ESXi PSOD in world "CAM taskq" when a Virtual CD-ROM is attached via older BMC versions (4.40 and below)

Products

• VMware vSphere ESXi

Issue/Introduction

An ESXi host running an outdated Baseboard Management Controller (BMC) firmware version experiences a Purple Screen of Death (PSOD). The crash signature and backtrace display the following error:

YYYY-MM-DDTHH:MM:SS cpu1:2098002)@BlueScreen: #PF Exception 14 in world 2098002:CAM taskq IP 0x0 addr 0x0

Environment

ESXi Servers running older BMC firmware versions. This issue has been specifically observed on BMC versions 3.68 and 4.40, while version 6.10 and later appear unaffected.

Cause

The PSOD is caused by a race condition within the taskqueue_run_locked thread. This condition is triggered when the BMC's Virtual CD-ROM device rapidly and repeatedly attaches and detaches over a short timeframe. Consequently, the thread attempts to process a corrupted task, resulting in a critical host failure.

Note: The current BMC version can be verified by executing the following command directly on the ESXi host:
esxcfg-info -a | grep "BMC Version"

Resolution

Contact hardware vendor to obtain and apply the latest BMC firmware update for server hardware.