HPC-Slurm-Hex

A GUI layer for SLURM, transforming traditional CLI-based SLURM into an intuitive web interface. Works seamlessly with pre-installed SLURM, requiring no removal or reconfiguration. Deployed as three independent modules: Master, Worker, and Frontend.

Key Features

• GUI for SLURM: Manage jobs, users, and resources via a web dashboard.
• Modular Deployment: Deploy Master, Worker, and Frontend separately.
• No SLURM Overhaul: Enhances preconfigured SLURM without replacing it.
• Admin Dashboard: Control user/group management, resource allocation, and job CRUD operations.
• Scalable: Connect multiple worker nodes and manage via the Frontend.

Architecture

HPC-Slurm-Hex consists of three modules:

• HPC-Slurm-Master: Central management node (Node.js, Redis).
• HPC-Slurm-Worker: Worker nodes for job execution (Python).
• HPC-Slurm-Frontend: Web interface for admin/user interaction (Docker).

Quick Setup Guide

1. HPC-Slurm-Master (Deploy on Master Node)

• Download HPC-Slurm-Master.zip from the latest release.

• Prerequisites: Node.js, Redis (running on port 6379).

• Navigate to the directory, run:

  npm install
  npm start

• Runs on port 5050. Keep it running.

2. HPC-Slurm-Worker (Deploy on Worker Nodes)

• Download HPC-Slurm-Worker.zip from the latest release.

• Prerequisite: Python 3.

• Navigate to the directory, run:

  pip install -r requirements.txt
  python3 app.py

• Runs on port 5050. Keep it running.

3. HPC-Slurm-Frontend (Deploy on Admin/User Machine)

• Prerequisite: Docker engine.

• Download docker-compose.yaml from the latest release.

• Run:

  docker compose up

• Visit localhost:5051, set up the admin user, and access the dashboard.

• Connect Master/Worker nodes:

  • Go to "Environment" in the sidebar.
  • Add Master node’s public IP and connect.
  • Add Worker nodes’ public IPs (specify quantity, e.g., 2) and connect.

• Manage users, groups, resources, and jobs via the sidebar.

Contributions welcome! Feel free to open issues or PRs.