1. Introduction to the NVIDIA DGX-1 Deep Learning System

The NVIDIA® DGX-1™ Deep Learning System is the world’s first purpose-built system for deep learning with fully integrated hardware and software that can be deployed quickly and easily.



Using the DGX-1: Overview

The NVIDIA DGX-1 is designed to operate in one of two modes — Base OS mode, and Cloud Managed mode. Cloud Management is currently not available, but will be available at a future date. Availability will vary by region.

Base OS mode provides the base operating system on the DGX-1 for customers who want to use their own on-site scheduling and management software and who will build and run their own applications.

Hardware Specifications

Components

Component Qty Description
Base Server 1 Dual Intel® Xeon® CPU motherboard with x2 9.6 GT/s QPI, 8 Channel with 2 DPC DDR4, Intel®X99 Chipset, AST2400 BMC
1 GPU Baseboard supporting 8 SXM2 modules (Cube Mesh) and 4 PCIE x16 slots for InfiniBand NICs
1 Chassis with 3+1 1600W Power supply and support for up to five 2.5 inch drives
1 10/100BASE-T IPMI Port
1 RS232 Serial Port
2 USB 3.0 Ports
Power Supply 4 1600 W each.
CPU 2 Intel® Xeon® E5-2698 v4, 20-core, 2.2GHz, 135W
GPU 8 (Option 1) Tesla P100, featuring
  • 170 teraflops, FP16
  • 16 GB memory per GPU
  • 28,672 NVIDIA CUDA® Cores
(Option 2) Tesla V100, featuring
  • 960 teraflops, FP16
  • 16 GB memory per GPU
  • 40,960 NVIDIA CUDA® Cores
  • 5120 NVIDIA Tensor Cores
System Memory 16 32 GB DDR4 LRDIMM (512 GB total)
SAS Raid Controller 1 8 port LSI SAS 3108 RAID Mezzanine
Storage (RAID 0) (Data) 4 1.92 TB, 6 Gb/s, SATA 3.0 SSD
Storage (OS) 1 480 GB, 6 Gb/s, SATA 3.0 SSD
10 GbE NIC 1 Dual port, 10GBASE-T, network adapter Mezzanine
InfiniBand EDR NIC 4 Single port, x16 PCIe, Mellanox ConnectX-4 VPI MCX455A-ECAT

Mechanical

Feature Description
Form Factor 3U Rackmount
Height 5.16” (13.1 cm)
Width 17.5" (44.4 cm)
Depth 34.1" (86.6 cm)
Gross Weight 134 lbs (61 kg)

Power

Input Specification for Each Power Supply Comments
200-240 V (ac) 3200 W max.

1600 W @ 200-240 V,

8 A, 50-60 Hz

The DGX-1 contains four load-balancing power supplies, with 3+1 redundancy.

Connections and Controls

ID Type Qty Description
1 Power button 1

Press to turn the DGX-1 on or off.

Blue: System power on

Off: System power off

Amber (blinking): DC Off and fault

Amber and blue (blinking): DC On and fault

2 ID button 1 Press to cause an LED on the back of the unit to flash as an identifier during servicing.
3 InfiniBand 4 QSFP28 port; Mellanox ConnectX-4 VPI MCX455A-ECAT, EDR IB (100Gb), x16 PCIe
4 USB 2 USB 3.0 ports are available to connect a keyboard.
5 VGA 1 The VGA port connects to a VGA capable monitor for local viewing of the DGX-1 setup console or base OS.
6 DB9 1 RS232 serial port for internal debugging
7 AC input 4 Power supply inputs
8 Ethernet (RJ45) 2 10GBASE-T dual port network adapter Mezzanine
9

IPMI (RJ45)

1 10/100BASE-T Intelligent Platform Management Interface (IPMI) port

Rear Panel Power Controls

ID Type Qty Description
1 Power button 1 Press and hold the power button for four seconds to turn off the motherboard. The BMC remains live.
2 Power LED 1

Off: Power off

Blue (steady): Power on

Blue (blinking): BMC reports system health fault.

3 Main Board Status LED 1

Off: Normal

Amber (blinking): BMC reports system health fault.

LAN LEDs

LEDs next to each Ethernet port indicate the connection status as described in the table below:

LED Status Description

1

(Port 1 Link/Activity)

Amber (steady) LAN link
Amber (blinking) LAN access (off when there is traffic)
Off Disconnected

2

(Port 1 Speed)

Green 10 Gb/s
Amber 1 Gb/s
Off 100 Mb/s

3

(Port 0 Link/Activity)

Amber (steady) LAN link
Amber (blinking) LAN access (off when there is traffic)
Off Disconnected

4

(Port 0 Speed)

Green 10 Gb/s
Amber 1 Gb/s
Off 100 Mb/s

IPMI Port LEDs

LEDs on the IPMI port indicate the connection status as described in the table below:

Link Activity Description
Off Off Unplugged
Green (steady) Green (blinking) 100M active link
Off Green (blinking) 10M active link

Hard Disk Indicators

ID Feature Description
1 Button and release lever for removing the HDD
2 HDD present LED

Blue (Steady): Drive present

Blue (Blinking twice/sec): Identification (such as when initializing or locating through the SBIOS)

Blue (Blinking once/sec): Rebuilding (such as when creating a RAID array)

Amber (Steady): Warning/failure

Off: Slot empty

3 HDD activity LED Blue: Access

Power Supply Unit (PSU) LED

The PSU LED indicates the operation status of the PSU as described in the table below:

Activity Description
Green Normal operation
Amber (blinking) Power off; Fault
Green (blinking) Power on; Standby mode

2. Installation and Setup

2.1. Registering Your DGX-1

Be sure to register your DGX-1 with NVIDIA as soon as you receive your purchase confirmation e-mail. Registration enables your hardware warranty and allows you to set up an NVIDIA DGX Container Registry account.

To register your DGX-1, you will need information provided in your purchase confirmation e-mail. If you do not have the information, send an e-mail to NVIDIA Enterprise Support at enterprisesupport@nvidia.com.

  1. From a browser, go to the NVIDIA DGX Product Registration page.
  2. Enter all required information and then click SUBMIT to complete the registration process and receive all warranty entitlements and, if applicable, DGX-1 support services entitlements.

Obtaining Software and Software Updates

You must register your DGX-1 in order to receive software updates. Once registered, you will receive an email notification whenever a new software update is available. You can access software update instructions as well as software downloads through the Enterprise Support site as follows:

  • From your browser, go to NVIDIA Enterprise Services, and log in.
  • Click the Announcements tab, which contains download links and supplemental documentation.
  • Refer to the DGX OS Server Software Release Notes for instructions on how to perform a software update.

Choosing a Setup Location / Site Preparation

Decide on a suitable location for setting up and operating the DGX-1. The location should be clean, dust-free, and well ventilated.

General Conditions

  • Prepare a sufficiently wide aisle to accommodate the unboxed chassis (chassis dimensions - 5.16”H x 17.5"W x 34.1"D).
  • The rack must accommodate a 134 lb, 3U rack mount system (chassis dimensions - 5.16”H x 17.5"W x 34.1"D).
  • The rack must have square mounting holes.
  • Leave enough clearance in front of the rack (36" (91.4 cm)) to enable you to install the unit into the rack.
  • Leave approximately 30" (76.2cm) of clearance in the back of the rack to allow for sufficient airflow and ease in servicing.
  • Always make sure the rack is secured and stable before adding or removing the appliance or any other component.
  • Prepare adequate sound-proofing: The equipment fans can generate 72-100 dBA.

Environmental Conditions

  • Operating environment
    • Temperature: 5 ◦ C to 35 ◦ C (41 ◦ F to 95 ◦ F)
    • Relative humidity: 20% to 85% noncondensing
  • Air flow
    • The chassis fans can produce a maximum of 340 CFM of air flow.
    • Do not block the ventilation areas at the front and rear of the chassis.
    • Minimize any restrictions on air flow around the chassis.

Connections

  • Power:
    • The DGX-1 is powered through four 1600W power supply units, each rated at 200-240VAC, 8A, 50/60 Hz. Total system power: 3200W
    • C13/C14 cables provided for each power supply to connect to a compatible PDU.
  • Network: Dual 10GBASE-T RJ45 connection
  • IPMI: 10/100BASE-T RJ45 connection
  • InfiniBand: Qty 4 - QSFP28 ports, InfiniBand and Ethernet compliant

Preparing for Network Access

  • The IPMI port and Ethernet ports can be connected to your local LAN.

    These ports are configured for DHCP by default.

    • To use DHCP, connect the port to a local DHCP server which should provide an IP address and assign a DNS configuration to the DGX-1.
    • If DHCP is not available, then you will need to set up a static IP for each Ethernet port.
  • NVIDIA recommends that customers follow best security practices for BMC management (IPMI port). These include, but are not limited to, such measures as:
    • Restricting the DGX-1 IPMI port to an isolated, dedicated, management network
    • Using a separate, firewalled subnet
    • Configuring a separate VLAN for BMC traffic if a dedicated network is not available
  • If you will be operating the DGX-1 in cloud-managed mode, then
    • Make sure that DNS is enabled
    • Make sure that the ports listed in the following table are open and available on your firewall to the DGX-1:
    Port (Protocol) Direction Use
    53 (UDP) Outbound DNS
    80 (TCP) Outbound HTTP, package updates
    123 (UDP)

    Outbound/

    Inbound

    NTP client
    443 (TCP)

    Inbound/

    Outbound

    For internet (HTTP/HTTPS) connection to DGX-1 Cloud Services

    If port 443 is proxied through a corporate firewall, then WebSocket protocol traffic must be supported

    2376 (TCP) Inbound For interacting with running containers using attach/exec commands
  • If you will be using the DGX-1 in Base OS mode, make sure your network can connect to the following:

    If access to those URLs requires use of a proxy, refer to Setting Up a System Proxy for setup instructions.

Unpacking the DGX-1

  1. Remove the shrinkwrap.
  2. Collapse the yellow "Do not stack" cone, if included.
  3. Open the main DGX-1 box, then remove the accessory and rail kit boxes.

    CAUTION: At least four people, or a mechanical assist, are required to remove the DGX-1 from the box. To reduce the risk of personal injury or damage to the equipment, always observe local occupational health and safety requirements and guidelines for material handling.

    DO NOT use the handles at the front of the DGX-1 to lift the unit. The handles are designed for sliding the unit out of a rack, and not for carrying the full weight of the DGX-1.

  4. Remove the protective plastic sheet from the top of the DGX-1.
  5. Preserve and retain packaging.
  6. Be sure to inspect each piece of equipment shipped in the packing box. If anything is missing or damaged, contact your supplier.

What's In the Box

The NVIDIA DGX-1 shipping box includes the following:

  • NVIDIA DGX-1
  • Bezel
  • Rail hardware kit
  • Accessory Box
    • AC Power Cables (qty 4 – IEC 60320 C13/14, compatible with data center PDUs)
    • Hard disk bay screws
    • Toxic Substance Notice & Safety Instructions
    • Quick Start Guide
    • DVD containing source files for open source software
Note: The four power cables included in the box are not optional. All power cables are necessary and must be plugged into individual 10 A capable sockets for optimal DGX-1 operation. Failure to do so can result in a reduction in power redundancy, a reduction in performance, or a complete system failure.

Installing the DGX-1 Into a Rack

CAUTION: To prevent bodily injury when mounting or servicing the DGX-1 in a rack, you must take special precautions to ensure that the system remains stable. The following guidelines are provided to ensure your safety.

• The DGX-1 should be mounted at the bottom of the rack if it is the only unit in the rack.

• When mounting the DGX-1 in a partially filled rack, load the rack from the bottom to the top with the heaviest component at the bottom of the rack.

• If the rack is provided with stabilizing devices, install the stabilizers before mounting or servicing the DGX-1 in the rack.

• The DGX-1 weighs approximately 134 lbs, so an equipment lift is required to safely lift the unit and then accurately align the chassis rails with the rack rails.

DO NOT use the handles at the front of the DGX-1 to lift the unit. The handles are designed for sliding the unit out of a rack, and not for carrying the full weight of the DGX-1.

Installing the Rails

Note: The rail assemblies shipped with the appliance fit into a standard 19” rack between 26-inches and 33.5-inches deep (66 cm to 85 cm). The outer rail is adjustable from approximately 23.5” to 34” (59.7 cm to 86.4 cm)

Refer to the instructions in the rail packaging for details on installing the rails onto the rack and chassis.

The following are supplemental instructions:

  1. Use a Phillips screwdriver to assist in mounting the rails to the rack.
  2. If necessary, detach the inner rails from the outer slide rails.
  3. Follow any designations on the inner rail (or its outer rail mate) to determine the proper orientation and positioning to connect to the chassis, then secure to the chassis.IMPORTANT: Make sure that the reinforced hole at the front end of the rail is positioned on the bottom side of the rail, and that it aligns with the thumbscrew on the front of the DGX-1. If the hole is positioned on the top side, then the rail is on the wrong side of the DGX-1 and the DGX-1 will not fit properly in the rack.
  4. Follow any designations on the outer slide rail to determine front/back and left-side/right-side positioning against the rack.
  5. Secure the back of one of the slide rails to the rack, then extend the rail until it fits securely to the front of the rack.
  6. Secure the slide rail to the front of the rack.
  7. Repeat steps 4-6 for the other slide rail.

Mounting the DGX-1

Note:CAUTION: Stability hazard — The rack stabilizing mechanism must be in place, or the rack must be bolted to the floor before you slide the DGX-1 out for servicing. Failure to stabilize the rack can cause the rack to tip over.
  1. Confirm that the DGX-1 has the inner rails attached and that you have already mounted the outer rails into the rack.
  2. With the front of the unit facing away from the rack, use an equipment lift to assist in sliding the unit into the rack as follows:

    CAUTION: The DGX-1 weighs approximately 134 lbs, so an equipment lift is required to safely lift the unit and then accurately align the chassis rails with the rack rails.

    1. Align the inner chassis rails with the front of the outer rack rails.
    2. Slide the inner rails into the outer rails, keeping the pressure even on both sides (you may have to depress the locking tabs when inserting).

      When the DGX-1 has been pushed completely into the rack, you should hear the locking tabs "click" into the locked position.



  3. Lock the unit in place using the thumb screws located on the front of the unit.





Attaching the Bezel

The bezel is designed to attach easily to the front of the DGX-1.

  1. Prepare the DGX-1 by making sure that the power supply handles (located at the power supply fans) are flipped up.



  2. Move any other obstructions, such as cable ties, away from the outer edge of the DGX-1.
  3. With the bezel positioned so that the NVIDIA logo is visible from the front and is on the left hand side, line up the pins near the corners of the DGX-1 with the holes in back of the bezel, then gently press the bezel against the DGX-1.



    The bezel is held in place magnetically .

Connecting the Power Cables

  1. Open the accessory box and remove the four C13/C14 power cables.
  2. Use the cables to connect each of the four plugs at the right-rear of the DGX-1 to a PDU.



    1. Secure each cable to the DGX-1, using the power cable retention clips attached to the power plugs.
    2. Connect each cable to the PDU. Ensure that the cables are distributed over at least two circuits and, if using 3-phase PDUs, they are balanced across all phases as much as possible. Ideally, each cable should connect to a different PDU.
    3. Verify that each cable is firmly inserted into the PDU. There is usually a click to indicate full insertion.

Connecting the Network Cables

  1. Using an Ethernet cable, connect one of the dual Ethernet ports (em1 or em2) to your LAN for internet access to the NVIDIA Cloud Portal, remote access to launched application containers on the DGX-1, or to connect to the DGX-1 using SSH.



    The left-side/right-side ethernet port designation depends on the Base OS software version installed on the DGX-1 as listed in the table below.
    Ethernet Port Position Port Designation: Base OS Software 2.x and earlier Port Designation: Base OS Software 3.x and later
    Right Side em1 enp1s0f0
    Left Side em2 enp1s0f1
    Note: NVIDIA recommends connecting only one of the Ethernet ports to your LAN. If you are connecting both Ethernet ports, they must each be connected to separate networks, The DGX-1 is not configured from the factory to have multiple Ethernet interfaces on the same network.
  2. Using an Ethernet cable, connect the IPMI (BMC) port to your LAN for remote access to the base management controllerr (BMC). Vefiy that all network cables are firmly inserted into the DGX-1 and the associated network switch.

Setting Up the DGX-1

These instructions describe the setup process that occurs the first time the DGX-1 is powered on after delivery. Be prepared to accept all EULAs and to set up your username and password.

  1. Connect a display to the VGA connector, and a keyboard to any of the USB ports.



  2. Power on the DGX-1.



    The system will take a few minutes to boot.

    You may be presented with end user license agreements (EULAs) for the NVIDIA software at this point in the setup, depending on the DGX-1 software version. Accept all EULAs to proceed with the installation.

    You are prompted to configure the DGX-1 software.

  3. Perform the steps to configure the DGX-1 software.
    • Select your time zone and keyboard layout.
    • Create a user account with your name, username, and password.

      You will need these credentials to log in to the DGX-1 as well as to log in to the BMC remotely. When logging in to the BMC, enter your username for both the User ID as well as the password. Be sure to create a unique BMC password at the first opportunity.

      Note: The BMC software will not accept "sysadmin" for a user name. If you create this user name for the system log in, "sysadmin" will not be available for logging in to the BMC.
    • Choose a primary network interface for the DGX-1.
      Note: After you select the primary network interface, the system attempts to configure the interface for DHCP and then asks you to enter a hostname for the system. If DHCP is not available, you will have the option to configure the network manually. If you need to configure a static IP address on a network interface connected to a DHCP network, select Cancel at the Network configuration – Please enter the hostname for the system screen. The system will then present a screen with the option to configure the network manually.
    • Choose a hostname for the DGX-1.
    • Choose to install predefined software.

      Press the space bar to select or deselect the software to install.

      Note: By default, the DGX-1 installs only minimal software packages necessary to ensure system functionality. You can deselect the OpenSSH package; however, NVIDIA recommends that you keep this package selected, and uninstall it only if required by your IT security policy.
  4. Select OK to continue. You may be presented with end user license agreements (EULAs) for the NVIDIA software at this point in the setup, depending on the DGX-1 software version. Accept all EULAs to complete the installation. The system completes the installation, reboots, then presents the system login prompt:
    <hostname> login:
    Password:
  5. Log in.

Post Setup Instructions for DGX OS Server Software Version 2.x and Earlier

These instructions apply if your DGX-1 is installed with software version 2.x or earlier.

To determine the DGX OS Server software version on your system, enter the following command.
$ grep VERSION /etc/dgx-release
DGX_SWBUILD_VERSION="3.1.1"
  1. If your network is configured for DHCP, then make sure that dynamic DNS updates are enabled. Check whether /etc/resolv.conf is a link to /run/resolvconf/resolv.conf.
    $ ls -l /etc/resolv.conf
    Expected output:
    lrwxrwxrwx 1 root root 29 Dec  1 21:19 /etc/resolv.conf ->
    ../run/resolvconf/resolv.conf
    • If the expected output appears, then skip to step 2.
    • If this does not appear, then enable dynamic DNS updates as follows:
    1. Launch the Resolvconf Reconfigure package.
      $ sudo dpkg-reconfigure resolvconf
      The Configuring resolvconf screen appears.
    2. Select <Yes>when asked whether to prepare /etc/resolv.conf for dynamic updates.
    3. Select <No> when asked whether to append original file to dynamic file.
    4. Select <OK> at the Reboot recommended screen. You do not need to reboot. You are returned to the command line.
    5. Bring down the interface, where <network interface> is em1 or em2, whichever you have set up as your primary network interface.
      $ sudo ifdown <network interface>
      Expected output:
      ifdown: interface <network interface> not configured
    6. Bring up the interface, where <network interface> is em1 or em2, whichever you have set up as your primary network interface.
      $ sudo ifup <network interface>
      Expected output (last line):
      …
      bound to <IP address> -- renewal in …
    7. Repeat step 1 to confirm that /etc/resolv.conf is a link to /run/resolvconf/resolv.conf.
  2. Make sure that the nvidia-peer-memory module is installed.
    $ lsmod | grep nv_peer_mem
    If the following output appears, then your DGX-1 setup is complete and you do not need to perform the next steps.
    nv_peer_mem            16384  0
    nvidia              11911168  30
    nv_peer_mem,nvidia_modeset,nvidia_uv
    mib_core               143360  13
    rdma_cm,ib_cm,ib_sa,iw_cm,nv_peer_mem,mlx4_ib,mlx5_ib,
    ib_mad,ib_ucm,ib_umad,ib_uverbs,rdma_ucm,ib_ipoib
  3. If there is no output to the lsmod command, then build and install the nvidia-peer-memory module.
    1. Get and install the module.
       $ sudo apt-get update
       $ sudo apt-get install --reinstall mlnx-ofed-kernel-dkms nvidia-peer-memory-dkms
      
      Expected output.
       DKMS: install completed.
       Processing triggers for initramfs-tools (0.103ubuntu4.2) ...
       update-initramfs: Generating /boot/initrd.img-4.4.0-45-generic
      
    2. Add the module to the Linux kernel.
       $ sudo modprobe nv_peer_mem
      
      There is no expected output for this command.
    3. Repeat step 2 to confirm that the nvidia-peer-memory module has been added.

Preparing for Using Docker Containers

This chapter presents an overview of the prerequisites for accessing NVIDIA Docker containers from the Docker command line for use on the NVIDIA® DGX-1™ in base OS mode. These containers include NVIDIA DGX-1 specific software to ensure the best performance for your applications. Using these containers as a basis for your applications should provide the best single-GPU performance and multi-GPU scaling.

Installing Docker and NVIDIA Docker on DGX OS Server Software 2.x or Earlier

To enable portability in Docker images that leverage GPUs, NVIDIA® developed nvidia-docker, an open-source project that provides a command line tool to mount the user mode components of the NVIDIA driver and the GPUs into the Docker container at launch.

As of DGX OS Server software version 3.1.1 and later, Docker and nvidia-docker are part of the base software installation and you do not need to perform the steps in this section. However, if your DGX-1 is installed with software version 2.x or earlier, then follow these instructions to install Docker and nvidia-docker on the system.

To determine the DGX OS Server software version on your system, enter the following command.
$ grep VERSION /etc/dgx-release
DGX_SWBUILD_VERSION="3.1.1"
Ensure your environment meets the prerequisites before installing Docker. For more information, see Getting Started with Docker.
  1. Install Docker.
    $ sudo apt-key adv --keyserver
    hkp://p80.pool.sks-keyservers.net:80 --recv-keys
    58118E89F3A912897C070ADBF76221572C52609D
    $ echo deb https://apt.dockerproject.org/repo ubuntu-trusty main
    | sudo tee /etc/apt/sources.list.d/docker.list
    $ sudo apt-get update
    $ sudo apt-get -y install docker-engine=1.12.6-0~ubuntu-trusty
  2. Edit the /etc/default/docker file to use the Overlay2 storage driver.
    1. Open the /etc/default/docker file for editing.
       $ sudo vi /etc/default/docker
    2. Add the following line:
      DOCKER_OPTS="--storage-driver=overlay2"
      If there is already a DOCKER_OPTS line, then add the parameters (text between the quote marks) to the DOCKER_OPTS environment variable.
    3. Save and close the /etc/default/docker file when done.
    4. Restart Docker with the new configuration.
      $ sudo service docker restart
  3. Install NVIDIA Docker. The following example installs both nvidia-docker and the nvidia-docker-plugin.
    $ wget -P /tmp  
    https://github.com/NVIDIA/nvidia-docker/releases/download/v1.0.1/nvidia-docker_1.0.1-1_amd64.deb
    
    $ sudo dpkg -i /tmp/nvidia-docker*.deb && rm
    /tmp/nvidia-docker*.deb

Configuring Docker IP Addresses

To ensure that the DGX-1 can access the network interfaces for nvidia-docker containers, the nvidia-docker containers should be configured to use a subnet distinct from other network resources used by the DGX-1.

By default, Docker uses the 172.17.0.0/16 subnet. Consult your network administrator to find out which IP addresses are used by your network. If your network does not conflict with the default Docker IP address range, then no changes are needed and you can skip this section.

However, ff your network uses the addresses within this range for the DGX-1, you should change the default nvidia-docker network addresses. The method for accomplishing this depends on the Base OS software version installed on the DGX-1.

  1. If you don't know the Base OS software version installed on the DGX-1, then enter the following and inspect the VERSION entry.
    $ cat /etc/dgx-release
    DGX_NAME="DGX Server"
    DGX_PRETTY_NAME="NVIDIA DGX Server"
    DGX_SWBUILD_DATE="2017-08-02"
    DGX_SWBUILD_VERSION="3.1.1"
    DGX_COMMIT_ID="0a0a8ec9e08836c5e99144dd19ae61690f2d9484"
    DGX_SERIAL_NUMBER=QTFCOU7080017
  2. Follow the instructions in the section appropriate for the software version installed.

Configuring Docker IP Addresses for DGX OS Server Software Version 2.x and Earlier

  1. Open the /etc/default/docker file for editing.
     $ sudo vi /etc/default/docker
  2. Modify the /etc/default/docker file, specifying the correct bridge IP address and IP address ranges for your network. Consult your IT administrator for the correct addresses.
    For example, if your DNS server exists at IP address 10.10.254.254, and the 192.168.0.0/24 subnet is not otherwise needed by the DGX-1, you can add the following line to the /etc/default/docker file:
    DOCKER_OPTS=”--dns 10.10.254.254 --bip=192.168.0.1/24 --fixedcidr=192.168.0.0/24”
    If there is already a DOCKER_OPTS line, then add the parameters (text between the quote marks) to the DOCKER_OPTS environment variable.
  3. Save and close the /etc/default/docker file when done.
  4. Restart Docker with the new configuration.
    $ sudo service docker restart

Configuring Docker IP Addresses for DGX OS Server Software Version 3.1.1 and Later

You can change the default Docker network addresses by either modifying the /etc/docker/daemon.json file or modifying the /etc/systemd/ system/docker.service.d/docker-override.conf file. These instructions provide an example of modifying the /etc/systemd/system/docker.service.d/docker-override.conf to override the default nvidia-docker network addresses.

  1. Open the docker-override.conf file for editing.
     $ sudo vi /etc/systemd/system/docker.service.d/docker-override.conf 
    [Service] 
    ExecStart= 
    ExecStart=/usr/bin/dockerd -H fd:// -s overlay2 --disable-legacy-registry=false
    LimitMEMLOCK=infinity 
    LimitSTACK=67108864
  2. Make the changes indicated in bold below, setting the correct bridge IP address and IP address ranges for your network. Consult your IT administrator for the correct addresses.
    [Service]
    ExecStart= 
    ExecStart=/usr/bin/dockerd -H fd:// -s overlay2 --bip=192.168.127.1/24
          --fixed-cidr=192.168.127.128/25 
           --disable-legacy-registry=false
    LimitMEMLOCK=infinity 
    LimitSTACK=67108864 
    Save and close the /etc/systemd/system/docker.service.d/docker-override.conf file when done.
  3. Reload the systemctl daemon.
     $ sudo systemctl daemon-reload
  4. Restart Docker.
    $ sudo systemctl restart docker

Letting Users Issue Docker Commands

To prevent the docker daemon from running without protection against escalation of privileges, the NVIDIA Docker software requires sudo privileges to run containers.

You can grant the required privileges to users who will run containers on the DGX-1 in one of the following ways:

  • Add each user as an administrator user with sudo privileges.

  • Add each user as a standard user without sudo privileges and then add the user to the docker group.

    This section provides instructions for adding users to the docker group.

Note:WARNING: Only add users to the docker group whom you would trust with root privilege. These instructions make it more convenient for users to access Docker containers; however, the resulting docker group is equivalent to the root user, because once a user is able to send commands to the Docker engine, they are able to escalate privilege and run root level operations. This may violate your organization's security policies. See the Docker Daemon Attack Surface for information on how this can impact security in your system. Always consult your IT department to make sure the installation is in accordance with the security policies of your data center.
Note: The commands in this section require sudo access, and should be performed by a system administrator.

Checking if a User is in the Docker Group

To check whether a user is already part of the docker group, enter the following:

$ groups username 

The output shows all the groups of which that user is a member. If docker is not listed, then add that user.

Creating a User

To create a new user in order to add them to the docker group, perform the following:

  1. Add the user.
    $ sudo useradd username
  2. Set up the password.
    $ sudo passwd username
    Enter a password at the prompts:
    Enter new UNIX password:
    Retype new UNIX password:
    passwd: password updated successfully

Adding a User to the Docker Group

For each user you want to add to the docker group, enter the following command:

$ sudo usermod -a -G docker username

Configuring a System Proxy

If you will be using the DGX-1 in base OS mode, and your network requires use of a proxy, then edit the file /etc/apt/apt.conf.d/proxy.conf and make sure the following lines are present, using the parameters that apply to your network:

Acquire::http::proxy "http://<username>:<password>@<host>:<port>/"; 
Acquire::ftp::proxy "ftp://<username>:<password>@<host>:<port>/"; 
Acquire::https::proxy "https://<username>:<password>@<host>:<port>/";

This is to ensure that Docker is able to access the DGX-1 Container Registry through the proxy. For best practice recommendations on configuring proxies for Docker, see https://docs.docker.com/engine/admin/systemd/#http-proxy.

Configuring NFS Mount and Cache

The DGX-1 includes four SSDs in a RAID 0 configuration. These SSDs are intended for application caching, so you must set up your own NFS drives for long term data storage. The following instructions describe how to mount the NFS onto the DGX-1, and how to cache the NFS using the DGX-1 SSDs for improved performance.

Make sure your DGX-1 is set up in Base OS mode, that you have an NFS server with one or more exports with data to be accessed by the DGX-1, and that there is network access between the DGX-1 and the NFS server.
Note: Skip this section if you are going to use the DGX-1 in cloud-managed mode. The DGX-1 Cloud Services software will set up the NFS cache for you as part of the cloud-managed mode configuration. Similarly, in cloud-managed mode, the person setting up the job will specify any NFS mount requirements for the job at that time.
  1. Check if the cache daemon is installed and configured.
    $ service cachefilesd status
    If the output indicates that cachefilesd is disabled, continue with the following steps. Otherwise, skip to step 7.
  2. Install the cache daemon.
    $ sudo apt-get install cachefilesd
  3. Edit the cache daemon startup file.
    $ sudo vi /etc/default/cachefilesd
    Uncomment the "RUN=yes" line in the startup file and then save the file.
  4. Configure the cache daemon for the DGX-1.
    1. Open the cache daemon configuration file.
      $ sudo vi /etc/cachefilesd.conf
    2. Edit the contents to match the following, then save the file.
      dir /raid
      tag dgx1cache
      brun 25%
      bcull 15%
      bstop 5%
      frun 10%
      fcull 7%
      fstop 3%
      These settings are optimized for Deep Learning workloads, and provide the best throughput for training from large datasets.
  5. Start the cache daemon.
    $ service cachefilesd start
  6. Verify the cache daemon started properly.
    $ service cachefilesd status
    Expected output.
    Checking status of FilesCache daemon cachefilesd
  7. Configure an NFS mount for the DGX-1.
    1. Edit the filesystem tables configuration.
      sudo vi /etc/fstab
    2. Add a new line for the NFS mount, using the local mount point of /mnt.
      <nfs_server>:<export_path> /mnt nfs rw,noatime,rsize=32768,wsize=32768,nolock,tcp,intr,fsc,nofail 0 0
      • /mnt is used here as an example mount point.
      • Consult your Network Administrator for the correct values for <nfs_server> and <export_path>.
      • The nfs arguments presented here are a list of recommended values based on typical use cases. However, "fsc" must always be included as that argument specifies use of FS-Cache.
    3. Save the changes.
  8. Verify the NFS server is reachable.
    ping <nfs_server>
    Use the server IP address or the server name provided by your network administrator.
  9. Mount the NFS export.
    sudo mount /mnt
    /mnt is the example mount point used in step 7.
  10. Verify caching is enabled.
    cat /proc/fs/nfsfs/volumes
    Look for the text FSC=yes in the output. Upon rebooting, the NFS should be mounted and cached on the DGX-1.

4. Configuring and Managing the DGX-1

Using the BMC

The DGX-1 includes a baseboard management controller (BMC) that lets you manage and monitor the DGX-1 independently of the CPU or operating system. You can access the BMC remotely through the Ethernet connection to the IPMI port.

This section describes how to access the BMC, and describes a few common tasks that you can accomplish through the BMC. It is not meant to be a comprehensive description of all the BMC capabilities.

To access the BMC remotely:

  1. Make sure you have connected the IPMI port on the DGX-1 to your LAN.
  2. Open a Java-enabled browser within your LAN and go to http://<IPMI IP Address>/. Use Firefox or Internet Explorer. Google Chrome is not officially supported by the BMC.

  3. Log in. Your initial log in credentials are based on the ones you created when you first set up the DGX-1. Enter your username for both the User ID as well as the Password. User ID: <your username> Password: <your username>.
  4. Be sure to change your password immediately to ensure the security of the BMC. See the next section for instructions on how to change your BMC password.

Creating a Unique BMC Password for Remote Access

When you set up the DGX-1 upon powering it on for the first time, you set up a username and password for the system. These credentials are also used to log in to the BMC remotely, except that the BMC password is the username.

It is strongly recommended that you create a unique password as soon as possible.

Create a unique BMC password as follows:

  1. Open a Java-enabled web browser within your LAN and go to http://<IPMI IP address>/.

    Use Firefox or Internet Explorer. Google Chrome is not officially supported by the BMC.

  2. Log in with the username that you created when you first set up the DGX-1. Enter your username for both the User ID as well as the password: User ID: <your username> Password: <your username>.
  3. From the top menu, click Configuration and then select User.
  4. Select your usename and then click Modify User.
  5. In the Modify User dialog, select Change Password, and then enter your new password in the Password and Confirm Password boxes.
    Note: The BMC software will not accept "sysadmin" for the user name.
  6. Click Modify when finished.

Viewing System Information

The BMC opens to the dashboard, which shows information about the system and system components, such as temperatures and voltages.

Submitting BMC Log Files

The BMC provides automatic logging of system activities and status. The NVIDIA Enterprise Support team uses the log files to assist in troubleshooting. Follow these instructions to obtain the log files to send to NVIDIA Enterprise Support.

  1. Log into the BMC, then click Server Health from the top menu and select Event Log.
  2. Make sure that Text is selected at Format of Download Event Logs.

  3. Click Save Event Logs to download the event logs.

Determining Total Power Consumption

You can use the BMC dashboard to determine total power consumption of the DGX-1 as follows:

  1. Log into the BMC.
  2. From the BMC dashboard, locate the Sensor Monitoring area and then scroll down the page until you see the PSU Input rows.
  3. Add the values for all the PSUs. In this example, the total power consumption would be 216+216+135+27 = 594 watts.

Accessing the DGX-1 Console

  1. Log into the BMC.
  2. From the top menu, click Remote Control and then select Console Redirection.
  3. Click Java Console to open the popup window. The window provides interactive control of the DGX-1 console.

Powering Off / Power Cycling the System Remotely

From the DGX-1 Console Window

If you have opened the Java Viewer (Remote Control->Console Redirection) to view the console window, then you can power cycle, reset, or shutdown the DGX-1 as follows:
  1. From the JViewer top menu, click Power and then select from the available options, depending on what you want to do.

  2. Click Yes and then OK at the Power Control dialog, then wait for the system to perform the intended action.

From the BMC UI

  1. Log into the BMC.
  2. From the top menu, click Remote Control and then select Server Power Control.

  3. Select from the available options according to what you want the system to do, then click Perform Action.

Configuring a Static IP Address for the BMC

This section explains how to set a static IP address for the BMC. You will need to do this if your network does not support DHCP.

Use one of the methods described in the following sections:

Configuring a BMC Static IP Address Using ipmitool

This section describes how to set a static IP address for the BMC from the Ubuntu command line.

Note: If you cannot access the DGX-1 remotely, then connect a display and keyboard directly to the DGX-1.
To view the current settings, enter the following command.
$ sudo ipmitool lan print 1
Set in Progress     : Set Complete
Auth Type Support    : MD5
Auth Type Enable     : Callback : MD5
                     : User  : MD5
                     : Operator : MD5
                     : Admin : MD5
                     : OEM   : MD5
IP Address Source    : DHCP Address
IP Address           : 10.31.241.190
Subnet Mask          : 255.255.255.0
MAC Address          : 54:ab:3a:72:08:a9
SNMP Community String   : Quanta
IP Header            : TTL=0x40 Flags=0x40 Precedence=0x00 TOS=0x10
BMC ARP Control      : ARP Responses Enabled, Gratuitous ARP Disabled
Gratituous ARP Intrvl   : 0.0 seconds
Default Gateway IP   : 10.31.241.1
Default Gateway MAC  : 00:00:00:00:00:00
Backup Gateway IP       : 0.0.0.0
Backup Gateway MAC   : 00:00:00:00:00:00
802.1q VLAN ID       : Disabled
802.1q VLAN Priority : 0RMCP+ Cipher Suites  : 0,1,2,3,6,7,8,11,12,15,16,17
Cipher Suite Priv Max   : XaaaaaaaaaaaXXX
                     :  X=Cipher Suite Unused
                     :  c=CALLBACK
                     :  u=USER
                     :  o=OPERATOR
                     :  a=ADMIN
                     :  O=OEM
To set a static IP address for the BMC, do the following.
  1. Set the IP address source to static.
    $ sudo ipmitool lan set 1 ipsrc static
  2. Set the appropriate address information.
    • To set the IP address (“Station IP address” in the BIOS settings), enter the following and replace the italicized text with your information.
      $  sudo ipmitool lan set 1 ipaddr  10.31.241.190
    • To set the subnet mask, enter the following and replace the italicized text with your information.
      $  sudo ipmitool lan set 1 netmask  255.255.255.0
    • To set the default gateway IP (“Router IP address” in the BIOS settings), enter the following and replace the italicized text with your information.
      $  sudo ipmitool lan set 1 defgw ipaddr  10.31.241.1

Configuring a BMC Static IP Address Using the System BIOS

This section describes how to set a static IP address for the BMC when you cannot access the DGX-1 remotely. This process involves setting the BMC IP address during system boot.

  1. Connect a keyboard and display to the DGX-1, then turn on the DGX-1.
  2. When you see the NVIDIA logo, press Del to enter the BIOS Utility Setup Screen.

  3. At the BIOS Setup Utility screen, navigate to the Server Mgmt tab on the top menu, then scroll to BMC network configuration and press Enter.

  4. Scroll to Configuration Address Source and press Enter , then at the Configuration Address source pop-up, select Static on next reset and then press Enter.

  5. Set the addresses for the Station IP address, Subnet mask, and Router IP address as needed by performing the following for each:
    1. Scroll to the specific item and press Enter.
    2. Enter the appropriate information at the pop-up, then press Enter.

  6. When finished making all your changes, press F10 to Save & Reset, then select Yes at the confirmation pop-up and press Enter.

    You can now access the BMC over the network.

Configuring a BMC Static IP Address Using the BMC Dashboard

  1. Log into the BMC, then click Configuration from the top menu and select Network Settings.
  2. In the IPv4 Configuration section of the Network Settings page, clear the Use DHCP check box, and then enter the appropriate values for the IPv4 Address , Subnet Mask , and Default Gateway fields.

  3. Click Save when done.

Configuring Static IP Addresses for the Network Ports

During the initial boot setup process for the DGX-1, you had an opportunity to configure static IP addresses for the network ports. If you did not set this up at that time, you can configure the static IP addresses from the Ubuntu command line according to the following instructions.

Note: If you cannot access the DGX-1 remotely, then connect a display and keyboard directly to the DGX-1.
  1. Determine the port designation that you want to configure, based on the physical ethernet port that you have connected to your network.

    Use the following port designations according to the DGX-1 Base OS software version installed on the DGX-1:

    Ethernet Port Position Port Designation: Base OS Software 2.x and earlier Port Designation: Base OS Software 3.x and later
    Right Side em1 enp1s0f0
    Left Side em2 enp1s0f1
  2. Edit the interfaces file.
    $ sudo vi /etc/network/interfaces
    ## Configure a static IP
    auto em1
    iface em1 inet static 
      address 192.168.1.14
      gateway 192.168.1.1
      netmask 255.255.255.0
      network 192.168.1.0
      broadcast 192.168.1.255
    Consult your network adiminstrator for the appropriate addresses for your network, and use the port designations that you determined in step 1.
  3. When finished with your edits, press ESC to switch to command mode, then save the file to the disk and exit the editor.
    :wq
  4. Restart the network services to put the changes into effect.
    $ sudo /etc/init.d/networking restart

Obtaining MAC Addresses

These instructions explain how to determine the MAC addresses for the IPMI port (BMC) as well as both ethernet ports of the DGX-1.

The ports are, from left to right, IPMI (BMC), em2 (or enp1s0f1), em1 (or enp1s0f0).

  1. Connect a display and keyboard to the DGX-1.
  2. Turn the DGX-1 on or reboot.
  3. At the NVIDIA logo boot screen, press [F2] or [Del] to enter the BIOS setup screen.

  4. Select the Advanced tab from the top menu, then scroll down to view the two Quanta Dual Port 10G BASE-T Mezzanine items.

    The first item shows the MAC address for ethernet port em1, and the second item shows the MAC address for em2.
  5. Navigate to and select Server Mgmt from the top menu, then scroll down to and select BMC network configuration.

  6. Scroll down to view the Station MAC address.

    This shows the MAC address for the BMC.

5. Maintaining and Servicing the NVIDIA DGX-1

Be sure to familiarize yourself with the NVIDIA Terms & Conditions documents before attempting to perform any modification or repair to the DGX-1. These Terms & Conditions for the DGX-1 can be found through the NVIDIA DGX Systems Support page.

5.1. Problem Resolution and Customer Care

Log on to the NVIDIA Enterprise Services site for assistance with troubleshooting, diagnostics, or to report problems with your DGX-1.

Refer to Submitting BMC Log Files for instructions on how to obtain the BMC log files to assist in troubleshooting.

Using the NVIDIA DGX-1 GPU Diagnostics Tool

The NVIDIA® DGX-1™ GPU Diagnostic tool is a powerful software program for testing the GPU hardware. The tool is provided as a download package that must be extracted to a bootable USB flash drive. After booting the DGX-1 from the USB flash drive, you can run a number of diagnostic tests on all the GPUs in the system.

Obtaining the DGX-1 GPU Diagnostics Tool

Obtain the DGX-1 GPU Diagnostics Tool from NVIDIA Support Enterprise Services.
  1. Log on to the NVIDIA Enterprise Services site.
  2. Click the Announcements tab to locate the download links for the DGX-1 GPU Diagnostics archive file.
  3. Download the archive file and save it to your local disk.

Creating a Bootable USB Flash Drive Under Windows

After obtaining the archive file that contains the GPU Diagnostics Tool from NVIDIA Support Enterprise Services, create a bootable USB flash drive that contains the diagnostics software.

On a Windows system, you can use the Akeo Reliable USB Formatting Utility (Rufus) to create a bootable USB flash drive that contains the DGX-1 GPU Diagnostics.

  1. Ensure that the following prerequisites are met:
    • The correct DGX-1 GPU Diagnostic archive is saved to your local disk. For more information, see Obtaining the DGX-1 GPU Diagnostics Tool .
    • You have a USB flash drive with a minimum capacity of 4 GB. All existing data on the drive will be deleted when creating the bootable drive.
  2. Plug the USB flash drive into one of the USB ports of your Windows system.
  3. Download and launch the Akeo Reliable USB Formatting Utility (Rufus).

  4. Under Partition scheme and target system type, select MBT partition scheme for BIOS or UEFI.
  5. Under File System, select FAT32.
  6. Click the Format Options arrow to view advanced options.
  7. Make sure the Create a bootable disk using checkbox is checked, then click the list arrow and select Syslinux 6.03.
  8. Click Start.
  9. When the formatting utility is finished, unzip the GPU Diagnostics archive file onto the USB flash drive. From Windows Explorer, right-click the archive file, select Extract All and then enter the drive letter corresponding to the USB flash drive.

Running the DGX-1 GPU Diagnostics

These instructions describe how to run the DGX-1 GPU Diagnostic Tool from a USB flash drive.

Before running the diagnostics from a USB flash drive, ensure that you have a bootable USB flash drive that contains the current DGX-1 GPU Diagnostics Tool.

  1. Plug the USB flash drive containing the GPU Diagnostics into the DGX-1.
  2. Connect a monitor and keyboard directly into the DGX-1.
  3. Boot the system, and when the NVIDIA logo appears, press F11 to get to the boot menu.
  4. At the boot menu, select the USB volume name that corresponds to the inserted USB flash drive, and boot the system from it.
  5. Unpack the GPU Diagnostics from the command line.
    $ tar xzf 629.367.67.4.tgz
    After unpacking, the diagnostics package includes the raw executable file fieldiag to be used along with the appropriate arguments.
  6. Run the complete diagnostics by entering the following on the command line according to the type of coverage you need.
    • Standard coverage (test runs at the GPUs' maximum performance level; two-hour approximate runtime)
      $ ./multifieldiag p0only
    • Enhanced coverage (test runs at several GPU performance levels; five-hour approximate runtime)
      $ ./multifieldiag
    LEDs on the keyboard will blink while the test runs.
  7. After the GPU Diagnostics test finishes, obtain the log file fieldiag_xxxx.log from the \home folder on the USB flash drive and send it as an attachment when you submit a service request through the Enterprise Services portal. The log file is not human-readable, so you must obtain the log file and send it to NVIDIA for evaluation.
  8. Reboot the DGX-1. You must reboot the system in order to restore drivers and softare modules that were removed as part of the diagnostics script.
Complete documentation for the GPU Diagnostics is available in the file NV_Field_Diag_Software.pdf, which you can find among the unpacked files on the \home directory. The document includes descriptions of optional command line arguments you can use with the test. Note that some arguments are not available with the DGX-1.

5.3. Restoring the DGX-1 Software Image

If the DGX-1 software image becomes corrupted or the OS SSD was replaced after a failure, restore the DGX-1 software image to its original factory condition from a pristine copy of the image.

The process for restoring the DGX-1 software image is as follows:

  1. Obtain an ISO file that contains the image from NVIDIA Support Enterprise Services as explained in Obtaining the DGX-1 Software ISO Image and Checksum File.
  2. Restore the DGX-1 software image from this file either remotely through the BMC or locally from a bootable USB flash drive.

Obtaining the DGX-1 Software ISO Image and Checksum File

To ensure that you restore the current version of the DGX-1 software image, obtain the correct ISO image file from NVIDIA Support Enterprise Services. A checksum file is provided for the image to enable you to verify the bootable installation medium that you create from the image file.
  1. Log on to the NVIDIA Enterprise Services site.
  2. Click the Announcements tab to locate the download links for the DGX-1 software image.
  3. Download the ISO image and its checksum file and save them to your local disk. The ISO image is also available in an archive file. If you download the archive file, be sure to extract the ISO image before proceeding.

Re-Imaging the System Remotely

These instructions describe how to re-image the system remotely through the BMC. For information about how to restore the system locally, see Re-Imaging the System From a USB Flash Drive.

Before re-imaging the system remotely, ensure that the correct DGX-1 software image is saved to your local disk. For more information, see Obtaining the DGX-1 Software ISO Image and Checksum File.
  1. Connect to the BMC and change user privileges.
    1. Open a Java-enabled web browser within your LAN and go to http://IPMI-IP-address/, then log in.

      Use Firefox or Internet Explorer. Google Chrome is not officially supported by the BMC.

    2. From the top menu, click Configuration and then select User Management.
    3. Select the user name that you created for the BMC, then click Modify User.
    4. In the Modify User dialog, select the VMedia check box to add it to the extended privileges for the user, then click Modify.



  2. Set up the ISO image as virtual media.
    1. From the top menu, click Remote Control and select Console Redirection.



    2. Click Java Console to open the remote JViewer window. Make sure pop-up blockers are disabled for this site.
    3. From the JViewer top menu bar, click Media and then select Virtual Media Wizard.





    4. From the CD/DVD Media: I section of the Virtual Media dialog, click Browse and then locate the re-image ISO file and click Open.

      You can ignore the device redirection warning at the bottom of the Virtual Media wizard as it does not affect the ability to re-image the system.

    5. Click Connect CD/DVD, then click OK at the Information dialog.

      The Virtual Media window shows that the ISO image is connected.





    6. Close the window.

      The CD ROM icon in the menu bar turns green to indicate that the ISO image is attached.





  3. Reboot, install the image, and complete the DGX-1 setup.
    1. From the top menu, click Power and then select Reset Server.



    2. Click Yes and then OK at the Power Control dialogs, then wait for the system to power down and then come back online.
    3. At the boot selection screen, select Install DGX-1 OS and then press [Enter].

      The DGX-1 will reboot from CDROM0 1.00, and proceed to install the image. This can take approximately 15 minutes.

      Note: The Mellanox InfiniBand driver installation may take up to 10 minutes.

      After the installation is completed, the system ejects the virtual CD and then reboots into the OS.

Refer to Setting Up the DGX-1 for the steps to take when booting up the DGX-1 for the first time after a fresh installation.

Creating a Bootable Installation Medium

After obtaining an ISO file that contains the software image from NVIDIA Support Enterprise Services, create a bootable installation medium, such as a USB flash drive or DVD-ROM, that contains the image.

Note: If you are restoring the software image remotely through the BMC, you do not need a bootable installation medium and you can omit this task.

Creating a Bootable USB Flash Drive by Using the dd Command

On a Linux system, you can use the dd command to create a bootable USB flash drive that contains the DGX-1 software image.

Note: Because the image is a hybrid ISO image, you must convert and copy the image to perform a device bit copy of the image. You cannot perform a simple file copy of the image.

Ensure that the following prerequisites are met:

  • The correct DGX-1 software image is saved to your local disk. For more information, see Obtaining the DGX-1 Software ISO Image and Checksum File.
  • The USB flash drive meets these requirements:
    • The USB flash drive has a capacity of at least 4 GB.
    • The partition scheme on the USB flash drive is a GPT partition scheme for UEFI.
  1. Plug the USB flash drive into one of the USB ports of your Linux system.
  2. Obtain the device name of the USB flash drive by running the lsblk command.
    lsblk

    You can identify the USB flash drive from its size, which is much smaller than the size of the SSDs in the DGX-1, and from the mount points of any partitions on the drive, which are under /media.

    In the following example, the device name of the USB flash drive is sde.

    ~$ lsblk
    NAME   MAJ:MIN RM   SIZE RO TYPE MOUNTPOINT
    sda      8:0    0   1.8T  0 disk 
    |_sda1   8:1    0   121M  0 part /boot/efi
    |_sda2   8:2    0   1.8T  0 part /
    sdb      8:16   0   1.8T  0 disk 
    |_sdb1   8:17   0   1.8T  0 part 
    sdc      8:32   0   1.8T  0 disk 
    sdd      8:48   0   1.8T  0 disk 
    sde      8:64   1   7.6G  0 disk 
    |_sde1   8:65   1   7.6G  0 part /media/deeplearner/DGXSTATION
    ~$
  3. As root, convert and copy the image to the USB flash drive.
    sudo dd if=path-to-software-image bs=2048 of=usb-drive-device-name
    CAUTION:
    The dd command erases all data on the device that you specify in the of option of the command. To avoid losing data, ensure that you specify the correct path to the USB flash drive.

Creating a Bootable USB Flash Drive by Using Akeo Rufus

On a Windows system, you can use the Akeo Reliable USB Formatting Utility (Rufus) to create a bootable USB flash drive that contains the DGX-1 software image.

Ensure that the following prerequisites are met:

  1. Plug the USB flash drive into one of the USB ports of your Windows system.
  2. Download and launch the Akeo Reliable USB Formatting Utility (Rufus).



  3. Under Partition scheme and target system type, select GPT partition scheme for UEFI.
  4. Select the Create a bootable disk using option and from the dropdown menu, select ISO image.
  5. Click the optical drive icon and open the DGX-1 software ISO image.
  6. Click Start. Because the image is a hybrid ISO file, you are prompted to select whether to write the image in ISO Image (file copy) mode or DD Image (disk image) mode.



  7. Select Write in ISO Image mode and click OK.

Re-Imaging the System From a USB Flash Drive

These instructions describe how to re-image the system from a USB flash drive. For information about how to restore the system remotely, see Re-Imaging the System Remotely.

Before re-imaging the system from a USB flash drive, ensure that you have a bootable USB flash drive that contains the current DGX-1 software image.

  1. Plug the USB flash drive containing the OS image into the DGX-1.
  2. Connect a monitor and keyboard directly to the DGX-1.
  3. Boot the system and press F11 when the NVIDIA logo appears to get to the boot menu.
  4. Select the USB volume name that corresponds to the inserted USB flash drive, and boot the system from it.
  5. When the system boots up, select Install DGX-1 OS on the startup screen and then press Enter.

The DGX-1 will reboot and proceed to install the image. This can take more than 15 minutes.

Note: The Mellanox InfiniBand driver installation may take up to 10 minutes.

After the installation is completed, the system then reboots into the OS.

Refer to Setting Up the DGX-1 for the steps to take when booting up the DGX-1 for the first time after a fresh installation.

Updating the System BIOS

You can update the system BIOS remotely through the BMC. Before updating the system BIOS, the system must be turned off through the BMC according to the instructions in this section.

  1. Obtain the BIOS image.
    1. Log on to NVIDIA Enterprise Services and click the Announcements tab to locate the DGX-1 software image archive.
    2. Download the image archive and then extract the .bin file.
  2. Log on to the BMC and shut down the DGX-1.
    1. Open a Java-enabled web browser within your LAN and go to http:\\<IPMI IP address>\, then log in.

      Use Firefox or Internet Explorer. Google Chrome is not officially supported by the BMC.

    2. From the top menu, click Remote Control and then select Server Power Control.
    3. At the Power Control and Status screen, select the Power Off Server - Orderly Shutdown option, then click Perform Action.

      You can verify that the DGX-1 is shut down by noting that all the Power Control and Status options are grayed out except for the Power On Server option.

  3. Update the system BIOS.
    1. From the top menu, click Firmware Update, select BIOS Update, and then click Enter Update Mode.

    2. Click OK at the Are you sure to enter update mode? dialog.
    3. From the BIOS Upload screen, click Browse at the Select Firmware to Upload step, then navigate the explorer windows to locate the file you downloaded and select it.

    4. Be sure all the check boxes under Select Preserve Configuration are cleared. This ensures that the BIOS reverts to its fail-safe default settings for a reliable update.
    5. Click Upload Firmware to start the process of installing the updated BIOS.

      You are asked to wait while the image is verified.

    6. Click OK at the Proceed? dialog to start the actual upgrade process.

      The BIOS Flash Status screen shows the upgrade progress, which should take a couple of minutes to complete.

      Note: Do not interrupt the upgrade process once it has started.
  4. After the upgrade process has completed, you can use the top menu to turn the system back on.
    1. From the top menu, click Remote Control and then select Server Power Control.
    2. Select the Power On Server option, and then click Perform Action.
  5. To verify that the BIOS was updated with the proper file, press [F2] or [Del] to enter the BIOS setup screen when the system reboots, then compare the Project Version with the update filename.

Updating the BMC

You can update the BMC remotely using the IPMI port, This can be done while the system is powered on and with applications running.

  1. Obtain the BMC image.
    1. Log on to NVIDIA Enterprise Services and click the Announcements tab to locate the DGX-1 software image archive.
    2. Download the image file.
  2. Open a Java-enabled web browser within your LAN and go to http:\\<IPMI IP address>\, then log in to the BMC.

    Use Firefox or Internet Explorer. Google Chrome is not officially supported by the BMC.

  3. If you’re using DHCP and choose not to preserve the network configuration, then obtain the MAC address for the BMC.

    If the BMC is connected to a network via DHCP, the IP address could change after the update. Follow these substeps to obtain the MAC address in order to connect to the BMC after the update, in case the IP address changes. You can skip these steps if a static IP is used.

    1. From the top menu, click Configuration and then select Network.
    2. Note the MAC address.
  4. From the top menu, click Firmware Update and then select Firmware Update from the drop-down menu.
  5. Click Enter Preserve Configuration, then set the IPMI Preserve Status to Preserve and all others to Overwrite.

    Note: Be sure to set IPMI to Preserve in order to preserve your BMC login credentials. If you fail to do this, the BMC username/password will be set to qct.admin/qct.admin. If this happens, then be sure to enter the BMC dashboard and go to Configuration->Users to add a new user account and disable the qct.admin account after updating the BMC.
  6. If necessary, click Firmware Update again from the top menu and then select Firmware Update from the drop-down menu to return to the Firmware Update page.
  7. Click Enter Update Mode, then click OK at the confirmation dialog.

    After entering Update Mode, aborting the operation or even resizing the browser windows will terminate the session and reset the BMC. If this happens, you will need to close and then reopen the browser to initiate a new session. You may need to wait several minutes for the BMC to reset.

  8. At the Upload Firmware prompt, click Browse to locate and select the firmware image file.

    Select the encrypted file (the file with the "_enc" suffix on the file extension), as the BMC requires the firmware image to be encrypted.

  9. Click Upload to transfer the image to the BMC.
  10. At the Select Based Firmware Update prompt. select Full Flash and then click Proceed.

  • When the BMC firmware update is completed, the BMC resets and the remote session terminates.
  • To initiate a new BMC session, close and then reopen the browser.
  • The BMC can take as much as 10 minutes to reset itself. During this time, the BMC will be unresponsive.

Replacing the System and Components

Be sure to familiarize yourself with the NVIDIA Terms & Conditions documents before attempting to perform any modification or repair to the DGX-1. These Terms & Conditions for the DGX-1 can be found through the NVIDIA DGX Systems Support page.

Contact NVIDIA Enterprise Customer support to obtain an RMA number for any system or component that needs to be returned for repair or replacement.

The following components are customer-replaceable:
  • Solid State Drives (SSDs)
  • Power Supplies
  • Fan Modules
  • DIMMs

Return the failed components to NVIDIA. Low-cost items such as power supplies and fans do not need to be returned.

5.6.1. Replacing the System

When returning a DGX-1 under RMA, consider the following points.

SSDs

If necessary, you can remove and keep the SSDs prior to shipping the system back for replacement. If you already received a replacement system and you want to keep the original SSDs, install the new SSDs into the defective system when shipping it back.

Bezel

Be sure to include the bezel when returning the system.

Replacing an SSD

Access the SSDs from the front of the DGX-1. You can hot swap the SSDs as follows:

  1. If not already removed, remove the bezel by grasping the bezel by the side handles and then pulling the bezel straight off the front of the DGX-1.
  2. Locate the SDD that you want to replace, then press the round button at the top edge to release the latch.
  3. Pull the latch down and then out to unseat the SSD assembly.



  4. Continue pulling the SSD assembly to completely remove it from the unit.
  5. Using a phillips screwdriver, remove the four screws attaching the SSD to the hot-swap tray.



  6. Save the screws for the replacement.
  7. Mount the replacement SSD to the hot-swap tray using the four screws. Make sure that the connector is on the open edge side of the tray.
  8. With the round button at the top, insert the assembly into the appropriate bay, then push the assembly all the way in.
  9. Press the latch against the assembly to completely seat the assembly.
  10. Reattach the bezel.

    With the bezel positioned so that the NVIDIA logo is visible from the front and is on the left-hand side, line up the pins near the corners of the DGX-1 with the holes in back of the bezel, then gently press the bezel against the DGX-1. The bezel is held in place magnetically.

Recreating the Virtual Drives

After you have replaced the OS SSD, with or without any of the cache SSDs, you need to recreate the virtual drives and then re-image the system in order to recreate the partitions on all the virtual drives.

The following is an overview of the process:

  1. Clear the drive group configuration
  2. Recreate the OS Virtual Drive
  3. Recreate the Cache Virtual Drive
  4. Re-image the System

These instructions apply only if you have replaced the OS SSD, with or without one or more of the cache SSDs. If you have replaced only one or more of the cache SSDs, and not the OS SSD, then follow the instructions in the section Recreating the RAID 0 Array

Access the BIOS Setup Utility

RAID configuration is accomplished through the BIOS setup utility.

  1. Connect a display and keyboard to the DGX-1.
  2. Turn the DGX-1 on or reboot.
  3. At the NVIDIA logo boot screen, press [F2] or [Del] to enter the BIOS setup screen.

  4. Select the Advanced tab from the top menu and then Scroll down and select the MegaRAID Configuration Utility.

    The RAID Configuration menu appears.

If you replaced the OS drive, follow the instructions in the section Clear the Drive Group Configuration .

Clear the Drive Group Configuration

These instructions apply when you have replaced the OS drive.

  1. Select Main Menu, then select Configuration Management.

  2. Select Clear Configuration.

  3. Select Confirm [Disabled] and then select Enabled at the confirmation popup.

  4. Select Yes, then select OK at the success screen.

  5. Follow the instructions in the sections Recreate the OS Virtual Drive and then Recreate the RAID0 Virtual Drive .

Recreate the OS Virtual Drive

These instructions apply when you have replaced the OS drive. Be sure to first complete the instructions in the section Clear the Drive Group Configuration.

  1. Navigate to the RAID Utility Main Menu, then under Actions, select Configure, then select Configuration Management.

  2. Select Create Virtual Drive, then select Select Drives at the next screen. Leave all other options at their default settings as shown below.

    The list of drives under CHOOSE UNCONFIGURED DRIVES will initially be empty.
  3. To view the available drives, select Select Media Type [HDD], then change to [SSD].

  4. Under CHOOSE UNCONFIGURED DRIVES, select the 446 GB drive, then change to [Enabled] at the pop-up dialog.

  5. Confirm that only the first drive at Drive Port 0 - 3:01:00 displays as [Enabled].

  6. Scroll up and select Apply Changes.

  7. Select OK at the success screen.

    The virtual drive creation page now displays a summary of your selection. The Virtual Drive Size should be approximately 446 GB.
  8. Select Save Configuration at the top of the menu.
  9. Change the Confirm [Disabled] field to [Enabled] and then select [Yes].

  10. Select [OK] at the success screen. You have successfully re-created Virtual Drive 0, where the OS will be installed.
  11. Follow the instructions in the section Recreate the RAID0 Virtual Drive

Recreate the RAID0 Virtual Drive

These instructions apply when you have replaced the OS drive and cleared the drive group configuration.

  1. Navigate to the RAID Utility Main Menu, then under Action, select Configure, then select Configuration Management.

  2. Select Create Virtual Drive.

  3. Scroll to Select RAID Level and switch to [RAID0], if not already set.

  4. Scroll to Select Media Type and switch to [SSD].

  5. Select Select Drives.

  6. Switch all unconfigured 1TB drives to [Enabled].

  7. Select Apply Changes.

  8. Change Confirm to [Enabled], then select Yes.
  9. Select OK at the success screen. The Create Virtual Drive screen displays a summary of your selection.
  10. Verify that the summary matches your selection, then select Save Configuration.

  11. Make sure Confirm is set to [Enabled], then select Yes to confirm the change.

  12. Select OK at the success screen.
  13. Confirm and exit.
    1. Select View Drive Group Properties to confirm the configuration.

    2. Verify that your configuration screen shows that you have two virtual drives with the following properties: Virtual Drive 0 of size 446 GB (or very similar) Virtual Drive 1 of size 7 TB (or very similar).

    3. If your Drive Groups match the above, press [F10] to save these settings and reset the system.
    4. Select Save Changes and Reset, then select Yes at the confirmation prompt.

  14. Follow the instructions in the section Restoring the DGX-1 Software Image to create the partitions.

Recreating the RAID 0 Array

After replacing one of the RAID 0 cache SSDs, you need to recreate the RAID 0 array. If you replaced only the cache and not the operating system SSD, then you can use a convenient script to recreate the RAID array. The script is part of the DGX-1 software as of version 2.0.4.
To use the script, you need to get and install the StorCLI utility. For instructions, see the document Using StorCLI to Recreate the NVIDIA DGX-1 RAID 0 Array, available from the Enterprise Services site.
Note: Connect a display and keyboard to the DGX-1 when booting the DGX-1 before recreating the RAID array. This is because the system may halt at the BIOS screen alerting you that the RAID array needs to be configured. Press C (or whichever key allows you to continue) to complete the boot process. You will be able to do this only if you are operating the DGX-1 through a direct display and keyboard connection.
  1. If you have installed the StorCLI utility, run the script by entering the following on the command line:
     $ sudo python /usr/local/bin/configure_raid_array.py -c -f 
  2. After the script has finished recreating the RAID 0 array, reboot the DGX-1 to verify that /raid is mounted and usable.
Refer to the document Using StorCLI to Recreate the NVIDIA DGX-1 RAID 0 Array for more information.

Replacing the Power Supplies

Access the power supplies from the front of the DGX-1. You can hot-swap the power supplies as follows:

  1. If not already removed, remove the bezel by grasping the bezel by the side handles and then pulling the bezel straight off the front of the DGX-1.
  2. Unplug the power cord from the power connector on the fan assembly.
  3. Flip the power supply handle out.



  4. Push the green release lever to the left and simultaneously use the power supply handle to pull out the power supply.
  5. Slide the replacement power supply into the bay and push until seated.
  6. Flip the power supply handle up against the power supply.
  7. Reconnect the power cord.
  8. Reattach the bezel. With the bezel positioned so that the NVIDIA logo is visible from the front and is on the left-hand side, line up the pins near the corners of the DGX-1 with the holes in back of the bezel, then gently press the bezel against the DGX-1. The bezel is held in place magnetically.

Replacing the Fan Module

CAUTION: To avoid overheating the system, the fan module should be replaced within 25 seconds after removal.

  1. Unscrew the thumbscrews at the front of the DGX-1, then slide the DGX-1 about half way out from the rack.
  2. Squeeze together the latches at the square access openings on the top of the chassis, then flip open the top of the chassis to expose the fan modules.



  3. Squeeze the release tabs on the outer edge of the fan module you want to replace, then pull up to lift the fan module out of the unit.



  4. Replace with a new fan module using the reverse steps.

Replacing the DIMMs

Before attempting to replace any of the dual inline memory modules (DIMMs), make sure that you know the location of the faulty DIMM needing replacment. The location ID is an alpha-numeric designator, such as A0, A1, B0, B1, etc., and is reported in the BMC log files.

CAUTION: Static Sensitive Devices: - Be sure to observe best practices for electrostatic discharge (ESD) protection. This includes making sure personnel and equipment are connected to a common ground, such as by wearing a wrist strap connected to the chassis ground, and placing components on static-free work surfaces.

The DIMMs are located on the motherboard tray, which is accessible from the rear of the DGX-1.

  1. Turn off the DGX-1 and disconnect all network and power cabling.
  2. Remove the motherboard tray.
    1. Locate the locking levers for the motherboard tray at the rear of the DGX-1. There are two sets of locking levers. The locking levers for the motherboard are the bottom set.

    2. Rotate the retention clasps inward towards the center of the unit.

      The retention clasps hold the locking levers in place. Rotating the clasps inward releases the locking levers.
    3. Swing the locking levers out and then use them to pull the motherboard tray out of the unit.

      Do not pull the unit by the blue retention clasps; they may break.
    4. Set the motherboard tray on a clean work surface, and position it so that the locking levers are at the top as you look down on the tray.

      The DIMMs are on a printed circuit board on the left side of the tray.
  3. Using the figure below as a guide, locate the DIMM corresponding to the ID of the faulty DIMM as reported in the BMC log.

  4. Remove the DIMM.
    1. Press down on the side latches at both ends of the DIMM socket to push them away from the DIMM. This should unseat the DIMM from the socket.
    2. Pull the DIMM straight up to remove it from the socket.
  5. Carefully insert the replacement DIMM.
    1. Make sure the socket latches are open.
    2. Positon the DIMM over the socket, making sure that the notch on the DIMM lines up with the key in the slot, then press the DIMM down into the socket until the side latches click in place.
    3. Make sure that the latches are up and locked in place.
  6. Carefully insert the motherboard tray back into the unit, then swing the locking levers flat against the tray and secure them in place with the retention clasps.

Replacing the InfiniBand Cards

The InfiniBand cards are located on the GPU tray which is accessible from the rear of the DGX-1. Be sure you have identified the faulty InfiniBand card needing to be replaced. The slots are identified as indicated in the following image.

CAUTION: Static Sensitive Devices: - Be sure to observe best practices for electrostatic discharge (ESD) protection. This includes making sure personnel and equipment are connected to a common ground, such as by wearing a wrist strap connected to the chassis ground, and placing components on static-free work surfaces.

  1. Turn off the DGX-1 and disconnect all network and power cabling.
  2. Remove the GPU tray.
    1. Locate the locking levers for the GPU tray at the rear of the DGX-1. There are two sets of locking levers. The locking levers for the GPU tray are the top set.

    2. Rotate the retention clasps inward towards the center of the unit.

      The retention clasps hold the locking levers in place. Rotating the clasps inward releases the locking levers.
    3. Swing the locking levers out and then use then to pull the GPU tray out of the unit.

      Do not pull the unit by the blue retention clasps; they may break.
  3. Set the GPU tray on a clean work surface.
  4. At the top edge of the bracket for the InfiniBand card that you want to replace, rotate the retention clasp to free the bracket.

  5. Firmly grasp the InfiniBand card and lift it straight up out of the PCIe slot.

  6. Position the replacement InfiniBand card over the empty PCIe slot and insert it into the slot.
  7. Swing the retention clasp over the bracket to secure the bracket in place.

  8. Carefully insert the GPU tray back into the unit, then swing the locking levers flat against the tray and secure them in place with the retention clasps.
  9. Reconnect all connectors, boot the system, then perform the verification and setup steps described in the next section.

Setting Up the InfiniBand Cards

This section describes the steps needed to verify that the InfiniBand card has been replaced correctly.

  1. With the DGX-1 turned on, verify that the card was installed correctly and is recognized by the system.
    $ lspci | grep -i mellanox
    The output should show all four InfiniBand cards. Example:
     05:00.0 Infiniband controller: Mellanox Technologies MT27700 Family [ConnectX-4]
     0c:00.0 Infiniband controller: Mellanox Technologies MT27700 Family [ConnectX-4]
     84:00.0 Infiniband controller: Mellanox Technologies MT27700 Family [ConnectX-4]
     8b:00.0 Infiniband controller: Mellanox Technologies MT27700 Family [ConnectX-4] 
    If all four cards are not reported, then the card was not installed properly and should be reseated. If a card other than the officially supported Mellanox family of adapters appears, contact NVIDIA Enterprise Support.
  2. Verify that the InfiniBand drivers are present.
    $ lsmod | grep -i ib_
    The output should be a list of lb_ and mlx_ driver components. Example:
     ib_ucm 20480 0
     ib_ipoib 131072 0
     ib_cm 45056 3 rdma_cm,ib_ucm,ib_ipoib
     ib_uverbs 73728 2 ib_ucm,rdma_ucm
     ib_umad 24576 0
     mlx5_ib 192512 0
     mlx4_ib 192512 0
     ib_sa 36864 5 rdma_cm,ib_cm,mlx4_ib,rdma_ucm,ib_ipoib
     ib_mad 57344 4 ib_cm,ib_sa,mlx4_ib,ib_umad
     ib_core 143360 13 rdma_cm,ib_cm,ib_sa,iw_cm,nv_peer_mem,mlx4_ib,mlx5_ib,ib_mad,ib_ucm,ib_umad,ib_uverbs,rdma_ucm,ib_ipoib
     ib_addr 20480 3 rdma_cm,ib_core,rdma_ucm
     ib_netlink 16384 3 rdma_cm,iw_cm,ib_addr
     mlx4_core 344064 2 mlx4_en,mlx4_ib
     mlx5_core 524288 1 mlx5_ib
     mlx_compat 16384 18 rdma_cm,ib_cm,ib_sa,iw_cm,mlx4_en,mlx4_ib,mlx5_ib,ib_mad,ib_ucm,ib_netlink,ib_addr,ib_core,ib_umad,ib_uverbs,mlx4_core,mlx5_core,rdma_ucm,ib_ipoib
  3. Verify that the OFED software was installed correctly.
     $ modinfo mlx5_core | grep -i version | head -1 
    Example output:
    Version : 3.4-1.0.0
    DGX-1 OS release 1.0 should have OFED software 3.2. DGX-1 OS release 2.0 should have OFED software 3.4.
  4. Restart the InfiniBand services so that the new card is recognized.
    1. Restart the InfiniBand service.
      $ sudo service openibd restart
    2. Restart the Service Manager service.
      $ sudo service opensmd restart
    3. Verify that the service has started.
      $ service openibd status
        openibd start/running
      $ service opensmd status
        OpenSM is running...
    4. If the services do not start, verify
      • That the drivers are loaded according to step 3.
      • That the associated cables are connected to the InfiniBand ports.
      • The state of ibstat (refer to step 7)
      • Whether errors are reported in /var/log/syslog.
      If these steps do not indicate a problem and yet the services still do not start, contact NVIDIA Enterprise Support and obtain an RMA for the card.
  5. Verify the firmware version.
    $ cat /sys/class/infiniband/mlx5*/fw_ver
    Example output:
     12.17.1010
     12.17.1010
     12.17.1010
     12.17.1010
    The latest InfiniBand firmware version supported on DGX-1 OS release 1.0 is 12.16.1020, and the latest supported on release 2.0 is 12.17.1010.
  6. If you need to update the firmware, follow these steps:
    1. Initiate the firmware update.
      $ sudo /opt/mellanox/mlnx-fw-updater/mlnx_fw_updater.pl
      The script will check the firmware version of each card and update where needed. If the firmware is updated for any card, you will need to reboot the system for the changes to take effect.
    2. Reboot the system if instructed.
    3. After rebooting the system, verify that all the Mellanox InfiniBand cards are using the current firmware.
      $ cat /sys/class/infiniband/mlx5*/fw_ver
       12.17.1010
       12.17.1010
       12.17.1010
       12.17.1010
  7. Verify the physical port state for the InfiniBand cards.
    $ ibstat
    In the output text, verify that the Physical State for each card with a cable connection is LinkUp and that the port for the card is configured with a GUID. The following example output shows one card in a non-connected state, and three cards in a connected state. Relevant text is highlighted in bold.
    CA 'mlx5_0'
     CA type: MT4115
     Number of ports: 1
     Firmware version: 12.17.1010
     Hardware version: 0
     Node GUID: 0x248a0703000de288
     System image GUID: 0x248a0703000de288
     Port 1:
           State: Down
           Physical state: Polling
           Rate: 10
           Base lid: 65535
           LMC: 0
           SM lid: 0
           Capability mask: 0x2651e848
           Port GUID: 0x248a0703000de288
           Link layer: InfiniBand
    CA 'mlx5_1'
     CA type: MT4115
     Number of ports: 1
     Firmware version: 12.17.1010
     Hardware version: 0
     Node GUID: 0x248a0703000de26c
     System image GUID: 0x248a0703000de26c
     Port 1:
            State: Initializing
            Physical state: LinkUp
            Rate: 100
            Base lid: 65535
            LMC: 0
            SM lid: 0
            Capability mask: 0x2651e848
            Port GUID: 0x248a0703000de26c
            Link layer: InfiniBand
    CA 'mlx5_2'
     CA type: MT4115
     Number of ports: 1
     Firmware version: 12.17.1010
     Hardware version: 0
     Node GUID: 0x248a0703001effde
     System image GUID: 0x248a0703001effde
     Port 1:
            State: Initializing
            Physical state: LinkUp
            Rate: 100
            Base lid: 65535
            LMC: 0
            SM lid: 0
            Capability mask: 0x2651e848
            Port GUID: 0x248a0703001effde
            Link layer: InfiniBand
    CA 'mlx5_3'
     CA type: MT4115
     Number of ports: 1
     Firmware version: 12.17.1010
     Hardware version: 0
     Node GUID: 0x7cfe900300118f22
     System image GUID: 0x7cfe900300118f22
     Port 1:
             State: Initializing
             Physical state: LinkUp
             Rate: 100
             Base lid: 65535
             LMC: 0
             SM lid: 0
             Capability mask: 0x2651e848
             Port GUID: 0x7cfe900300118f22
             Link layer: InfiniBand 

Installing Software on Air-Gapped NVIDIA DGX-1 Systems

For security purposes, some installations require that systems be isolated from the internet or outside networks. Since most DGX-1 software updates are accomplished through an over-the-network process with NVIDIA servers, this section explains how updates can be made when using an over-the-network method is not an option. It includes a process for installing Docker containers as well.

Installing NVIDIA DGX-1 Software

One method for updating DGX-1 software on an air-gapped DGX-1 is to download the ISO image, copy it to removable media and then re-image the DGX-1 from the media. This method is available only for software versions that are available as ISO images for download.

Alternately, you can update the DGX-1 software by performing a network update from a local repository. This method is available only for software versions that are available for over-the-network updates.

Re-Imaging the System

Note:WARNING:This process destroys all data and software customizations that you have made on the DGX-1. Be sure to back up any data that you want to preserve, and push any Docker images that you want to keep to a trusted registry.
  1. Obtain the ISO image from the Enterprise Support site.
    1. Log on to the NVIDIA Enterprise Services site and click the Announcements tab to locate the DGX-1 Base OS Image ISO file.
    2. Download the image ISO file.
  2. Refer to the instructions in the Restoring the DGX-1 Software Image section for additional instructions.

Creating a Local Mirror of the NVIDIA and Canonical Repositories

Instructions for setting up a private repository or mirroring the NVIDIA and Canonical repositories are beyond the scope of this document. It is expected that users are knowledgeable about those processes.

  1. Create a private repository that mirrors the NVIDIA as well as the Canonical repositories.

    Consult /etc/apt/sources.list and the contents of /etc/apt.sources.list.d/ on your running DGX-1 for the repository locations.

  2. Modify /etc/apt/sources.list and appropriate contents of /etc/apt.sources.list.d/ to point to your private repository.
  3. Perform the update from the private repository, starting with Get the new package list step (sudo apt-get update) of the instructions found in the DGX-1 Software Release Notes and Upgrade Guide, which you can obtain from the Enterprise Support site.

Installing Docker Containers

This method applies to Docker containers hosted on the NVIDIA DGX Container Registry, and requires that you have an active DGX Cloud Services account.

  1. On a system with internet access, log in to the DGX Container Registry by entering the following command and credentials.
    $ docker login nvcr.io
    Username: $oauthtoken 
    Password: apikey

    Type “$oauthtoken” exactly as shown for the Username. This is a special username that enables API key authentication. In place of apikey, paste in the API Key text that you obtained from the Cloud Services website.

  2. Enter the docker pull command, specifying the image registry, image repository, and tag:
    $ docker pull nvcr.io/nvidia/repository:tag
  3. Verify the image is on your system using docker images.
    $ docker images
  4. Save the Docker image as an archive. .
    $ docker save nvcr.io/nvidia/repository:tag > framework.tar
  5. Transfer the image to the air-gapped system using removable media such as a USB flash drive.
  6. Load the NVIDIA Docker image.
     $ docker load –i framework.tar
  7. Verify the image is on your system.
    $ docker images

7. Safety

To reduce the risk of bodily injury, electrical shock, fire, and equipment damage, read this document and observe all warnings and precautions in this guide before installing or maintaining your server product.

In the event of a conflict between the information in this document and information provided with the product or on the website for a particular product, the product documentation takes precedence.

Your server should be integrated and serviced only by technically qualified persons.

You must adhere to the guidelines in this guide and the assembly instructions in your server manuals to ensure and maintain compliance with existing product certifications and approvals. Use only the described, regulated components specified in this guide. Use of other products I components will void the UL Listing and other regulatory approvals of the product, and may result in noncompliance with product regulations in the region(s) in which the product is sold.

Safety Warnings and Cautions

To avoid personal injury or property damage, before you begin installing the product, read, observe, and adhere to all of the following safety instructions and information. The following safety symbols may be used throughout the documentation and may be marked on the product and I or the product packaging.
Symbol Meaning
CAUTION Indicates the presence of a hazard that may cause minor personal injury or property damage if the CAUTION is ignored.
WARNING Indicates the presence of a hazard that may result in serious personal injury if the WARNING is ignored.
Indicates potential hazard if indicated information is ignored.
Indicates shock hazards that result in serious injury or death if safety instructions are not followed
Indicates hot components or surfaces.
Indicates do not touch fan blades, may result in injury.
Indicates to unplug all AC power cord(s) to disconnect AC power.
Recycle the battery.
The rail racks are designed to carry only the weight of the server system. Do not use rail-mounted equipment as a workspace. Do not place additional load onto any rail-mounted equipment.
Indicates two people are required to safely handle the system

Intended Application Uses

This product was evaluated as Information Technology Equipment (ITE), which may be installed in offices, schools, computer rooms, and similar commercial type locations.The suitability of this product for other product categories and environments (such as medical, industrial, residential, alarm systems, and test equipment), other than an ITE application, may require further evaluation.

Site Selection

Choose a site that is:
  • Clean, dry, and free of airborne particles (other than normal room dust).
  • Well-ventilated and away from sources of heat including direct sunlight and radiators.
  • Away from sources of vibration or physical shock.
  • In regions that are susceptible to electrical storms, we recommend you plug your system into a surge suppressor and disconnect telecommunication lines to your modem during an electrical storm.
  • Provided with a properly grounded wall outlet.
  • Provided with sufficient space to access the power supply cord(s), because they serve as the product's main power disconnect.

Equipment Handling Practices

Reduce the risk of personal injury or equipment damage:
  • Conform to local occupational health and safety requirements when moving and lifting equipment.
  • Use mechanical assistance or other suitable assistance when moving and lifting equipment .

Electrical Precautions

Power and Electrical Warnings

Caution: The power button, indicated by the stand-by power marking, DOES NOT completely tum off the system AC power, SV standby power is active whenever the system is plugged in. To remove power from system, you must unplug the AC power cord from the wall outlet. Your system may use more than one AC power cord. Make sure all AC power cords are unplugged. Make sure the AC power cord(s) is/are unplugged before you open the chassis, or add or remove any non hot-plug components.

Do not attempt to modify or use an AC power cord if it is not the exact type required. A separate AC cord is required for each system power supply.

Some power supplies in servers use Neutral Pole Fusing. To avoid risk of shock use caution when working with power supplies that use Neutral Pole Fusing.

The power supply in this product contains no user-serviceable parts. Do not open the power supply. Hazardous voltage, current and energy levels are present inside the power supply. Return to manufacturer for servicing.

When replacing a hot-plug power supply, unplug the power cord to the power supply being replaced before removing it from the server.

To avoid risk of electric shock, tum off the server and disconnect the power cord, telecommunications systems, networks, and modems attached to the server before opening it.

Power Cord Warnings

Use certified AC power cords to connect to both the power distribution unit (PDU) and server system installed in your rack.

Do not attempt to modify or use the AC power cord(s) if they are not the exact type required to fit into the grounded electrical outlets.

Caution: To avoid electrical shock or fire, check the power cord(s) that will be used with the product as follows:

  • The power cord must have an electrical rating that is greater than that of the electrical current rating marked on the product.
  • The power cord must have safety ground pin or contact that is suitable for the electrical outlet.
  • The power supply cord(s) is/are the main disconnect device to AC power.
  • The socket outlet(s) must be near the equipment and readily accessible for disconnection.
  • The power supply cord(s) must be plugged into socket-outlet(s) that is /are provided with a suitable earth ground.

System Access Warnings

Caution: To avoid personal injury or property damage, the following safety instructions apply whenever accessing the inside of the product:

  • Turn off all peripheral devices connected to this product.
  • Turn off the system by pressing the power button to off.
  • Disconnect the AC power by unplugging all AC power cords from the system or wall outlet.
  • Disconnect all cables and telecommunication lines that are connected to the system.
  • Retain all screws or other fasteners when removing access cover(s). Upon completion of accessing inside the product, refasten access cover with original screws or fasteners.
  • Do not access the inside of the power supply. There are no serviceable parts in the power supply. Return to manufacturer for servicing.
  • Power down the server and disconnect all power cords before adding or replacing any non hot-plug component.
  • When replacing a hot-plug power supply, unplug the power cord to the power supply being replaced before removing the power supply from the server.

Caution: If the server has been running, any installed processor(s) and heat sink(s) may be hot.

Unless you are adding or removing a hot-plug component, allow the system to cool before opening the covers. To avoid the possibility of coming into contact with hot component(s) during a hot-plug installation, be careful when removing or installing the hot-plug component(s).

Caution: To avoid injury do not contact moving fan blades. Your system is supplied with a guard over the fan, do not operate the system without the fan guard in place.

Rack Mount Warnings

Note: The following installation guidelines are required by UL for maintaining safety compliance when installing your system into a rack.

The equipment rack must be anchored to an unmovable support to prevent it from tipping when a server or piece of equipment is extended from it. The equipment rack must be installed according to the rack manufacturer's instructions.

Install equipment in the rack from the bottom up with the heaviest equipment at the bottom of the rack.

Extend only one piece of equipment from the rack at a time.

You are responsible for installing a main power disconnect for the entire rack unit. This main disconnect must be readily accessible, and it must be labeled as controlling power to the entire unit, not just to the server(s).

To avoid risk of potential electric shock, a proper safety ground must be implemented for the rack and each piece of equipment installed in it.

Elevated Operating Ambient- If installed in a closed or multi-unit rack assembly, the operating ambient temperature of the rack environment may be greater than room ambient. Therefore, consideration should be given to installing the equipment in an environment compatible with the maximum ambient temperature (Tma) specified by the manufacturer.

Reduced Air Flow -Installation of the equipment in a rack should be such that the amount of air flow required for safe operation of the equipment is not compromised.

Mechanical Loading- Mounting of the equipment in the rack should be such that a hazardous condition is not achieved due to uneven mechanical loading.

Circuit Overloading- Consideration should be given to the connection of the equipment to the supply circuit and the effect that overloading of the circuits might have on overcurrent protection and supply wiring. Appropriate consideration of equipment nameplate ratings should be used when addressing this concern.

Reliable Earthing- Reliable earthing of rack-mounted equipment should be maintained.

Particular attention should be given to supply connections other than direct connections to the branch circuit (e.g. use of power strips).Topic paragraph

Electrostatic Discharge

Caution: ESD can damage drives, boards, and other parts. We recommend that you perform all procedures at an ESD workstation. If one is not available, provide some ESD protection by wearing an antistatic wrist strap attached to chassis ground -- any unpainted metal surface -- on your server when handling parts.

Always handle boards carefully. They can be extremely sensitive to ESO. Hold boards only by their edges. After removing a board from its protective wrapper or from the server, place the board component side up on a grounded, static free surface. Use a conductive foam pad if available but not the board wrapper. Do not slide board over any surface.

Other Hazards

PROPOSITION 65 WARNING

This product contains chemicals known to the State of California to cause cancer and birth defects or other reproductive harm.

CALIFORNIA DEPARTMENT OF TOXIC SUBSTANCES CONTROL

Perchlorate Material – special handling may apply. See www.dtsc.ca.gov/hazardouswaste/perchlorate.

Perchlorate Material: Lithium battery (CR2032) contains perchlorate. Please follow instructions for disposal.

NICKEL

NVIDIA Bezel: The bezel’s decorative metal foam contains some nickel. The metal foam is not intended for direct and prolonged skin contact. Please use the handles to remove, attach or carry the bezel. While nickel exposure is unlikely to be a problem, you should be aware of the possibility in case you’re susceptible to nickel-related reactions.

Battery Replacement

Caution: There is the danger of explosion if the battery is incorrectly replaced. When replacing the battery, use only the battery recommended by the equipment manufacturer.

Dispose of batteries according to local ordinances and regulations. Do not attempt to recharge a battery.

Do not attempt to disassemble, puncture, or otherwise damage a battery.

Cooling and Airflow

Caution: Carefully route cables as directed to minimize airflow blockage and cooling problems. For proper cooling and airflow, operate the system only with the chassis covers installed. Operating the system without the covers in place can damage system parts. To install the covers:

  • Check first to make sure you have not left loose tools or parts inside the system.
  • Check that cables, add-in cards, and other components are properly installed.
  • Attach the covers to the chassis according to the product instructions.

8. Compliance

The NVIDIA DGX-1 is compliant with the regulations listed in this section.

United States

Federal Communications Commission (FCC)

FCC Marking (Class A)

This device complies with part 15 of the FCC Rules. Operation is subject to the following two conditions: (1) this device may not cause harmful interference, and (2) this device must accept any interference received, including any interference that may cause undesired operation of the device.

NOTE: This equipment has been tested and found to comply with the limits for a Class A digital device, pursuant to part 15 of the FCC Rules. These limits are designed to provide reasonable protection against harmful interference when the equipment is operated in a commercial environment. This equipment generates, uses, and can radiate radio frequency energy and, if not installed and used in accordance with the instruction manual, may cause harmful interference to radio communications. Operation of this equipment in a residential area is likely to cause harmful interference in which case the user will be required to correct the interference at his own expense.

United States / Canada

cULus Listing Mark

Canada

Industry Canada (IC)

CAN ICES-3(A)/NMB-3(A)

The Class A digital apparatus meets all requirements of the Canadian Interference-Causing Equipment Regulation.

Cet appareil numerique de la class A respecte toutes les exigences du Reglement sur le materiel brouilleur du Canada.

CE

European Conformity; Conformité Européenne (CE)

This is a Class A product. In a domestic environment this product may cause radio frequency interference in which case the user may be required to take adequate measures.

The product has been marked with the CE Mark to illustrate its compliance.

This device complies with the following Directives:

  • EMC Directive (2014/30/EU) for Class A, I.T.E equipment.
  • Low Voltage Directive (2014/35/EU) for electrical safety.
  • RoHS Directive (2011/65/EU) for hazardous substances.
  • ErP Directive (2009/125/EC) for European Ecodesign.

A copy of the Declaration of Conformity to the essential requirements may be obtained directly from NVIDIA GmbH (Floessergasse 2, 81369 Munich, Germany).

Japan

VCCI

This is a Class A product.

In a domestic environment this product may cause radio interference, in which case the user may be required to take corrective actions. VCCI-A

China

RoHS Material Content

South Korea

KC

Class A Equipment (Industrial Broadcasting & Communication Equipment). This equipment Industrial (Class A) electromagnetic wave suitability equipment and seller or user should take notice of it, and this equipment is to be used in the places except for home.

Notices

Notice

THE INFORMATION IN THIS GUIDE AND ALL OTHER INFORMATION CONTAINED IN NVIDIA DOCUMENTATION REFERENCED IN THIS GUIDE IS PROVIDED “AS IS.” NVIDIA MAKES NO WARRANTIES, EXPRESSED, IMPLIED, STATUTORY, OR OTHERWISE WITH RESPECT TO THE INFORMATION FOR THE PRODUCT, AND EXPRESSLY DISCLAIMS ALL IMPLIED WARRANTIES OF NONINFRINGEMENT, MERCHANTABILITY, AND FITNESS FOR A PARTICULAR PURPOSE. Notwithstanding any damages that customer might incur for any reason whatsoever, NVIDIA’s aggregate and cumulative liability towards customer for the product described in this guide shall be limited in accordance with the NVIDIA terms and conditions of sale for the product.

THE NVIDIA PRODUCT DESCRIBED IN THIS GUIDE IS NOT FAULT TOLERANT AND IS NOT DESIGNED, MANUFACTURED OR INTENDED FOR USE IN CONNECTION WITH THE DESIGN, CONSTRUCTION, MAINTENANCE, AND/OR OPERATION OF ANY SYSTEM WHERE THE USE OR A FAILURE OF SUCH SYSTEM COULD RESULT IN A SITUATION THAT THREATENS THE SAFETY OF HUMAN LIFE OR SEVERE PHYSICAL HARM OR PROPERTY DAMAGE (INCLUDING, FOR EXAMPLE, USE IN CONNECTION WITH ANY NUCLEAR, AVIONICS, LIFE SUPPORT OR OTHER LIFE CRITICAL APPLICATION). NVIDIA EXPRESSLY DISCLAIMS ANY EXPRESS OR IMPLIED WARRANTY OF FITNESS FOR SUCH HIGH RISK USES. NVIDIA SHALL NOT BE LIABLE TO CUSTOMER OR ANY THIRD PARTY, IN WHOLE OR IN PART, FOR ANY CLAIMS OR DAMAGES ARISING FROM SUCH HIGH RISK USES.

NVIDIA makes no representation or warranty that the product described in this guide will be suitable for any specified use without further testing or modification. Testing of all parameters of each product is not necessarily performed by NVIDIA. It is customer’s sole responsibility to ensure the product is suitable and fit for the application planned by customer and to do the necessary testing for the application in order to avoid a default of the application or the product. Weaknesses in customer’s product designs may affect the quality and reliability of the NVIDIA product and may result in additional or different conditions and/or requirements beyond those contained in this guide. NVIDIA does not accept any liability related to any default, damage, costs or problem which may be based on or attributable to: (i) the use of the NVIDIA product in any manner that is contrary to this guide, or (ii) customer product designs.

Other than the right for customer to use the information in this guide with the product, no other license, either expressed or implied, is hereby granted by NVIDIA under this guide. Reproduction of information in this guide is permissible only if reproduction is approved by NVIDIA in writing, is reproduced without alteration, and is accompanied by all associated conditions, limitations, and notices.

Trademarks

NVIDIA, the NVIDIA logo, DGX, DGX-1, and DGX Station are trademarks and/or registered trademarks of NVIDIA Corporation in the Unites States and other countries. Other company and product names may be trademarks of the respective companies with which they are associated.