ESXi host multiple PSODs frequently with PF Execption 14

Products

• VMware vSphere ESXi

Issue/Introduction

Multiple PSODs are due to different scenarios on Physical CPU Cores.

Example:

PCPU 216 accessed 0x0 as the instruction address, the thread jumped to such an invalid address when calling Timer_Interrupt handler. 

/var/run/log/vmkernel.log

YYYY-MM-DDTHH:mm:SS.961Z cpu216:2097368)@BlueScreen: #PF Exception 14 in world 2097368:idle216 IP 0x0 addr 0x0

The thread on PCPU 217 received #PF when running 'sub' instruction, that shouldn't have accessed memory.

/var/run/log/vmkernel.log

YYYY-MM-DDTHH:mm:SS.863Z cpu217:2097369)Panic: 630: Panic from another CPU (cpu 217, world 2097369): ip=0x4200146a6de0 randomOff=0x14000000:
#PF Exception 14 in world 2097369:idle217 IP 0x4200146cf93c addr 0x0
PTEs:0x507e21c023;0x507e21d023;0x307ebfb023;0x0;

PCPU 216 and 217 are the physical processor core.

Environment

ESXi 8.x

Cause

It was caused by misbehaving processor or bad CPU Core due to which multiple PSODs caused by different scenarios on the CPU.

 

/var/run/log/LogEFI.log

YYYY-MM-DDTHH:mm:SS.957Z In(14) LogEFI[2100143]: PTEs:0x0;Halting PCPU 216.2025-08-30T05:37:35.592Z cpu217:2105157)VMware ESXi 8.0.3 [Releasebuild-24022510 x86_64]
YYYY-MM-DDTHH:mm:SS.613Z In(14) LogEFI: cpu217:2100663)VMware ESXi 8.0.3 [Releasebuild-24022510 x86_64]
YYYY-MM-DDTHH:mm:SS.079Z In(14) LogEFI: cpu217:2097369)VMware ESXi 8.0.3 [Releasebuild-24022510 x86_64]
YYYY-MM-DDTHH:mm:SS.415Z In(14) LogEFI: cpu217:2099549)VMware ESXi 8.0.3 [Releasebuild-24022510 x86_64]
YYYY-MM-DDTHH:mm:SS.356Z In(14) LogEFI: cpu216:2116355)VMware ESXi 8.0.3 [Releasebuild-24022510 x86_64]
YYYY-MM-DDTHH:mm:SS.514Z In(14) LogEFI[2100068]: PTEs:0x507e21c023;0x507e21d023;0x307ebfb023;0x0;Halting PCPU 217.YYYY-MM-DDTHH:mm:SS.955Z cpu216:2097368)VMware ESXi 8.0.3 [Releasebuild-24022510 x86_64]

Resolution

Please engage with hardware Vendor for further assistance such as updating BIOS, processor replacement, etc.