Question 930
A server operations team encounters the following situation: A blade's host crashes when its GPU workload starts, but temperature logs reveal rapid GPU overheating and an obstructed air duct. After making the corrective change, which result best demonstrates that this particular problem is resolved?
Explanation
Correct Answer
This is incident-specific acceptance evidence: Thermal telemetry remains within limits while the application executes. It verifies the action restore the approved airflow path and retest the GPU cooling system. The failure under test was: The workload exposes insufficient GPU cooling rather than an application-only defect.
Incorrect Answers
This observation tests an unrelated Given a scenario, troubleshoot common hardware failures case: A new CPU cooler has a loose mounting bracket, and server logs show intermittent overheating and unexplained kernel crashes. It may be true while the current failure persists; the decisive current cause is: The workload exposes insufficient GPU cooling rather than an application-only defect.
This observation tests an unrelated Given a scenario, troubleshoot common hardware failures case: A compute server produces repeated uncorrectable ECC memory errors followed by kernel panics, while its processor temperatures remain normal. It may be true while the current failure persists; the decisive current cause is: The workload exposes insufficient GPU cooling rather than an application-only defect.
This observation tests an unrelated Given a scenario, troubleshoot common hardware failures case: A maintenance technician blows dust deeper into server fan bearings with uncontrolled shop compressed air, after which fans vibrate and overheat. It may be true while the current failure persists; the decisive current cause is: The workload exposes insufficient GPU cooling rather than an application-only defect.