ESXi host experience PSOD HPE ProLiant XL645d Gen10 Plus Servers Running VMware ESXi 8.0 When a VM Is Started With an NVIDIA H100 PCIe GPU Configured as a PCI Pass-Through Device.

Products

• VMware vSphere ESX 8.x

Issue/Introduction

ESXi host might enter PSOD with a similar trace:

Environment

VMware vSphere 8.0

Resolution

To resolve the issue, perform the following steps:

1. Navigate to the file "/etc/vmware/passthru.map" that controls how specific devices should reset.
   
   
   
   This file includes
   
   ----------------------
   
   # passthrough attributes for devices
   #
   # file format: vendor-id device-id resetMethod fptShareable
   # vendor/device id: xxxx (in hex) (ffff can be used for wildchar match)
   # reset methods: flr, d3d0, link, bridge, default
   # fptShareable: true/default, false
   ...
   # NVIDIA
   10de ffff bridge false
   
   ----------------------
2. Use "Bridge Reset" to reset any NVIDIA pass-through device.
3. Remove or comment out this line "# 10de ffff bridge false"
   
   
   

    

4. Reboot the VMware ESXi system.
5. After the VMware ESXi system is rebooted, the next time the Virtual Machine starts with NVIDIA pass-through device, the system should use "Function Level Reset (FLR)" rather than "Bridge Reset" which avoids the Purple Diagnostic Screen (PSOD).

Additional Information

Refer vendor document for further details:

https://support.hpe.com/hpesc/public/docDisplay?docId=a00133193en_us&docLocale=en_US