Introduction

1.1 NVIDIA Mission Control: Overview And Licensing

The NVIDIA Mission Control features are extended features that are only available in the NVIDIA Mission Control edition of NVIDIA Base Command Manager (BCM). NVIDIA Mission Control features are targeted at NVIDIA B200, GB200, and other OEM platforms. The NVIDIA sales team must be contacted in case an NVIDIA Mission Control edition is required.

The NVIDIA Mission Control features of the NVIDIA Mission Control edition of BCM can be enabled on a BCM cluster with an active BCM license. Activation of the BCM license is described in Chapter 4 of the Installation Manual of BCM

1.2 NVIDIA Mission Control Features

The NVIDIA Mission Control edition is capable of running the following:

• NVIDIA-conforming BMS: a Building Management System software that aids in cluster management in the building.

• Leak detection: monitoring for leak detection in the fluid-cooled processors. Associated with this are the possible automated actions taken.

• Autonomous hardware recovery: Automates some hardware management operations in order to keep the cluster up.

• Autonomous job recovery: integrates a suite of autonomous software services to enhance job submission, restart procedures, carry out telemetry collection, and carry out anomaly resolution processes.

• NVLink Management Software (NetQ): Integrated with NMX, to manage Ethernet and NVLink.

• IMEX (Internode Memory Exchange Service): a secure service that facilitates the mapping of GPU memory over NVLink between the GPUs in an NVLink domain. IMEX configuration is described in section 7.5.1.4 of the Administrator Manual as part of workload management configuration.

• GB200 rack management: Automated management of rack power up and power down sequence.

• GB200 firmware management: GB200 firmware and BIOS management.

• Power reservation steering: Predicting power consumption, and capping the power to active components based on that prediction, to reduce data center energy costs.

• Run:ai: Kubernetes-based workload management and orchestration using Run:ai.

The point release version (section 9.1.2 of the Administrator Manual) of BCM, and the Kubernetes version in use (kubeadm version -o short), must both be compatible with the NVIDIA Mission Control version that is used.

Details on compatibility for the various versions of NVIDIA Mission Control can be found online at:

https://docs.nvidia.com/mission-control/docs/nmc-software-installation-guide

1.3 Checking That The NVIDIA Mission Control Edition Is Running

Whether the NVIDIA Mission Control edition is running on the cluster can be checked:

• in cmsh:

  • using the licenseinfo command within cmsh

    Example

    root@basecm11:~# cmsh -c 'main licenseinfo | egrep Version\|Edition'
    Version                              10 and above
    Edition                              NVIDIA Mission Control

    or

  • using the verify-license utility from a command line:

    Example

    root@basecm11:~# verify-license info
    ========= Certificate Information ========
    Version:                    10
    Edition:                    NVIDIA Mission Control
    ...

    or

  • using the editiondisabledfeatures command.

    • If the NVIDIA Mission Control edition is running, then the features are all available, as indicated by:

      Example

      root@basecm11:~# cmsh -c 'main editiondisabledfeatures'
      No disabled features

    • If the NVIDIA Mission Control edition is not running, then the features are not available, as indicated by:

      root@basecm11:~# cmsh -c 'main editiondisabledfeatures'
      LeakDetection
      DisplayRack
      RackPower
      AutonomousHardwareRecovery
      AutonomousJobRecovery
      Firmware.flash:GB200
      BMS
      Run:ai
      IMEX
      NMX
      PRS

• in Base View the status of the NVIDIA Mission Control features can be seen in the navigation paths:

  Cluster > License Information > Mission Control

  or

  Cluster > Overview > License Information

1.4 NVIDIA Mission Control Installation Wizards

Installation of NVIDIA Mission Control features can be carried out with:

• a cm-mission-control-setup TUI wizard, which comes with the cm-setup package

• a Base View GUI wizard, if available.

Starting with BCM release version 11.32, the cm-mission-control-setup wizard has some of its components decoupled, so that they are now plugins. This allows minor BCM releases to come out more frequently, since developing NVIDIA Mission Control plugin components can be carried out more independently from the BCM release schedule, compared with developing a monolithic cm-mission-control-setup.

When plugins are selected, the decoupled version automatically installs a new system package, cm-config-apt-nmc. The package lets the cluster administrator select updated plugins for installation on all BCM nodes from the NMC package repository (edge.urm.nvidia.com/artifactory/sw-dgxc-nmc-cm-setup-plugins-debian-local). By default, updating is skipped, unless explicitly chosen, when the wizard is run.

The plugins are implemented as TUI plugins, and there are additionally some GUI (Base View) plugins which run the TUI version in their backend.

Decoupled plugin components at the time of writing (February 2026) are:

• NetQ (Chapter 2) (TUI only)

• AJR (Chapter 5) (TUI and Base View)

• AHR (Chapter 6) (TUI and Base View)

• DPS (Chapter 11) (TUI only)

The plugins are packages containing Python scripts:

• The TUI plugin package name format is:

  cm-setup- <component>-bcm<major version> (for example: cm-setup-netq-bcm11)

• The Base View plugin package names format is:

  cm-base-view- <component>-bcm<major version> (for example: cm-base-view-netq-bcm11)

Download this manual as a PDF