VPXD crashes with core.vpxd-worker after expanding disk size which is configured with NVMe controller on vCenter Server 8.0 U3

Products

• VMware vCenter Server 8.0
• VMware vSphere ESXi 8.0

Issue/Introduction

• There are two conditions to encounter this vpxd crash:

1. 1. vCenter Server is running on 8.0 U3 while ESXi is 8.0 U1 or lower.
   2. Performing a disk expansion of a VM which is using a NVMe controller through the vSphere client.

• A vpxd crash can be detected from the /var/log/vMonCoredumper.log on vCenter Server:

YYYY-MM-DDT##:##:##:###Z In(05) host-####### Notify vMon about vpxd-worker dumping core. Pid : #######
YYYY-MM-DDT##:##:##:###Z In(05) host-####### Successfully notified vMon.
YYYY-MM-DDT##:##:##:###Z In(05) host-####### Successfully generated core file /var/core/core.vpxd-worker.#######.

• A reconfigure task is observed in the vpxd logs on vCenter Server (var/log/vmware/vpxd/vpxd-###.log) before the crash:

YYYY-MM-DDT##:##:##:###Z info vpxd[#######] [Originator@#### sub=vpxLro opID=######-#######-auto-######-h5:########-##] [VpxLRO] -- BEGIN task-###### -- vm-##### -- vim.VirtualMachine.reconfigure 

• Virtual machine NVMe controller usage can be validated through the vSphere UI or via the vmx file (see Locating virtual machine log files on an ESXi host). Search the vmx file for references to nvme similar to the line below:

nvme0.present = "TRUE"

Environment

• VMware vCenter Server 8.0 U3
• VMware vSphere ESXi 8.0 U1 and lower

Cause

vpxd service crashes due to a sharedBus property check added in vCenter Server 8.0 U3. This property is unset on 8.0 U1 ESXi hosts or lower. This issue is not seen if the ESX host is 8.0.2 and above.

Resolution

This issue is resolved in vCenter to 8.0 U3b or later.

Workaround:

Preform the disk size expansion using host UI.

1. Log in to Host UI client directly.
2. Extend Disk size from the Edit settings of the VM.