ProDeploy — Learn the fabric that keeps AI clusters alive

← Labs ← Chapter 3: The CLILab 0A - Fabric CLI orientation

Desktop recommendedChecking session

Lab 0A - Fabric CLI orientation

beginner

Elapsed

00:00

Scenario

You are standing in front of a healthy two-rail RoCE fabric. This is a calibration lab, not a failure lab. Objective: Build a correct mental map before you troubleshoot incidents. Topology: DGX Node 01 mlx5_0 / eth0 -> Rail 0 -> leaf-rail0 swp1 DGX Node 01 mlx5_1 / eth1 -> Rail 1 -> leaf-rail1 swp1 Your task: 1. Start on UFM and read the rail-to-endpoint mapping 2. On the DGX, inspect the HCA inventory and RDMA-to-netdev mapping 3. On Linux, verify eth0 and eth1 map to different rails 4. On each Cumulus leaf, verify swp1 is up and cabled to DGX Node 01 5. Use 'help' on each tab if you are unsure which command belongs to which platform

Conditions

UFM rail map checked
DGX HCA inventory checked
RDMA to netdev mapping checked
eth0 Linux link state checked
eth1 Linux link state checked
leaf-rail0 swp1 checked
leaf-rail1 swp1 checked

Need Help?

Open the step-by-step command guide for this lab.

Topology

Filled green circle = Active

Unfilled red circle = Down

Solid line = 400G link

Dashed line = degraded

Click a device node to open its terminal session

Preparing lab environment...