NCP-AII Question 16
Select 4You are managing an NVIDIA DGX system in a data center running critical AI workloads. One of the GPUs has stopped responding, and the logs indicate persistent hardware errors. What actions should you take to replace the faulty GPU and restore system functionality?
- A
Power off the system and disconnect it from the power source before replacing the GPU.
- B
Run a hardware diagnostic tool to confirm the GPU failure before proceeding with replacement.
- C
Replace the faulty GPU with a spare GPU while the system is powered on to minimize downtime.
- D
Verify that the replacement GPU is compatible with the system and has the correct firmware version.
- E
Document the replacement process and update the inventory records after completing the task.
Show answer and explanation
Correct answers: A, B, D, E
Explanation
Replacing a faulty GPU in an NVIDIA DGX system requires careful planning and adherence to best practices for hardware safety and system compatibility. Powering off the system, confirming the failure, verifying compatibility, and documenting the process are essential steps to ensure a successful replacement while maintaining system reliability and operational integrity.
- A. Correct.
Powering off the system and disconnecting it from the power source is a critical safety step to avoid electrical hazards and prevent further hardware damage during the replacement process.
- B. Correct.
Running a hardware diagnostic tool helps confirm the root cause of the problem and ensures that the GPU is indeed faulty before performing a replacement.
- C. Incorrect.
Replacing the GPU while the system is powered on is not recommended because it can cause electrical damage and void warranty or support agreements.
- D. Correct.
Verifying compatibility and firmware version ensures the replacement GPU will function correctly within the system and avoid compatibility issues with other components or software.
- E. Correct.
Documenting the process and updating inventory records is essential for accurate tracking of hardware changes and maintaining system documentation.