NVIDIA NEO User Manual and Release Notes v2.7.20
NVIDIA NEO User Manual and Release Notes v2.7.20

Release Notes Changes and Features History

This section lists the changes and new features of the previous versions of Mellanox NEO®.

Feature

Description

2.7.10

WebUI

Aligned NEO web user interface with NVIDIA template

2.7

New system support

Added support for SN3420, SN3700, and SN4600C switch systems

What Just Happened®

WJH is now enabled on Mellanox NEO by default upon Telemetry Agent installation

WJH now supports up to 5K WJH drops per second

Added ability to enable/disable WJH events per severity

Added the ability to filter out WJH events by category, severity or reason (Streaming Settings)

Enhanced wording for WJH events

Updated WJH dashboard

Connectivity check

Added ability to verify device connectivity across multiple communication protocols

Rev 2.6

Infrastructure

  • Transition to Python 3

  • Changing all UI tables to Grid Tables

Health monitoring

Added support for monitoring NEO processes and restarting them if they fail

Dashboard inventory

Display all managed devices inventory at NEO dashboard

Services

Network map

Added support for viewing sites in Network Map screen

Auto-provisioning

  • Automatically register for SNMP traps for every newly added Onyx switch

  • Automatically configure SNMP and LLDP for every newly added Cumulus switch

User management

Added support for limited view for NEO read only user

UI enhancements

  • Added alternative to right click operations

  • Display switch names whenever switches are displayed

  • Aligned devices status and health for all managed devices

  • Aligned the format of all error handling messages

NEO container HA

Add support for NEO docker container to run in high availability mode

VLAN auto-provisioning

  • Updating switch information (VLANs) once VLAN auto-provisioning is done

  • Added support for auto-VLAN provisioning according to the switchport mode detected

  • Configuring the VLAN name as part of the auto VLAN provisioning

Switch configuration comparison

Added option to compare configurations between two or more switches

Telemetry enhancements

  • Added protection for users to not use localhost (127.0.0.1) if they are deploying telemetry agent

  • Added option to configure telemetry collectors via the telemetry streaming page

NEO container upgrade

Added support for upgrading NEO container using docker volumes

NEO discovery

Added pop-up indication and notification summary for every device being discovered (by scanning)

Events policy enhancement

Added 3 predefined recipient lists per severity for users who want to send all NEO events for specific severity to a specified list of recipients

Rev 2.5.1

Telemetry Agent

Upgraded the Telemetry Agent to version 2.5.1-5

Telemetry Agent

Added provisioning templates to configure the Telemetry Agent via NEO

Telemetry Agent

Telemetry Agent LAG and MLAG discovery mechanism is now configurable and is disabled by default

NEO Discovery

Improved NEO switch interfaces discovery mechanism to be more efficient

WJH

The legend in WJH timeline graph is now displayed by reason category

Rev 2.5

WJH Enhancements

Added support for collecting and viewing ACL dropped packets.

Added support for collecting and viewing L1 dropped packets.

Added support for filtering the WJH reasons categories when querying WJH from the switch.

Bring up wizard enhancements

Added support for multiple MLAGs configuration within the Bring Up wizard

Added support for VMWare vCenter and Nutanix Prism integrations setting via the Bring Up wizard.

Auto VLAN Provisioning

Added support for automatic VLAN provisioning (on both Onyx and Cumulus switches) in case of networking or VM event, triggered by the vCenter.

Added support for automatic VLAN provisioning over remote clusters (VXLAN) using Onyx switches connected to Nutanix AHV hypervisors.

Added support for extending the amount of hypervisors in the Nutanix setup when working in “Global VLAN Provisioning” mode.

Transition to InfluxDB

Enabled collecting and inserting telemetry and monitoring data into InfluxDB (instead of Graphite).

RoCE Cleanup

Added support for removing RoCE configuration from switches which are part of NEO RoCE service.

NEO on Switch

Integrated the InfluxDB as part of NEO on Switch (NEO docker container) capability.

Enhanced Data collection

NEO was optimized to collect data from either Monitoring or Telemetry. Once Telemetry is enabled, Monitoring is disabled and vise versa.

Telemetry Agent

Performance improvements and more efficient way to collect telemetry data and stream it to external collector.

Added support for upgrading the telemetry agent version via NEO Web UI.

Added a mechanism which prevents the telemetry agent deployment in case the switch and NEO clocks are not synchronized.

Added support for retrieving telemetry data for logical switch ports like LAG and MLAG.

Added support for streaming telemetry data to Kafka Consumer.

Added RAM limitation of 300M for the docker container running the telemetry agent.

Added a new notification which lists the current switches that are using old telemetry agent version (which require upgrade).

Rev 2.4

Snapshot Differences and Notifications/Events

Allows the user to create a telemetry snapshots of periodic CLI command output. These snapshots can be used to alert the user when a difference is found in the output.

gRPC Streaming Enhancements

Streams the Telemetry data to NEO using a unified collector with gRPC protocol.

Optimize NEO Refresh Rate

Shortens the refresh time of the device's Docker data.

Bring-up wizard for 2-switch MLAG for ESF

Allows the user to bring-up a network from scratch in a few quick steps.

VMware vCenter Virtualization Support (DVS)

Allows the user to manage ESXi hosts in NEO.

Generate System Dump

Allows the user to generate debug dumps for Onyx switches and upload them to a remote folder.

Support L2 in the ToR use-case

Nutanix integration for L2 networks.

Nutanix Calm BP update

Updated NEO Blueprint on Nutanix Calm.

Nutanix Plugin API Change to Webhook v3

Updated NEO Nutanix plugin to use Nutanix v3 REST APIs.

Package InfluxDB as an External Container with NEO

Provides a Docker container with InfluxDB and Grafana to be used as a Telemetry data collector.

Grafana Visualization Pack (on top of InfluxDB)

What-Just-Happened (WJH) Debug

Provides the user more information on WJH failures.

Link Monitoring

Provides the user better visibility on the device and the ports' traffic, and the network behavior.

WJH integration with SDK

Integrated the Telemetry Agent with SDK for WJH data.

Onyx EVPN Templates

Added templates for EVPN configuration on Onyx.

WJH Enhancements

Enhanced the WJH dashboard user interface.

Telemetry

Added support for modifying the collectors to an existing session.

Power Management

Removed support for the eco mode as this capability is not supported in Spectrum based switches.

Rev 2.3.1

Telemetry Agent

Telemetry sessions performance improvements

NEO Telemetry Management

Telemetry status reflection improvement

Discovery

Add option for disabling VLAN discovery

Task Management

Added an option to attach additional NEO data model attributes to the collected switch information

NEO VM Deployment

Improved the NEO OVA image for easier VM deployments

Rev 2.3

RoCE Dashboard

The RoCE Dashboard contains a snapshot of the RoCE related network state, including information on service state, traffic and events. RoCE services can also be added and managed from this dashboard.

Nutanix Virtualization Improvements

Allows discovering the virtualization information (VM name) using prism central API.

Cisco Model 2960 Switches Support

Cisco switches model 2960 can be managed by NEO, apart from provisioning related operations.

What Just Happened

A dashboard that contains information about packet drops in the fabric.

MAC Address table for Onyx

Allows the Telemetry Agent to retrieve MAC Tables data.

Switch Agent (Onyx): Routing Tables

Allows the Telemetry Agent to retrieve Routing Tables data.

Telemetry Agent for SwitchDev (interface and port counters)

Allows running the telemetry agent as a docker container on SwitchDev switches. The data is collected using ethtool.

Telemetry Agent for Linux Host

Allows running the telemetry agent as a docker container on Linux hosts. The data is collected using ethtool.

LLDP Discovery APIAutomatic SNMP v3 Registration

Enables the user to attach the 'Register for SNMP v3 traps for Mellanox switch' built-in task to the ‘Device Added’ event so it will run every time a new device has been added.

Port Unsilenced Admin State

Added support for ‘Unlicensed’ port admin state.

Planned-Reload Event

Added a new event policy - ‘Planned Reload’.

MAC Address Table Scheduled Task

Allows the user to set a scheduled task to obtain the MAC address table for each Onyx device.

Rev 2.2

Telemetry Agent Enhancements

  • Added calculated counters (rate and normalized counters)

  • Added error handling and telemetry sessions status

  • Added support for multiple sessions and multiple destinations

  • Added a mechanism for triggering threshold crossing events

Telemetry Agent on Cumulus Linux

Added support for deploying and running telemetry agent on Cumulus Linux.

Telemetry Integration with ELK

Added support for switch telemetry data integration with ELK.

One Click RoCE

Added support for deploying RoCE on Windows and Cumulus Linux systems.

Network Map Enhancements

  • Added support for link utilization (showing bandwidth utilization per link)

  • Added a new hierarchical view of managed devices

Log Debug level via Web UI

Added support for NEO controller log debug level.

Multi-site Support

Added support for grouping and managing devices per site

Network Path

Added a view of the optional network paths between a selected switch and a target host.

Nutanix Integration

Added support for NEO integration with Nutanix Prism Central and Multi-Cloud environment.

Rev 2.1

One-click RoCE

  • Updated the RoCE recipes for:

  1. ECN only

  2. ECN + QoS

  3. ECN + QoS + PFC.

  • Added the ability to select specific switches, hosts and ports to RoCE provision.

  • Added the ability to easily grow the RoCE fabric after the initial provisioning with additional switches, hosts and ports.

Telemetry Agent (Early Availability)

The NEO telemetry agent is a software module designed to run inside a Docker container operated on a Mellanox Spectrum switch system. The agent collects data on the switch, and streams it out to an external data collector for processing, analysis and presentation.

The telemetry agent can be centrally deployed and managed from the NEO application or from a 3rd party controller.

Mellanox Spectrum Containers Management

Added the ability to orchestrate Docker containers on Mellanox Spectrum switch systems operated with Mellanox Onyx software.

Cable Inventory

Added the ability to view and manage an inventory of cables across the fabric through one pane of glass.

NEO Docker Appliance

The NEO Docker appliance is a pre-packaged Docker container image operating on CentOS and installed with the NEO 2.1 software.

Backup & Restore

Added the ability to backup and restore a NEO instance for data protection and recovery purposes.

Event Management

  • The condition ID was changed to be numeric instead of textual data

  • The event description contains the condition message instead of the reason

  • The reason field is generated by NEO, and contains Reason information per event ID. The user cannot modify the Reason information

  • The 'RelatedObjectID', 'RelatedObjectName' and 'RelatedObjectType' new attributes were added to each event

  • The 'Parent' attribute was changed to 'RelatedSystem'. Only one related system is supplied

  • The 'ConditionID' attribute was added

  • Additional log data entries were added - 'ConditionID‘, Event sequence ID and related object data

  • The corrective description is generated by NEO, and contains corrective information per condition ID

Remote Storage

Added the ability to change the default path where NEO stores all configuration files from the local host to a remote storage.

NEO-Host Notification

Added new notification listing hosts that are not installed with NEO-Host, as well as the ability to install the latest NEO-Host software.

Device Access

Added the ability to set http/https device access for every Mellanox Onyx operated device.

Export/Import Network Map Views

Added the ability to export Network Map views as files and import them to other NEO user environments.

QoS Monitoring

Added the ability to set traffic priority for monitoring, with priority 3 being the default for RoCE traffic.

Network Services

Added an option to automatically create a network snapshot before applying a service configuration to allow a restoration point in case of a failure.

Provisioning

Added several new provisioning templates.

Rev 2.0.5

RoCE Profiles Support

Added the ability to use MLNX_OS 3.6.5000 traffic pool APIs for RoCE service configuration.

HTTPS Communication

Added the option to replace HTTP access to the switches with HTTPS access.

Configuration Management Module

Added a new module to consolidate all related configuration management functions, such as network provisioning templates, global configuration templates, etc. in a single pane of glass.

Running-Config Variables

Added the ability to use variables in running-config templates such that a single running-config can be provisioned on multiple switches in a single operation.

Network Snapshots

Added support for creating network-wide snapshot of switch running-configurations to use as restoration points. By default, a network snapshot is created automatically by the system every 24 hours.

One-click MLAG

Expanded the existing MLAG service to provision and validate IP networks and servers on the switch and host.

Nutanix AHV Support

Added support for auto-discovery and visibility of Nutanix Acropolis (AHV) virtual machines across the network fabric.

Linux Virtual Networking

Added modeling and visibility into a Linux host virtual networking and virtual functions bridging to physical adapters.

NEO Virtual Appliance for PowerKVM

Added support for new NEO Virtual Appliance compatible with PowerKVM based systems.

Device MAC

Added a new identifier for switch systems.

Device Access

Added SNMP-based connection settings per switch and changed “Device Credentials” to “Device Access”.

Syslog Server

Added configuration options to set up to five servers to export syslog messages to, each has its own service/port settings.

Live Monitoring Interval Improvement

Reduced the minimal counters monitoring interval of Spectrum and SwitchX devices from 20 to 2 seconds.

NMOS API

Added support for Networked Media Open Specifications (NMOS) API.

For further information on NMOS, visit the following links: https://www.nmos.tv/about_NMI.html

Rev 1.9

Enhanced Server/NIC Support

Added Mellanox adapter and port information through integration with NEO-Host software operated on the host.

Enhanced Device Discovery

Disabled the default automated device discovery mechanism, and enabled the user to create and manage a discovery policy of allowed IP subnets and/or ranges.

Added a new discovery solution of IP subnet/range scan.

Configuration Management

Added the option to schedule a configuration swap/rollback on a Mellanox switch or a group of switches using an Apply Config task operation.

Added the option to edit switch configuration through a built-in config editor.

One-Click RoCE Service

Added built-in automation for the provisioning and validation of RDMA over Converged Ethernet (RoCE).

This feature is designed for Mellanox Spectrum switch systems operated with Mellanox Onyx, and Linux servers operated with Mellanox ConnectX-4 or ConnectX-5.

RoCE Monitoring

Added sampling of additional counters on Mellanox switches, used for RoCE traffic.

This feature is designed for Mellanox Spectrum switch systems operated with Mellanox Onyx.

NEO Software Upgrade

Added an enhanced solution for NEO software upgrade from NEO v1.9 and up.

Enhanced Device State

Added the following new device states:

  • Pending Reboot – indicates that a device is pending reboot (as a result of a software upgrade or apply configuration)

  • Reboot Status – indicates the status of a device upon reboot operation

Enhanced LAG/MLAG Support

Added an enhanced support for LAG/MLAG port-channels on Mellanox Onyx powered switch systems with:

  • One-click provisioning – configuring LAG/MLAG ports configuration based on physical connectivity

  • LAG/MLAG information – added information on LAG/MLAG port members/peers, state, MTU, etc.

  • One-click validation – monitoring existing LAG/MLAG ports through network services

Rev 1.8

Network Health Enhancements

Added new network health tests for:

  • IP connectivity between all nodes to all nodes, or random selection of one node per switch/rack.

Multicast latency and bandwidth tests between all nodes to all nodes, or random selection of one node per switch/rack.

Network Health Report

Allows the user to execute data traffic tests across the network, to ensure utmost network performance

Task Sequence

Allows the user to select multiple tasks that will run in a sequence, to achieve end-to-end network automation in a single run-time

End to End RoCE Automation

Added a new wizard that configures Lossless RoCE/Resilient RoCE on an entire fabric comprised of Mellanox switches, Linux and Windows hosts

Ports Group

Enablse the user to create a group of ports, to achieve port-level network automation and create monitoring policies

Software Upgrade Profiles

Allows the user to manage software upgrade profiles for a more intuitive way to upgrade Mellanox software on switches and servers

VLAN Auto-Provisioning for Nutanix Acropolis Virtualization

Added network automation for configuring VLAN networks on Mellanox switches, for Nutanix Acropolis VM life-cycle, VM creation, VM migration and deletion.

This is provided as a software add-on/plug-in.

Rev 1.7

High Availability

Added the option to install and run NEO on a cluster of nodes that are synchronized and can take over each other in cases of failovers

Cable Information

Added the option to view cable information of selected devices, ports and groups

Performance Monitoring

Added the ability to run performance checks between two hosts

New Supported Virtual Appliances

Added new VM images of NEO for the following hypervisors:

  • VMWare Workstation

  • ESXI Server

  • Virtual-BOX

  • Hyper-V

Port Level Provision

Added the ability to run provisioning actions for Mellanox Onyx switch ports

L3 Network Provisioning

Added the option to define a dynamic layer 3 network configuration which can be automatically applied for new discovered systems

Syslog Configuration

Added the option to change NEO syslog configuration settings and filter for events to be written to Syslog

Support for 3rd Party Switches

Added support for managing HP and Brocade switches via monitoring and provisioning

Saving Topology Layouts

Added the option to save a user-defined layout of the discovered topology

Skin Personalization

Added the option to personalize the UI view of NEO in terms of colors and theme

Rev 1.6.1

NEO Installation and Start-up Enhancements

Made various enhancements to the NEO installation process to make the installation experience easier for the user. These enhancements include:

  • An error message that appears once Python packages conflict with NEO packages during installation.

  • An error message that appears once RPM conflict with NEO packages during installation.

  • An error message that appears when trying to access NEO GUI while SELinux is enabled. This message asks for disabling SELinux to be able to access the GUI.

Mellanox Virtual Modular Switch (VMS)

Added support for configuring SN2410 and SN2700 switch systems as members in VMS.

Rev 1.6.0

Enhanced Network Awareness

Presents VLAN and LAG configuration data per Mellanox switch on the fabric.

Virtualization Support

Enables NEO to automatically identify and extract data from Linux KVM hypervisos regarding virtual machines’ properties and virtual switches. This data is presented to the user per device and on a fabric level.

Services View

Services View is a single pane of glass for initial network provisioning and monitoring. NEO services are used to provision complex configurations on multiple devices, and upon completion, constantly monitors these configurations. It includes an Out-of-the-box support for Virtual Modular Switch (VMS), Lossless Fabric, MLAG, and MTU services.

Live Monitoring

Enables the user to create graph reports on device and traffic counters with intensive polling rates, whenever real-time monitoring capabilities are needed.

Configuration Management

Enhanced configuration management with the following capabilities:

  • Provision new switches by pushing full device config

  • Rollback switch config to a last known good configuration

  • New global config repository to manage global device configurations

MLNX_OFED Software Upgrade

Added the option to upgrade MLNX_OFED software on Linux servers.

Maintenance Mode

Enables the user to set NEO to maintenance mode where all NEO events will be disabled.

Dashboard Enhancements

Added the option to create custom device heatmap views based on various types of attributes (CPU, memory, traffic, etc.), and user-defined thresholds.

Network Map Enhancements

Enhanced network map with the following capabilities:

  • Added filtering capabilities based on device type (switch, host, etc.), severity (warning, error, etc.) and VLAN number.

  • Enhanced search capabilities to use standard strings.

Mellanox Care Support

Enables Mellanox NEO customers to subscribe to Mellanox Care support service, where all configured events/alerts will be relayed to the Mellanox Care support team constantly.

Logs

Added the option to download log files.

Kerberos Authentication

Added support for Kerberos authentication for Windows hosts management.

Rev 1.5

Topology Map

Added a visual view of the physical connectivity between managed devices.

VMS Configuration

Added a wizard for Virtual Modular Switch™ (VMS) configuration.

Provisioning of 3rd Party Switches

Added the option to provision configurations to 3rd party switches using templates.

Traffic Monitoring of 3rd Party Switches

Added the option to monitor traffic counters for 3rd party switches using SNMP protocol.

NEO SDK Enhancement

Added support for additional SDK scripts for NEO integrations.

Task Management Enhancement

Enhanced performance of task management and task editing.

RH7.0 Support

Added the option to install NEO on an RH7.0 system.

VM Deployment

Added a pre-configured Virtual Machine (VM) image with NEO for deploying it on a Linux KVM hypervisor system.

© Copyright 2023, NVIDIA. Last updated on Nov 16, 2023.