https://github.com/ahrefs/dgx-helper-scripts

Nvidia DGX H100 helper scripts for support related tasks
https://github.com/ahrefs/dgx-helper-scripts

Last synced: about 1 year ago
JSON representation

Nvidia DGX H100 helper scripts for support related tasks

• Host: GitHub
• URL: https://github.com/ahrefs/dgx-helper-scripts
• Owner: ahrefs
• License: gpl-3.0
• Created: 2025-01-03T16:16:22.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2025-01-04T10:13:07.000Z (over 1 year ago)
• Last Synced: 2025-06-30T05:44:36.000Z (about 1 year ago)
• Language: Shell
• Homepage:
• Size: 24.4 KB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# dgx-helper-scripts

This is a collection of shell scripts for helping on support tasks for Nvidia DGX H100 cluster

It includes scripts for firmware upgrades, for debugging issues, and for several common actions needed to generate common logs requested by Nvidia enterprise support

Install these into ~/DGX folder on bright nodes.

Add nvfwupd and firmware files into some subfolder of ~/DGX

run ./00-config.sh and fill all missing info into "config" file

Check that ipmi.txt was properly filled with ipmi IPs of all dgx nodes (it should be ok if category of all dgx nodes is dgx-h100 in bright)