"LiveInstallationError" error reported while removing NVIDIA VIBs from ESXi host

Products

• VMware vSphere ESXi
• VMware vSphere ESXi 7.0
• VMware vSphere ESXi 8.0

Issue/Introduction

• The following error was reported while uninstalling NVIDIA VIBs from the ESXi host.
  
  #esxcli software vib remove -n <NVIDIA_driver_VIB_name>

[LiveInstallationError]",
--> " NVD_bootbank_NVD-AIE_ESXi_7.0.2_Driver_525.105.14-1OEM.702.0.0.17xxx0552: Failed to unmount tardisk nvd_boot.v00 of VIB NVD_bootbank_NVD-AIE_ESXi_7.0.2_Driver_525.105.14-1OEM.702.0.0.17630552: Error in running [rm /tardisks/nvd_boot.v00]:",

• VIB removal fails with the same error even after stopping xorg service along with unloading the vmk module.

#/etc/init.d/xorg stop

Remove the NVIDIA VMkernel driver by running the command:

#vmkload_mod -u nvidia

#esxcli software vib remove -n <NVIDIA_driver_VIB_name>

Environment

VMware vSphere ESXi 7.x
VMware vSphere ESXi 8.x

Cause

NVIDIA introduced the optional NVIDIA GPU Management Daemon.

As per their documentation: https://docs.nvidia.com/vgpu/16.0/grid-vgpu-user-guide/index.html#update-vgpu-vib-vmware-vsphere, this daemon must be stopped for VIB removal/install/update.

Resolution

Stop the NVIDIA daemon before removing the NVIDIA VIB by following the steps below:

1. Command to stop the NVIDIA GPU Management Daemon 

   #/etc/init.d/nvdGpuMgmtDaemon stop

2. Remove NVIDIA VIB from ESXi host

    # esxcli software vib remove -n <NVIDIA_driver_VIB_name>

Additional Information

Installing and configuring the NVIDIA VIB on ESXi