NVIDIA GRID – Could not initialize plugin ‘/usr/lib64/vmware/plugin/libnvidia-vgx.so’ for vGPU “profile_name”
If you are using the NVIDIA A30 GPU graphic card based on the Ampere architecture, especially if you are building a NVIDIA AI Enterprise platform, you need to activate SR-IOV in the BIOS settings of your servers, otherwise your VMs will not power on and you will be left with the error:
In my case I have two Dell R750 and I can activate SR-IOV directly over iDRAC
(BIOS Settings->Integrated Devices) and reboot the Server.
Single Root I/O Virtualization (SR-IOV) enables multiple virtualized guest operating systems to natively access a shared PCI-e device.
In other situations, you may need to disable ECC Mode on all GPUs if you are using NVIDIA vGPU. Check ECC Mode status by:
nvidia-smi -q
To disable ECC Mode of your GPU put your Host into the maintenance mode, start an SSH session to your ESX host and apply the following commands:
nvidia-smi -e 0
reboot
After the reboot check with the command “nvidia-smi -q” if the ECC Mode is disabled. Your VM should now be able to power on and use the assigned vGPU. And that’s it, if you have any questions please use the comment section below. 🙂