How should you gracefully shut down a hung server?

• A. Use out-of-band BMC power cycle after capturing whatever evidence is possible.
• B. Force power off via the power switch without BMC.
• C. Reboot the OS and monitor for recovery.
• D. Unplug the power supply to force shutdown immediately.

Answer: (A)

When a server is hung, the priority is to recover safely while preserving evidence. Use the out-of-band management controller to perform a controlled power cycle after you’ve captured what you can observe from the system. The BMC lets you initiate an orderly shutdown or a clean power cycle without relying on the unresponsive OS, which reduces the risk of data loss compared to pulling power or pressing the physical switch without remote control. Before cycling power, collect logs and evidence you can access: console output, last messages from the kernel, syslog or event logs, hardware health status, and any active alerts. Then let the BMC handle the power cycle. This approach provides a salvageable restart path and keeps a verifiable record of what happened.

Unplugging the power, forcing a switch-only reset, or rebooting the OS from within a hung system can lead to data corruption and loss of evidence, so they’re less appropriate in this scenario.