Yes, you can update the firmware on the DGX A100 without using PXE booting. Here are the detailed steps for updating the firmware using alternative methods:
Using the DGX A100 FW Update Utility
The DGX A100 System Firmware Update utility is available in a tarball and as a `.run` file. You can update the firmware using one of the following methods:
1. NVSM (NVIDIA System Management): This method provides convenient commands to update the firmware using the firmware update container.
- Example command to update all firmware components:
bash
nvsm(/system/localhost/firmware/install)-> set Flags=update_fw\ all
- Note: An escape is needed before blank spaces when setting the flags.
2. Using Docker: This method involves running the firmware update container using Docker.
- Example command to update all firmware components:
bash
sudo docker run --rm --privileged -ti -v /:/hostfs nvfw-dgxa100:24.11.1 update_fw all
- Ensure Docker is installed and running on your system.
3. Using the `.run` file: This is a self-extracting package that embeds the firmware update container tarball.
- Example command to update all firmware components:
bash
sudo ./nvfw-dgxa100_24.11.1_241107.run update_fw all
Updating Specific Firmware Components
If you want to update specific firmware components, such as the BMC (Baseboard Management Controller) or SBIOS (System BIOS), you can specify them in the command. For example:
- NVSM Example:
bash
nvsm(/system/localhost/firmware/install)-> set Flags=update_fw\ BMC\ SBIOS
- Docker Run Example:
bash
sudo docker run --rm --privileged -ti -v /:/hostfs nvfw-dgxa100:24.11.1 update_fw BMC SBIOS
- `.run` File Example:
bash
sudo ./nvfw-dgxa100_24.11.1_241107.run update_fw BMC SBIOS
Non-Interactive Mode
To run the firmware update in non-interactive mode, you can use the `set_flags auto=1` option. This is particularly useful for automating the update process.
Requirements
Ensure that the necessary modules are installed on your system, including `nvidia_vgpu_vfio`, `nvidia-uvm`, and `nvidia-drm`, as required by the firmware update container[2].
Downloading Firmware Updates
Firmware updates can be downloaded from the NVIDIA Enterprise Support Portal. Navigate to the section for DGX A100 firmware updates to obtain the latest versions[3].
By following these steps, you can update the firmware on your DGX A100 system without relying on PXE booting.
Citations:
[1] https://docs.nvidia.com/dgx/dgxa100-fw-container-release-notes/using-utility.html
[2] https://docs.nvidia.com/dgx/pdf/dgxa100-fw-container-release-notes.pdf
[3] https://kb.brightcomputing.com/knowledge-base/how-to-upgrade-dgx-a100-firmware-from-headnode/
[4] https://docs.nvidia.com/dgx/dgxa100-fw-container-release-notes/dgxa100-fw-container-release-notes.pdf
[5] https://docs.nvidia.com/dgx/dgxa100-user-guide/updating-restoring-sw.html
[6] https://support.brightcomputing.com/manuals/10/admin-manual.pdf
[7] https://github.com/NVIDIA/deepops/blob/master/docs/deepops/dgx-diagnostic-firmware.md
[8] https://craigulmer.com/data/2023/231004_glinda_standup.pdf