fuzzforge_ai/docs/docs/how-to/docker-setup.md

Docker Requirements for FuzzForge

FuzzForge runs entirely in Docker containers with Temporal orchestration. This guide covers system requirements, worker profiles, and resource management to help you run FuzzForge efficiently.

---

System Requirements

Docker Version

• Docker Engine: 20.10.0 or later
• Docker Compose: 2.0.0 or later

Verify your installation:

docker --version
docker compose version

Hardware Requirements

Component	Minimum	Recommended
CPU	2 cores	4+ cores
RAM	4 GB	8 GB+
Disk	20 GB free	50 GB+ free
Network	Internet access	Stable connection

Port Requirements

FuzzForge services use these ports (must be available):

Port	Service	Purpose
8000	Backend API	FastAPI server
8080	Temporal UI	Workflow monitoring
7233	Temporal gRPC	Workflow execution
9000	MinIO API	S3-compatible storage
9001	MinIO Console	Storage management
5432	PostgreSQL	Temporal database

Check for port conflicts:

# macOS/Linux
lsof -i :8000,8080,7233,9000,9001,5432

# Or just try starting FuzzForge
docker compose -f docker-compose.yml up -d

---

Worker Profiles (Resource Optimization)

FuzzForge uses Docker Compose profiles to prevent workers from auto-starting. This saves 5-7GB of RAM by only running workers when needed.

Profile Configuration

Workers are configured with profiles in docker-compose.yml:

worker-ossfuzz:
  profiles:
    - workers    # For starting all workers
    - ossfuzz    # For starting just this worker
  restart: "no"  # Don't auto-restart

worker-python:
  profiles:
    - workers
    - python
  restart: "no"

worker-rust:
  profiles:
    - workers
    - rust
  restart: "no"

Default Behavior

docker compose up -d starts only core services:

• temporal
• temporal-ui
• postgresql
• minio
• minio-setup
• backend
• task-agent

Workers remain stopped until needed.

On-Demand Worker Startup

When you run a workflow via the CLI, FuzzForge automatically starts the required worker:

# Automatically starts worker-python
fuzzforge workflow run atheris_fuzzing ./project

# Automatically starts worker-rust
fuzzforge workflow run cargo_fuzzing ./rust-project

# Automatically starts worker-secrets
fuzzforge workflow run secret_detection ./codebase

Manual Worker Management

FuzzForge CLI provides convenient commands for managing workers:

# List all workers and their status
ff worker list
ff worker list --all  # Include stopped workers

# Start a specific worker
ff worker start python
ff worker start android --build  # Rebuild before starting

# Stop all workers
ff worker stop

You can also use Docker commands directly:

# Start a single worker
docker start fuzzforge-worker-python

# Start all workers at once (uses more RAM)
docker compose --profile workers up -d

# Stop a worker to free resources
docker stop fuzzforge-worker-ossfuzz

Stopping Workers Properly

The easiest way to stop workers is using the CLI:

# Stop all running workers (recommended)
ff worker stop

This command safely stops all worker containers without affecting core services.

Alternatively, you can use Docker commands:

# Stop individual worker
docker stop fuzzforge-worker-python

# Stop all workers using docker compose
# Note: This requires the --profile flag because workers are in profiles
docker compose down --profile workers

Important: Workers use Docker Compose profiles to prevent auto-starting. When using Docker commands directly:

• docker compose down (without --profile workers) does NOT stop workers
• Workers remain running unless explicitly stopped with the profile flag or docker stop
• Use ff worker stop for the safest option that won't affect core services

Resource Comparison

Command	Workers Started	RAM Usage
docker compose up -d	None (core only)	~1.2 GB
fuzzforge workflow run atheris_fuzzing .	Python worker only	~2-3 GB
fuzzforge workflow run ossfuzz_campaign .	OSS-Fuzz worker only	~3-5 GB
docker compose --profile workers up -d	All workers	~8 GB

---

Storage Management

Docker Volume Cleanup

FuzzForge creates Docker volumes for persistent data. Clean them up periodically:

# Remove unused volumes (safe - keeps volumes for running containers)
docker volume prune

# List FuzzForge volumes
docker volume ls | grep fuzzforge

# Remove specific volume (WARNING: deletes data)
docker volume rm fuzzforge_minio-data

Cache Directory

Workers use /cache inside containers for downloaded targets. This is managed automatically with LRU eviction, but you can check usage:

# Check cache usage in a worker
docker exec fuzzforge-worker-python du -sh /cache

# Clear cache manually if needed (safe - will re-download targets)
docker exec fuzzforge-worker-python rm -rf /cache/*

---

Environment Configuration

FuzzForge requires volumes/env/.env to start. This file contains API keys and configuration:

# Copy the example file
cp volumes/env/.env.template volumes/env/.env

# Edit to add your API keys (if using AI features)
nano volumes/env/.env

See Getting Started for detailed environment setup.

---

Troubleshooting

Services Won't Start

Check ports are available:

docker compose -f docker-compose.yml ps
lsof -i :8000,8080,7233,9000,9001

Check Docker resources:

docker system df
docker system prune  # Free up space if needed

Worker Memory Issues

If workers crash with OOM (out of memory) errors:

1. Close other applications to free RAM
2. Start only needed workers (don't use --profile workers)
3. Increase Docker Desktop memory limit (Settings → Resources)
4. Monitor usage: docker stats

Slow Worker Startup

Workers pull large images (~2-5GB each) on first run:

# Check download progress
docker compose logs worker-python -f

# Pre-pull images (optional)
docker compose pull

---

Best Practices

1. Default startup: Use docker compose up -d (core services only)
2. Let CLI manage workers: Workers start automatically when workflows run
3. Stop unused workers: Free RAM when not running workflows
4. Monitor resources: docker stats shows real-time usage
5. Regular cleanup: docker system prune removes unused images/containers
6. Backup volumes: docker volume ls shows persistent data locations

---

Next Steps

• Getting Started Guide: Complete setup walkthrough
• Troubleshooting: Fix common issues

---

Remember: FuzzForge's on-demand worker startup saves resources. You don't need to manually manage workers - the CLI does it automatically!