Fan Manager for Dell PowerEdge Server with NVIDIA GPUs

Enterprise-grade fan controller for 13th Gen Dell PowerEdge Server using iDRAC. Controlling the fans by monitoring the temperature of passive cooled NVIDIA GPUs and the CPUs, built with C# and .NET 10 running in Docker.

🎯 Features

• ✅ Testable Architecture - Interfaces for all external dependencies (IIpmiService, INvidiaSmiService, IProcessService)
• ✅ Clean Architecture - Separation of Core, Infrastructure, and Worker layers
• ✅ Modern C# 14 - Records, pattern matching, nullable reference types
• ✅ Full Unit Test Coverage - Demonstrates mocking with Moq
• ✅ Docker Support - Multi-stage build with nvidia/cuda base image
• ✅ Production Ready - Logging, health checks, graceful shutdown

🚀 Quick Start

Prerequisites

• Docker
• Nvidia Container Runtime (for GPU support in Docker)
• Dell PowerEdge Server (Gen 13) with iDRAC 8 and IPMI over LAN Enabled
  • To configure IPMI over LAN, go to Overview > iDRAC Settings > Network in the iDRAC Web interface. Under IPMI Settings, specify the values for the attributes and click Apply.

⚙️ Configuration

Configuration via environment variables with prefix FANMANAGER__:

Variable	Default	Description
FANMANAGER__IPMI__HOST	local	iDRAC host ('local' or IP address)
FANMANAGER__IPMI__USERNAME	root	iDRAC username
FANMANAGER__IPMI__PASSWORD	calvin	iDRAC password
FANMANAGER__BASE_FAN_SPEED	20	Base fan speed (%)
FANMANAGER__CPU_TEMPERATURE_THRESHOLD	60	CPU threshold (°C)
FANMANAGER__GPU_TEMPERATURE_THRESHOLD	45	GPU threshold (°C)
FANMANAGER__GPU_TEMPERATURE_MAX	-	Optional: Maximum GPU temperature (°C)
FANMANAGER__GPU_IGNORE_ACTIVE_COOLED	true	Ignore active cooled gpus on highest gpu temp detection
FANMANAGER__CHECK_INTERVAL	00:00:30	Check interval (TimeSpan format)
FANMANAGER__ENABLE_DELL_THIRD_PARTY_PCIE_CARD_COOLING_BEHAVIOR	false	Enable Dell third-party PCIe card cooling
FANMANAGER__RESTORE_DELL_THIRD_PARTY_PCIE_CARD_COOLING_BEHAVIOR_ON_EXIT	true	Restore Dell cooling behavior on exit

Development

Local Development

# Restore dependencies
dotnet restore

# Build solution
dotnet build

# Run tests
dotnet test

# Run locally (requires IPMI and GPU access)
dotnet run --project src/FanManager/FanManager.csproj 

Docker Build & Run

# Build Docker image
docker build -t fan-manager:latest .

🧪 Running Tests

# Run all tests
dotnet test

# Run with coverage
dotnet test --collect:"XPlat Code Coverage"

# Run specific test
dotnet test --filter "FanControlStrategyTests"

🏗️ Building from Source

# Clone repository
git clone <repo-url>
cd FanManager

# Restore dependencies
dotnet restore

# Build Release
dotnet build -c Release

# Publish self-contained
dotnet publish src/FanManager -c Release 

🐳 Docker Details

Multi-Stage Build

1. Build Stage - Uses .NET SDK to compile
2. Runtime Stage - Uses nvidia/cuda with ipmitool installed

Health Check

Container includes health check monitoring the worker process every 60 seconds.

---

Best Practices Implemented

• ✅ SOLID Principles - Clean separation of concerns
• ✅ Dependency Injection - All dependencies injected
• ✅ Interface Segregation - Small, focused interfaces
• ✅ Testability - All business logic unit testable
• ✅ Async/Await - Proper async patterns
• ✅ Nullable Reference Types - Compile-time null safety
• ✅ Records - Immutable data structures
• ✅ Pattern Matching - Modern C# patterns
• ✅ Structured Logging - Microsoft.Extensions.Logging
• ✅ Options Pattern - Type-safe configuration with validation

---

Roadmap

• Web API – A REST API to expose sensor values and stats for integration with dashboards like Homepage or Home Assistant
• Manual Override via Web API – Ability to manually override fan settings through the API
• GPU Selection
  • Option to select specific NVIDIA GPUs to include/exclude
  • Option to ignore actively cooled GPUs for highest temperature detection
• Extended Hardware Metrics – Additional server and GPU statistics via nvidia-smi, including GPU utilization and VRAM usage
• Custom Fan Curve – Support for defining a custom fan curve to fine-tune fan behavior based on temperature thresholds
• Additional Device Support – Support for other devices beyond GPUs that are cooled by the server fans (e.g. PCIe cards, storage, etc.)
• Inlet & Exhaust Temperature – Include inlet and exhaust air temperatures in the fan control logic for more accurate thermal management
• Controller & Provider Architecture – Architectural extension introducing a central controller and provider model, allowing resources running in other VMs or remote systems to report their metrics to the central controller
• External Webhooks – Ability to trigger external webhooks on temperature thresholds (e.g. to activate an air conditioning unit or send alerts)
• Centralized Log Forwarding – Support for forwarding logs to a central logging service (e.g. Loki, Graylog, or similar)

---

Contributing

Contributions are what make the open source community such an amazing place to learn, inspire, and create. Any contributions you make are greatly appreciated.

If you have a suggestion that would make this better, please fork the repo and create a pull request. You can also simply open an issue. Remember to give the project a star! Thanks again!

Contributing Guidelines

1. Fork the Project
2. Create your Feature Branch (git checkout -b feature/123456_Amazing-Feature)
3. Commit your Changes (git commit -m 'feat(something): Add some AmazingFeature')
4. Push to the Branch (git push origin feature/123456_Amazing-Feature)
5. Open a Pull Request

Branch Naming

Branch naming conventions are enforced via git hooks. All branches must follow one of these patterns:

Pattern 1: <type>/<ticket-number>_<description>

• type: Must be one of: bugfix, feature, impediment
• ticket-number: 5 or more digits (e.g., 00001)
• description: Lowercase alphanumeric with hyphens (e.g., amazing-feature)

Pattern 2: develop

• Protected main development branch

Valid examples:

• feature/12345_add-gpu-monitoring
• bugfix/67890_fix-memory-leak
• impediment/11111_update-dependencies
• develop

Invalid examples:

• feature/1234_short-ticket (ticket number is too short)
• Feature/12345_uppercase (type must be lowercase)
• feature/12345_Invalid_Name (description must be lowercase with hyphens only)

Commit Messages

We follow the Conventional Commits specification for all commit messages. This standard is enforced via git hooks.

Format: <type>(<scope>): <description>

Commit message length: Maximum 100 characters

Common types:

• feat: New feature
• fix: Bug fix
• docs: Documentation changes
• refactor: Code refactoring
• test: Test additions or modifications
• chore: Maintenance tasks

---

📄 License

Inspired by: tigerblue77/Dell_iDRAC_fan_controller_Docker

Distributed under the MIT License. See LICENSE for more information.

Legal Disclaimer

This project is an independent open-source utility.

• iDRAC and PowerEdge are registered trademarks of Dell Inc.
• NVIDIA is a registered trademark of NVIDIA Corporation.

This software is not affiliated with, sponsored by, or endorsed by Dell Inc. or NVIDIA Corporation. Use it at your own risk.

---

Built with ❤️ using .NET 10 and C# 14