Introduction#

This installation guide provides step-by-step instructions for deploying and configuring NVIDIA Mission Control software on NVIDIA DGX B200/B300 and GB200/GB300 NVL72 systems, with early access support for Vera Rubin NVL72. It covers the following:

  • Installation and configuration of all software components required to enable full NVIDIA Mission Control functionality

  • Software dependencies for each feature

  • Installation and deployment of the features themselves

  • Verification and testing procedures to ensure proper feature functionality

For control plane hardware requirements, refer to the NVIDIA Mission Control Hardware Requirements Guide.

The following figures show the NVIDIA Mission Control 2.5 architecture. Feature availability varies by platform. Refer to NVIDIA Mission Control Feature Support before selecting components for your deployment.

NVIDIA Mission Control 2.5 architecture for DGX B200/B300, with x86 admin and user control planes, DGX compute nodes, and supporting systems

Figure 1 NVIDIA Mission Control 2.5 Software Architecture for DGX B200/B300#

NVIDIA Mission Control 2.5 architecture for GB200/GB300 and early access Vera Rubin NVL72, with admin and user control planes, compute trays, and supporting systems

Figure 2 NVIDIA Mission Control 2.5 Software Architecture for GB200/GB300 and Vera Rubin NVL72#

Vera Rubin NVL72 support is early access. Although the shared NVL72 diagram includes Run:ai and NetQ, these components are not supported on Vera Rubin NVL72 in this release.

Assumptions and Prerequisites#

Before installing the NVIDIA Mission Control software, you must complete the following tasks within BCM 11, as outlined in the NVIDIA Mission Control Management Plane and Rack Setup Installation Guide:

  • All networks are defined and all switches (in-band and out-of-band) are “Up”

  • All control plane nodes are configured and “Up”:

    • slogin

    • K8s-admin (Admin Kubernetes nodes)

    • K8s-user (User Kubernetes nodes)

  • High Availability (HA) setup is configured and failover is verified.

  • For GB200/GB300 systems, the GB200 NVL72 rack(s) setup is complete and the following items are verified:

    • All NVLink switch chips are online with NMX-C/T enabled

    • NVLink switch leader is assigned

    • Each GB200/GB300 NVL72 rack contains 9 NVLink switch trays

    • All 18 compute trays per rack are provisioned and in an “Up” state

    • Power control is established at rack, compute tray, and NVLink switch tray levels

  • Shared storage meets Shared Storage Requirements.

  • A valid BCM license with Mission Control enabled is installed

Mission Control Components – DGX B200/B300#

The DGX B200/B300 architecture uses x86-based control plane nodes. The number of DGX nodes per scaling unit depends on the system model and network fabric, as shown in the diagram.

Control Plane#

The B200/B300 architecture separates management responsibilities across four sets of x86 control-plane node groups.

BCM Head Nodes – x86 (×2, Cluster deployment, management, and monitoring)

Runs Base Command Manager (BCM) providing:

  • GUI (Base Command View), CLI (CMSH), and API interfaces

  • OS provisioning

  • Observability

  • Firmware (FW) Update

  • Network provisioning

  • Health Check

  • Rack and inventory management

  • Power Profiles

  • Leak monitoring and control

  • Slurm Workflow Software

Admin Kubernetes Nodes – x86 (×3, K8s and BCM-integrated services)

  • BCM-Integrated Services:

    • NVIDIA Mission Control-autonomous hardware recovery

    • Observability Stack

    • NVIDIA Mission Control-autonomous job recovery

  • Common Kubernetes services, including Loki, Prometheus, Keycloak, and OpenBao

  • BCM-Provisioned K8s

Slurm Nodes – x86 (×2, User access to Slurm cluster)

  • BCM-Provisioned Slurm

  • Slurm Submission Software

User Kubernetes Nodes - x86 (×3, Run:ai and other user K8s workloads)

  • Run:ai

  • Common Kubernetes services, including GPU Operator, DRA, and Network Operator

  • BCM-Provisioned K8s

Compute Plane#

DGX B200/B300 Scaling Unit (SU)

Each DGX B200/B300 node runs as either a Slurm worker or a K8s + Run:ai worker, depending on its assigned role. Each node contains 8 GPUs.

Network and Storage Systems#

  • InfiniBand (IB) Switches and UFM

  • Ethernet Switches

  • NFS Storage

  • High-Speed Storage

  • RCP Plugin for Spectrum-X (experimental)

Note

Spectrum-X support is experimental. Although the diagram includes Dynamic Power Software, version 0.9.0 does not support DGX B200/B300 because topology activation fails during power policy setup.

Mission Control Components - GB200/GB300 and Vera Rubin NVL72#

GB200/GB300 and Vera Rubin NVL72 systems use 8 racks per scaling unit. DGX deployments require Arm-based BCM head nodes and Slurm nodes, with exceptions permitted for x86 BCM head nodes. OEM deployments recommend Arm-based Slurm nodes and can use x86 for the remaining control plane nodes. The diagram shows the shared layout and labels control plane nodes as Arm/x86. Refer to NVIDIA Mission Control Feature Support for the requirements that apply to your platform.

Admin Control Plane#

BCM Head Nodes (×2, Cluster deployment, management, and monitoring)

Runs Base Command Manager (BCM) providing:

  • GUI (Base Command View), CLI (CMSH), and API interfaces

  • OS provisioning

  • Observability

  • Firmware (FW) Update

  • Network provisioning

  • Health Check

  • Rack and inventory management

  • Power Profiles

  • Leak monitoring and control

  • Slurm Workflow Software

Admin Service Nodes (×3, K8s and BCM-integrated services)

  • BCM-Integrated Services:

    • NetQ (GB200/GB300 only)

    • Observability Stack

    • Dynamic Power Software (DPS), available separately through a scoped beta engagement

    • NVIDIA Mission Control-autonomous hardware recovery

    • NVIDIA Mission Control-autonomous job recovery

  • Common Kubernetes services, including Loki, Prometheus, Keycloak, and OpenBao

  • BCM-Provisioned K8s

User Control Plane#

Slurm Nodes (×2, User access to Slurm cluster)

  • BCM-Provisioned Slurm

  • Slurm Submission Software

User Service Nodes (×3, user K8s workloads)

  • Run:ai (GB200/GB300 only):

    • Control Plane

    • Scheduler

  • Common Kubernetes services, including GPU Operator, DRA, and Network Operator

  • BCM-Provisioned K8s

Compute Plane#

NVL72 Rack - 8 Racks per SU

  • Execution hosts composed of Compute Trays

  • Each Compute Tray contains CPU (C), GPU (U), and integrated memory

  • Trays run as Slurm workers or, on GB200/GB300, K8s + Run:ai workers depending on their assigned role

Additional Systems#

The control plane integrates with the following additional infrastructure components:

  • BMS (Building Management Service) – Customer-provided API-compliant BMS

  • NVLink Switches

  • InfiniBand (IB) Switches and UFM

  • Ethernet Switches

  • NFS Storage

  • High-Speed Storage

Note

Dynamic Power Software is a beta component available through a scoped engagement and is not included in the release. Contact your NVIDIA account manager to participate. BMS must be customer-provided and API-compliant.

Key Features#

NVIDIA Mission Control provides the following key capabilities:

  • Supports both training and inference workloads across DGX B200/B300 and GB200/GB300 NVL72 platforms

  • Provides centralized management for configuration and observability through BCM GUI, CLI, and API

  • Delivers a standardized, scalable, and secure control plane for all supported NVIDIA systems

  • Includes NVIDIA Mission Control-autonomous hardware recovery and NVIDIA Mission Control-autonomous job recovery for resilient, self-healing operations

  • Integrates Run:ai and Slurm workload managers for flexible GPU orchestration

  • Supports Dynamic Power Software for GB200/GB300 NVL72 and early access Vera Rubin NVL72 through a separate beta engagement

  • Provides NetQ unified observability for GB200/GB300 systems

  • Offers RCP Plugin for Spectrum-X for B200/B300 networking (experimental)