ESXi host fails with a purple screen of death "PF#14 in world vsi_traverse during CpuMetricsLoadHistorySnapshotStats"

Products

• VMware vSphere ESXi

Issue/Introduction

ESXi host experienced PSOD

cpu1:123503206)FMS=06/cf/2 uCode=0x21000291
LogEFI: cpu1:123503206)r13=0xdc r14=0x420010cdb9bc r15=0x0
LogEFI[2098616]: *PCPU1:123503206/esxcfg-info
LogEFI[2098616]: PCPU  0: SUVUUSVVVUVVVSUVSVVVVVVSVVVVUVVUUUVSVVVVVVVVUVVVVSUVVUVVSVSVUVVS
cpu1:123503206)0x4539fff9b590:[0x420010a5bd95]CpuMetricsLoadHistoryDemandGet@vmkernel
cpu1:123503206)0x4539fff9b598:[0x420010a5d8a7]CpuMetricsSR_VmCpuCapacityStatsGet
cpu1:123503206)0x4539fff9b5a0:[0x420010a95edb]Sched_GetGroupCpuStats@vmkernel
cpu1:123503206)0x4539fff9b5d0:[0x420010cdb9e9]SchedVsi_GroupVsiGetCPUStats
cpu1:123503206)0x4539fff9b5f0:[0x420010606139]VSI_GetInfo@vmkernel
cpu1:123503206)0x4539fff9b660:[0x420010c353b8]UW64VMKSyscallUnpackVSI_Get@vmkernel
cpu1:123503206)0x4539fff9bef0:[0x420010bd06bb]User_UWVMK64SyscallHandler@vmkernel
cpu1:123503206)0x4539fff9bf40:[0x420010c98105]SyscallUWVMK64@vmkernel
cpu1:123503206)base fs=0x0 gs=0x420040400000 Kgs=0x0
LogEFI: cpu3:2097740)vmsyslog logger ###.##.##:### lost 369899 log messages
LogEFI: cpu53:2097740)vmsyslog logger  ###.##.##:### lost 1 log messages
LogEFI: cpu58:2097740)vmsyslog logger  ###.##.##:### lost 1 log messages
LogEFI: cpu34:2097740)vmsyslog logger  ###.##.##:### lost 688836 log messages
LogEFI: cpu34:2097740)vmsyslog logger  ###.##.##:### lost 680598 log messages
LogEFI: cpu7:2097740)vmsyslog logger  ###.##.##:### lost 1 log messages
LogEFI: cpu28:2097740)vmsyslog logger  ###.##.##:### lost 1 log messages
LogEFI: cpu0:2097740)vmsyslog logger  ###.##.##:### lost 572130 log messages
LogEFI: cpu0:2097740)vmsyslog logger  ###.##.##:### lost 573743 log messages

Environment

VMware ESXi 8.0.3

Cause

This issue occurs due to a race condition during the collection of CPU usage statistics. 

 

Resolution

This issue is fixed in ESXi 8.0U3h or later. 

Additional Information

VMware ESXi 8.0 Update 3h Release Notes