Before Troubleshooting: Check the Logs
Most issues leave clear error messages in the logs. Always start here:
# View recent logs
pnpm openclaw logs --tail
# View logs for a specific component
pnpm openclaw gateway logs
pnpm openclaw channels telegram logs
# Check system status
pnpm openclaw gateway status
pnpm openclaw models list
Category 1: Installation Errors
"node: command not found" or "npm: command not found"
Cause: Node.js is not installed or not in PATH.
Solution:
# Install via nvm (recommended)
curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.39.0/install.sh | bash
source ~/.bashrc
nvm install 22
nvm use 22
# Verify
node -v # Should show v22.x.x
"pnpm: command not found"
Solution:
npm install -g pnpm
source ~/.bashrc
pnpm -v # Should show version number
"pnpm install" fails with ENOENT or EACCES
Cause: Permission issues or missing directories.
Solution:
# Fix npm global directory permissions
mkdir -p ~/.npm-global
npm config set prefix '~/.npm-global'
export PATH=~/.npm-global/bin:$PATH
echo 'export PATH=~/.npm-global/bin:$PATH' >> ~/.bashrc
"ENOTFOUND" during install (can't reach npm registry)
Cause: Network issue reaching the npm registry (common in China).
Solution:
# Switch to Taobao mirror
pnpm config set registry https://registry.npmmirror.com
pnpm install
Category 2: Gateway Errors
Gateway fails to start: "port already in use"
Cause: Another process is using port 18789.
Solution:
# Find what's using the port
lsof -i :18789
# Kill the process (replace PID with the actual PID)
kill -9 PID
# Or change OpenClaw's port in openclaw.json:
# "gateway": { "port": 18790 }
"Unable to acquire lock" error
Cause: A previous OpenClaw process didn't exit cleanly.
Solution:
# Remove the lock file
rm -f ~/.openclaw/.next/dev/lock
# Or kill all OpenClaw processes
pkill -f "openclaw"
pnpm openclaw gateway start
Gateway starts but WebChat shows "Connection refused"
Cause: Firewall blocking the port, or gateway only listening on localhost.
Solution:
# Open the port in UFW
sudo ufw allow 18789/tcp
# Verify gateway is listening on the right interface
ss -tlnp | grep 18789
# Should show 0.0.0.0:18789, not 127.0.0.1:18789
# If only on localhost, update openclaw.json:
# "gateway": { "host": "0.0.0.0" }
Category 3: Model Configuration Errors
"API Key not set" or "Missing required environment variable"
Solution:
# Check the variable is set
printenv | grep API_KEY
# If missing, add to your .env file
echo 'MINIMAX_API_KEY=sk-your-key' >> ~/.openclaw/.env
# Reload environment
source ~/.openclaw/.env
pnpm openclaw gateway restart
"401 Unauthorized" from the model API
Cause: Invalid or expired API Key.
Solution:
- Log in to the model provider's console (platform.minimax.io, console.anthropic.com, etc.)
- Verify the key is active and hasn't been deleted
- Create a new key if needed
- Update your
.envfile and restart OpenClaw
"429 Too Many Requests" / Rate Limit
Cause: You've exceeded the API rate limit or free-tier quota.
Solution:
// Add a fallback model in openclaw.json
{
"agents": {
"defaults": {
"model": {
"primary": "minimax/MiniMax-M2.1",
"fallback": ["ollama/llama3.3", "deepseek/deepseek-v3"]
}
}
}
}
"Model not found" or "Unknown model ID"
Cause: Incorrect model ID in config.
Solution: Check the exact model ID in the provider's documentation. Common correct IDs (2026):
- MiniMax:
MiniMax-M2.1(notminimax-m2.1) - Anthropic:
claude-sonnet-4-6(notclaude-sonnet) - OpenAI:
gpt-4o(notgpt4o)
Category 4: Skills Errors
Skill installs but doesn't appear in conversation
Cause: Skill not activated or requires restart.
Solution:
# List installed skills
pnpm openclaw skills list
# Activate a specific skill
pnpm openclaw skills activate skill-name
# Restart gateway
pnpm openclaw gateway restart
"Skill execution failed" error
Cause: Skill has a bug, missing dependency, or incorrect configuration.
Solution:
# Check skill-specific logs
pnpm openclaw skills logs skill-name
# Reinstall the skill
pnpm openclaw skills uninstall skill-name
pnpm openclaw skills install skill-name
Skill uses wrong model
Cause: Skill doesn't have a model specified and is using a default that doesn't meet its requirements.
Solution: Specify the model for the skill in openclaw.json:
{
"skills": {
"image-analyzer": {
"model": "minimax/MiniMax-VL-01"
}
}
}
Category 5: Telegram Errors
Bot doesn't respond to messages
Checklist:
# 1. Is the gateway running?
pnpm openclaw gateway status
# 2. Is the Telegram channel enabled?
pnpm openclaw channels list
# 3. Is the webhook configured?
pnpm openclaw channels telegram webhook status
# 4. Re-setup the webhook
pnpm openclaw channels telegram setup-webhook
"Webhook SSL certificate error"
Cause: Telegram requires HTTPS for webhooks. HTTP is rejected.
Solution: Set up a TLS certificate for your domain using Caddy or Certbot:
# Using Caddy (easiest โ handles HTTPS automatically)
sudo apt install caddy
# In /etc/caddy/Caddyfile:
# your-domain.com {
# reverse_proxy localhost:18789
# }
sudo systemctl restart caddy
"Conflict: terminated by other getUpdates request"
Cause: Multiple OpenClaw instances are running with the same Bot Token.
Solution:
# Kill all OpenClaw processes
pkill -f "openclaw"
# Verify only one instance is running
ps aux | grep openclaw
# Restart cleanly
pnpm openclaw gateway start
Still Stuck?
If the above doesn't resolve your issue:
- Check the full error log:
pnpm openclaw logs --all > openclaw-debug.log - Search GitHub Issues: github.com/openclaw/openclaw/issues
- Ask in the community: Join the OpenClaw Discord or Telegram group (links in the GitHub repo)
- Open a new issue with your log file and
openclaw.json(with API keys redacted)