Hyper Converged Infrastructure (HCI/aSV)

Sangfor HCI and aSV provide a unified infrastructure combining compute, storage, networking, and built-in security to simplify deployment, operations, and services.
{{ $t('productDocDetail.guideClickSwitch') }}
{{ $t('productDocDetail.know') }}
{{ $t('productDocDetail.dontRemind') }}
6.11.1R1
Hyper Converged Infrastructure (HCI/aSV) {{ breadTitle }} User Manual Virtual Machine Configuration Guidance Virtual Machine Management Manage vGPU Graphics Driver Installation and Licensing for Linux VMs
{{sendMatomoQuery("Hyper Converged Infrastructure (HCI/aSV)","Graphics Driver Installation and Licensing for Linux VMs")}}

Graphics Driver Installation and Licensing for Linux VMs

{{ $t('productDocDetail.updateTime') }}: 2026-01-05

Function Description

After configuring vGPU for a VM, you need to install a graphics driver for the VM before proper use.

Precautions

None

Prerequisites

If the GPU is compatible with HCI, refer to the NVIDIA official website for the compatibility between the Linux operating system and GPUs.

Procedure

  1. Configure the IP address, netmask, and gateway for the Linus VM.
  2. Disable nouveau, which is an open-source NVIDIA driver that may interfere with subsequent driver installation. The procedure for disabling nouveau is as follows.

• In CentOS Linux 7 or Red Hat Enterprise Linux 7, add the following content to the dist-blacklist.conf file in the /usr/lib/modprobe.d or /lib/modprobe.d directory.

blacklist nouveau

options nouveau modeset=0

• In Ubuntu 16.04, create the blacklist.conf file containing the following content in the /etc/modprobe.d directory.

blacklist nouveau

options nouveau modeset=0

Run the $ sudo update-initramfs –u command.

• Run the dracut --force command and then restart the VM.

dracut --force

reboot

• Run the lsmod | grep nouveau command. If no output is returned, nouveau is disabled successfully.

  1. Close the X window by running the init 3 command or by other means.
  2. Install the driver.

• Copy the correct driver file to the system. For example, in CentOS, the file is NVIDIA-Linux-x86_64-390.75-grid.run.

• Modify file permissions as the super administrator and run the following command. Note that you need to add the following parameters before running the command.

sudo chmod +x NVIDIA-Linux-x86_64-390.75-grid.run

./NVIDIA-Linux-x86_64-390.75-grid.run -Z --no-opengl-files

• After the installation starts, you will see multiple blue dialog boxes. Proceed as prompted and select default options, as shown in the figures below.

• If you are asked whether to install compatibility libraries, select Yes.

• If you are asked whether to automatically upgrade the X configuration file, select No.

• The installation is completed.

  1. Obtain the license.

• Download the client license key file (.tok file) as instructed in section 5.2.15.1 "Deploy an NVIDIA Licensing Server."

• Run the cd /ect/nvidia command to enter the /nvidia directory and change gridd.conf.template to gridd.conf.

• Edit gridd.conf. You only need to set FeatureType to the number indicating the vGPU license type and can leave ServerAddress unspecified.

• Copy the downloaded token file to the /etc/nvidia/ClientConfig Token directory.

For example,

copy the client_configuration_token_11-15-2021-11-53-00.tok file to /etc.nvidia/ClientConfigToken/.

• Run the systemctl restart nvidia-gridd command to restart the nvidia-gridd service. Then, you will see that the license is applied successfully.

• You can run the NVIDIA-smi -q command to query the license status.