"Error: PF Exception 14 in world ####:vmsyslogd IP #### addr #### due to Faulty Processor"

Products

• VMware vSphere ESXi 8.0

Issue/Introduction

• An ESXi host crashes with a Purple Screen of Death (PSOD) displaying a Page Fault Exception 14. The crash signature matches the following redacted backtrace:

#PF Exception 14 in world ####:vmsyslogd IP #### addr ####

PTEs:0x814aa53027:0x101a6c1f027:0x816a62c027:0x800cf85ed984d140:

Module(s) involved in panic: [vmkernel Version Releasebuild-24674464]

cr0=0x80010031 cr2=0xa0c52a21c0 cr3=0x1019f241000 cr4=0x142768

FMS=06/8f/8 uCode=0x2b000639

framc=0x453adc09bc30 ip=0x420030bc482f crr=0x0 rflags=0x10206

rax=0xa0e52a0700 rbx=0x431a36602010 rcx=0x10

rdx=0x2003d0 rbp=0x453ade09bf40 rsi=0x453ade09be58

rdi=0x420030cdec61 r8=0x4314e94017f0 r9=0x4314e94066b0

r10=0x4314e9400e58 r11=0x2 r12=0x453ade09f000

r13=0x431a36617080 r14=0x431a36617080 r15=0x420030bd326

*PCPU29:2098128/vmsys logd

         Code start: 0x420030600000 VMK uptimc: 2:09:23:45.539

0x453ade09bef0:[0x420030be482f ]UserDump_InDumperCheck@vmkernellnover+0xf stack: 0x453ade09bf 40

0x453ade09bf10:[0x420030bd3373]User_LinuxSyscallHandler@vnkernel#nover+0x110 stack: 0x3

0x453ade09bf40: [0x420030ca10c6]gate_entry@vnkernel#nover+0xa7 stack: 0x0

 

• Backtrace:

UserDump_InDumperCheck () at bora/vmkernel/user/userDump.c:3630
0x00004200313d3c44 in User_UWVMK64SyscallHandler () at bora/vmkernel/user/user.c:2248
0x000042003149d446 in Syscall_InstructionHandler () at bora/vmkernel/main/x86/common64.S:324
0x0000000000000103 in ?? ()
0x0000000000000000 in ?? ()

Environment

VMware ESXi 8.x

Cause

The PSOD occurs because the thread obtains an invalid world address. The address should be loaded from the PRDA corresponding PCPU. However, the value in the register is incorrect despite the valid value residing in memory. This directly indicates a hardware-level fault within the physical processor.

Resolution

To resolve this issue, perform the following steps:

1. Analyze /var/run/log/vmkernel.log or /var/run/log/LogEFI.log to verify if the system crashes consistently map to a specific Physical CPU (PCPU).

2. Collect the ESXi host diagnostic information bundle. This can be generated using the vSphere Client Collecting diagnostic information for VMware ESXi using vSphere Client  or by executing the vm-support command directly on the ESXi host via SSH Collecting diagnostic information for VMware ESX/ESXi using vm-support command

3. Engage the hardware vendor to replace the physical processor package containing the identified PCPU.

Additional Information

• Collecting diagnostic information for VMware ESX/ESXi using vSphere Client

• Collecting diagnostic information for VMware ESX/ESXi using vm-support command