DGX-2 System Firmware Update Container Version 18.09.4

The DGX Firmware Update container version 18.09.4 is available.

  • Package name:nvfw-dgx2_18.09.4.tar.gz
  • Image name: nvfw-dgx2_18.09.4

Contents of the DGX-2 System Firmware Container

This container includes the firmware binaries and update utilities for the firmware listed in the following table.

Component Version
BMC 0.97.01
SBIOS 0.14
M.2 SSD (Samsung) CXV8601Q
U.2 SSD (Micron) 101008R0
VBIOS 88.00.6B.00.01

Known Issues

BMC Update Timeout

Issue

The container update may hang and report a BMC update timeout.

Workaround

If the container does not recover, stop the container as follows:
  1. From another terminal session, find the CONTAINER ID of the firmware container instance.
    # sudo docker ps | grep nvfw-dgx2

    Example output:

    CONTAINER ID    IMAGE                 COMMAND                           CREATED          STATUS        
    2e76a51fd85b    nvfw-dgx2_08.19.1     "/usr/bin/python /sr\u2026"       5 seconds ago     Up 4 seconds 
  2. Using the CONTAINER ID, terminate the instance.
    # sudo docker kill <container-id>

    Example:

    # sudo docker kill 2e76a51fd85b
  3. Determine whether the updates were performed by querying the currently installed firmware using the show_version option.
    # sudo docker run --privileged -v /:/hostfs <image-name> show_version
  4. If the BMC is still downlevel, then force the BMC update by using the -f option.
    # sudo docker run --rm --privileged -ti -v /:/hostfs <image-name> update_fw -f bmc
  5. If the issue still occurs, then reboot the system and try to perform the update.
  6. If the issue still occurs, then run nvsm dump health and submit the log files to NVIDIA Enterprise Support.

VBIOS Not Updated on DGX KVM Host

Issue

On a DGX-2 System that has been converted to a DGX KVM host, the VBIOS will not get updated if the GPU is being used by a guest GPU VM.

Explanation

All guest GPU VMs must be stopped before running the container to update the VBIOS. To stop the VMs, run the following from the KVM host for each guest GPU VM.

virsh shutdown <vm-domain>