layout-3 layout-6 layout-1 layout-top
NVIDIA Docs Hubchevron_right NVIDIA Networkingchevron_right Networking Softwarechevron_right Management Softwarechevron_right

NVIDIA NetQ

layout-top layout-6 layout-1

If you are using the current version of Cumulus NetQ, the content on this page may not be up to date. The current version of the documentation is available here. If you are redirected to the main page of the user guide, then this page may have been renamed; please search for it there.

Interfaces

Physical Interfaces Commands

Use the CLI to monitor OSI Layer 1 physical components on network devices, including interfaces, ports, links, and peers. You can monitor transceivers and cabling deployed per port (interface), per vendor, per part number, and so forth.

This information can help you:

  • Determine which ports are empty versus which ones have cables plugged in to help validate expected connectivity.
  • Audit transceiver and cable components by vendor, helping you estimate replacement costs, repair costs, and overall maintenance costs.
  • Identify mismatched links.
  • Identify when physical layer changes (for example, bonds and links going down or flapping) occurred.

NetQ uses LLDP (Link Layer Discovery Protocol) to collect port information. NetQ can also identify peer ports connected to DACs (Direct Attached Cables) and AOCs (Active Optical Cables) without using LLDP, even if the link is not UP.

View performance and status information about cables, transceiver modules, and interfaces with netq show interfaces physical:

netq show interfaces physical

View Utilization Statistics Networkwide

Utilization statistics can indicate whether resources are becoming dangerously close to their maximum capacity or other, user-defined thresholds. Depending on the function of the switch, the acceptable thresholds can vary.

Compute Resources Utilization

View how many compute resources—CPU, disk, and memory—the switches on your network consume with netq show resource-util:

netq show resource-util

Port Statistics

View statistics about a given node and interface, including frame errors, ACL drops, and buffer drops with netq show ethtool-stats:

netq show ethtool-stats

Interface Statistics and Utilization

NetQ Agents collect performance statistics every 30 seconds for the physical interfaces on switches in your network. The NetQ Agent does not collect statistics for non-physical interfaces, such as bonds, bridges, and VXLANs. The NetQ Agent collects the following statistics:

  • Statistics
    • Transmit: tx_bytes, tx_carrier, tx_colls, tx_drop, tx_errs, tx_packets
    • Receive: rx_bytes, rx_drop, rx_errs, rx_frame, rx_multicast, rx_packets
  • Utilization
    • rx_util, tx_util
    • port speed

To view interface statistics and utilization, run the netq show interface-stats or netq show interface-utilization commands:

netq show interface-stats 
netq show interface-utilization

ACL Resource Utilization Networkwide

View incoming and outgoing access control lists (ACLs) configured on all switches and host with netq show cl-resource acl:

netq show cl-resource acl

Forwarding Resources Utilization Networkwide

View forwarding resources on all devices with netq show cl-resource forwarding:

netq show cl-resource forwarding

SSD Utilization Networkwide

For NetQ Appliances that have 3ME3 solid state drives (SSDs) installed (primarily in on-premises deployments), you can view the utilization of the drive on demand. A warning is generated when a drive drops below 10% health, or has more than a 2% loss of health in 24 hours, indicating the need to rebalance the drive. Tracking SSD utilization over time lets you see any downward trend or drive instability before you receive a warning message.

View SDD utilization with netq show cl-ssd-util:

netq show cl-ssd-util

Disk Storage After BTRFS Allocation Networkwide

Customers running Cumulus Linux 3 which uses the BTRFS (b-tree file system) might experience issues with disk space management. This is a known problem of BTRFS because it does not perform periodic garbage collection, or rebalancing. If left unattended, these errors can make it impossible to rebalance the partitions on the disk. To avoid this issue, NVIDIA recommends rebalancing the BTRFS partitions in a preemptive manner, but only when absolutely needed to avoid reduction in the lifetime of the disk. By tracking the state of the disk space usage, users can determine when to rebalance.

For details about when to rebalance a partition, refer to When to Rebalance BTRFS Partitions.

View BTRFS disk utilization with netq show cl-btrfs-info:

netq show cl-btrfs-info

View interface (link) state, type, count, aliases, and additional information with variations of the netq show interfaces command, including netq show interfaces type and netq show events message_type interfaces:

netq show interfaces
netq show interfaces type
netq show events message_type interfaces

The netq check interfaces command verifies interface communication status for all nodes (leafs, spines, and hosts) or an interface between specific nodes in your network fabric. This command only checks the physical interfaces; it does not check bridges, bonds, or other software constructs.

netq check interfaces

To view interface information, select the Menu, then enter Interfaces in the search field.

Check for MTU Inconsistencies

The maximum transmission unit (MTU) determines the largest size packet or frame that can be transmitted across a given communication link. When the MTU is not configured to the same value on both ends of the link, communication problems can occur. Use the netq check mtu command to verify that the MTU is correctly specified for each link.

    Copyright © 2024  NVIDIA Corporation