vSAN host reported PSOD while rebooting

Products

• VMware vSAN

Issue/Introduction

• vSAN ESA host reported PSOD when reboot was initiated.
  
  
• Below is the PSOD backtrace:
  
  2025-09-10T12:35:32.213Z cpu73:2099542)@BlueScreen: #PF Exception 14 in world 2099542:VSAN_0x433e2 IP 0x42002179653d addr 0x453b29c9b360
PTEs:0x8000084023;0x8299bfb063;0x873a0db063;0x0;
2025-09-10T12:35:32.213Z cpu73:2099542)Code start: 0x42001f200000 VMK uptime: 11:07:04:23.731
2025-09-10T12:35:32.214Z cpu73:2099542)0x453b0181beb0:[0x42002179653d][email protected]#0.0.0.1+0x61 stack: 0x0
2025-09-10T12:35:32.214Z cpu73:2099542)0x453b0181bed0:[0x420020f16ec8][email protected]#0.0.0.1+0x6c1 stack: 0x433e22a72d00
2025-09-10T12:35:32.214Z cpu73:2099542)0x453b0181bfa0:[0x42001f39f7f8]vmkWorldFunc@vmkernel#nover+0x31 stack: 0x42001f39f7f4
2025-09-10T12:35:32.214Z cpu73:2099542)0x453b0181bfe0:[0x42001f8d67b2]CpuSched_StartWorld@vmkernel#nover+0xbf stack: 0x0
2025-09-10T12:35:32.215Z cpu73:2099542)0x453b0181c000:[0x42001f344cef]Debug_IsInitialized@vmkernel#nover+0xc stack: 0x0
2025-09-10T12:35:32.219Z cpu73:2099542)base fs=0x0 gs=0x420052400000 Kgs=0x0




Environment

VMware vSAN ESA 8.x

Cause

PSOD was reported when the host was manually rebooted which killed an existing VSI thread(world). Later after reboot, vSAN server picked the lingering zdom request sent from that VSI thread and tried to access its stack memory by running its callback which was already gone.

It is very rare for a ESXi host to run into such case where a VSI thread of that particular VSI node killed in the middle.



/var/run/log/vmkernel log of the vSAN host reports "Failed to wait for DOMVsiGetZdomSnapChainCB to run in vsan server context: World is marked for death".

2025-09-10T12:32:32.552Z cpu5:2101833 opID=85840e0c)DOM: DOMVsiGetZdomObjectSnapDepthCountIterCB:15107: a82b7166-6af7-54c9-6fe6-############: Failed to wait for DOMVsiGetZdomSnapChainCB to run in vsan server context: World is marked for death
2025-09-10T12:32:32.552Z cpu5:2101833 opID=85840e0c)DOM: DOMVsiIterateZdomObjectsCB:8447: a82b7166-6af7-54c9-6fe6-############: Failed to run callback on zdom object: World is marked for death

Resolution

Currently there is no resolution for the PSOD issue. VMware engineering already is aware about this issue.