Purpose

This article explains how to collect diagnostic logs from NVIDIA Unified Fabric Manager (UFM) using the web-based GUI. The UFM system dump is typically required by NVIDIA Support for troubleshooting fabric health, host visibility, alerts, and UFM service issues.

Scope

This procedure applies to:

• NVIDIA UFM Enterprise deployments

• Environments using the UFM Web GUI

• Troubleshooting InfiniBand fabric, hosts, switches, and UFM-related issues

Prerequisites

• Access to the UFM Web GUI

• Administrative or equivalent privileges in UFM

Procedure

Step 1: Log in to the UFM Web GUI

1. Open a web browser.

2. Navigate to the UFM management URL:

   https://<ufm-management-ip>

3. Log in using your UFM administrator credentials.

Step 2: Navigate to System Health

1. From the left-hand navigation menu, click System Health.
   

Step 3: Open UFM System Dump

1. Within the System Health section, locate UFM System Dump.

2. Click Create System Dump.
   

---

Step 4: Create the System Dump

1. After clicking Create System Dump, wait for the dump generation to complete.

2. The process may take several minutes depending on fabric size and UFM load.

Note: Do not refresh the browser or navigate away while the dump is being generated.

Step 5: Download the System Dump

1. Once the system dump is ready, click the Click Here button.

2. The system dump file will be downloaded to your local system automatically.
   

Notes and Best Practices

• Always collect the UFM system dump as close to the incident time as possible.

• Avoid restarting UFM services before collecting the dump unless instructed.

• Store system dumps securely, as they may contain fabric topology and configuration details.

• Collect UFM system dumps in addition to host and switch logs for complete analysis.