

Correct Answer

B:

This is the measurable acceptance result for the described case: A test technician can execute only permitted BMC tasks and each administrative action is attributable to a distinct authorized principal. It validates the corrective operation (Replace shared privileged access with named role-scoped identities and centralized authentication and audit logging where supported.) rather than a related but different system check.

Incorrect Answers

A:

This observation would help verify a different incident: An AI appliance is deployed with measured boot enabled but platform ownership and TPM-backed trust measurements were never recorded. It does not directly establish the required result for the current case: A test technician can execute only permitted BMC tasks and each administrative action is attributable to a distinct authorized principal.

C:

This observation would help verify a different incident: A rack has a functioning GPU data fabric but its remote operators cannot reach the BMC after a workload network outage. It does not directly establish the required result for the current case: A test technician can execute only permitted BMC tasks and each administrative action is attributable to a distinct authorized principal.

D:

This observation would help verify a different incident: A GPU server presents PCIe link-width degradation immediately after a platform firmware change, but technicians plan to replace the accelerator first. It does not directly establish the required result for the current case: A test technician can execute only permitted BMC tasks and each administrative action is attributable to a distinct authorized principal.

NCP-AI Infrastructure (NCP-AII) Practice Tests