NVIDIA UFM Enterprise User Manual

Changes and New Features

UFM Enterprise Changes and New Features

Notes:

  • For an archive of changes and features from previous releases, please refer to Changes and New Features History.

  • The items listed in the table below apply to all UFM license types.

  • For bare metal installation of UFM, it is required to install DOCA_HOST before the UFM installation. Please make sure to use the UFM installation package that is compatible with your setup, as detailed in Bare Metal Deployment Requirements.

Feature

Description

ASIC Health Monitoring on NVOS-Managed Switches

Added support for ASIC health monitoring on NVOS-managed switches, flagging ASIC faults as critical by marking the switch red and generating a Critical event (previously shown as healthy). For more information, refer to Threshold-Crossing Events Reference (Event ID #409).

Enhanced ASIC/Plane Port Visibility

Enhanced Threshold-Crossing Event #1604 (Non-optimal aggregated port bandwidth) to identify and list specific inactive, missing, or non-functional plane ports (ASICs), enabling faster root-cause analysis instead of reporting only the active port count. For more information, refer to Threshold-Crossing Events Reference.

Secured BareMetal Cloud Configuration Validation

Added the ability to validate secured BareMetal Cloud configurations for CSP/NCP deployments, including verification of IB security settings (opensm.conf, gv.cfg) and partitioning (PKeys configurations), with errors, warnings, and info reported based on verbosity level. For more information, refer to Security Tab. For REST API, refer to IB Security Verification REST API (TBD Link).

NVOS Switch Health 

Added the ability to retrieve and expose health status from NVOS switches in UFM, enabling clients to monitor switch health directly through UFM, with automatic support detection on NVOS-compatible switches. For more information, refer to Switches Health Window (for all switches) or Health Tab (for selected switches).

UFM Prime Topology Compare

Added the ability (via REST API and UI) to perform topology comparisons, both manual and scheduled, for UFM Providers managed by UFM Prime. For more information, refer to UFM Prime. For the dedicated REST API, refer to REST API Complementary Information.

UFM Prime XDR Router Support (Beta Level)

Added discovery of XDR routers connecting different IB subnets (managed by UFM providers), with support for visualizing them in the unified Network Map, inspecting router internals, and exposing XDR router telemetry data. For more information, refer to UFM Prime (XDR router provisioning is not supported in this UFM Enterprise version). 

UFM Prime - Cable FW Management

Added support for managing cable transceiver firmware (upload/list/delete/burn) across all providers in one Web UI operation. For more information, refer to Cable Transceiver Firmware Upgrade Across Providers.

UFM Prime - Tenant Performance Monitoring

Added support for collecting tenant performance counters across UFM providers in a single aggregated job. For more information, refer to Tenant Performance Monitoring Across Managed UFM Providers.

UFM Prime - PKey Management

Added support for viewing and managing PKeys and member GUIDs across UFM providers from one unified interface. For more information, refer to PKey Management Across Managed UFM Providers.

UFM Prime

Added the ability in UFM Prime to verify all-to-all connectivity and assess link quality. For more information, refer to UFM Cable Validation Tool Plugin.

Added the ability in UFM Prime to alert on connectivity mismatches across subnets. For more information, refer to UFM Cable Validation Tool Plugin.

NDR Switch Ports View

Added support for displaying switch ports in mixed split mode (split and non-split ports) when the switch profile is configured as split-ready. For more information, refer to Ports Window.

Telemetry Microservice

Added the ability to run telemetry data collection and computation as a dedicated microservice by default, improving UFM main process responsiveness, with an option to disable via [TelemetryService] enabled = false in gv.cfg. By default, this feature is enabled (allowing improved performance in large-scale setups). For more information, refer to Telemetry Microservice.

Kubernetes (Helm) Deployment

Added the ability to deploy UFM Enterprise on Kubernetes using Helm charts, enabling declarative configuration, simplified operations, and plugin deployment as separate pods. For more information, refer to UFM on Kubernetes.

RAMP Plugin

Introduced the UFM RAMP plugin, a UFM-managed Docker-based plugin with REST API support that enables fabric-level monitoring of PKey assignments and exposes them for subnet nodes. As part of the plugin, end-to-end validation of PKey configuration is supported. For more information, refer RAMP Plugin.

Plugins Changes and New Features

For plugin changes, new features, bug fixes or known issues, refer to plugin documentation under UFM Plugins.

Unsupported Functionalities/Features

The following distributions are no longer supported in UFM:

  • Ubuntu18.04/Unbuntu20.04

Last updated: