Skip to main content
Mole fixing issues

Common Issues

Quick solutions for frequently encountered problems.

Agent Won't Start​

Port Already in Use​

Error: listen tcp 0.0.0.0:4433: bind: address already in use

Solution:

# Find what's using the port
lsof -i :4433
netstat -tlnp | grep 4433

# Kill the process or use a different port
kill <PID>

# Or change config
listeners:
  - address: "0.0.0.0:4434"   # Different port

Configuration Error​

Error: invalid configuration: ...

Solution:

# Validate YAML syntax
yamllint config.yaml

# Check for common issues:
# - Incorrect indentation
# - Missing quotes around values with special chars
# - Invalid CIDR notation
# - Missing required fields

Certificate Not Found​

Error: open ./certs/agent.crt: no such file or directory

Solution:

# Generate certificates
muti-metroo cert ca -n "My CA" -o ./certs
muti-metroo cert agent -n "agent-1" \
  --ca ./certs/ca.crt \
  --ca-key ./certs/ca.key \
  -o ./certs

# Or fix path in config
tls:
  cert: "/absolute/path/to/agent.crt"
  key: "/absolute/path/to/agent.key"

Permission Denied​

Error: open ./data/agent_id: permission denied

Solution:

# Fix ownership
chown -R $(whoami) ./data

# Or fix permissions
chmod 700 ./data

# If running as service user
chown -R muti-metroo:muti-metroo /var/lib/muti-metroo

Permission Denied for Agent Keys​

Error: failed to read private key: open /opt/muti-metroo/data/agent_key: permission denied

This often happens when the agent was previously run as root or a different user.

Solution:

# Check current ownership
ls -la /opt/muti-metroo/data/

# Fix ownership (replace 'andris' with your user)
sudo chown andris:andris /opt/muti-metroo/data/agent_key
sudo chown andris:andris /opt/muti-metroo/data/agent_key.pub

# Or fix for all data files
sudo chown -R andris:andris /opt/muti-metroo/data/

Stale Process from Previous Run​

Error: bind: address already in use

A previous instance may still be running.

Solution:

# Find and kill stale processes
pgrep -la muti-metroo
pkill -f muti-metroo

# Verify ports are free
lsof -i :4433
lsof -i :1080

# On Linux, check for zombie listeners
ss -tlnp | grep muti

Connection Issues​

Peer Won't Connect​

Symptoms:

  • peers: 0 in health check
  • Logs show connection attempts but no success

Solutions:

  1. Check network connectivity:

    # Can you reach the peer?
    nc -zv peer-address 4433
    ping peer-address

  2. Check firewall:

    # Is port open?
    sudo iptables -L -n | grep 4433

  3. Check TLS:

    # Verify certificate
    openssl s_client -connect peer-address:4433

  4. Check peer ID:

    # Is the ID correct?
    peers:
      - id: "correct-peer-id..."  # Must match peer's agent_id

Connection Refused​

Error: connection refused to peer-address:4433

Solutions:

  1. Peer not running or not listening
  2. Wrong address/port
  3. Firewall blocking connection
  4. Wrong transport type

TLS Handshake Failed​

Error: tls: handshake failure

Solutions:

  1. Certificate not signed by trusted CA:

    openssl verify -CAfile ca.crt peer.crt

  2. Certificate expired:

    openssl x509 -enddate -noout -in agent.crt

  3. Wrong hostname/IP in certificate:

    openssl x509 -text -noout -in agent.crt | grep -A1 "Subject Alternative Name"

SOCKS5 Issues​

No Route to Host​

Error: no route to 1.2.3.4

Cause: No exit agent with matching route.

Solutions:

  1. Check exit is running and connected:

    curl http://localhost:8080/healthz | jq '.peers, .routes'

  2. Check exit has route configured:

    exit:
      enabled: true
      routes:
        - "0.0.0.0/0"  # Or specific CIDR

  3. Trigger route advertisement:

    curl -X POST http://exit-agent:8080/routes/advertise

Connection Timeout​

Error: connection timeout

Causes:

  • Slow network
  • Too many hops
  • Exit agent overloaded

Solutions:

  1. Increase timeout:

    limits:
      stream_open_timeout: 60s  # Default 30s

  2. Reduce hop count

  3. Check exit agent health

Authentication Failed​

Error: SOCKS5 authentication failed

Solutions:

  1. Check username/password:

    curl -x socks5://user:password@localhost:1080 https://example.com

  2. Verify password hash in config is correct

  3. Check for special characters needing escaping

Exit Issues​

DNS Resolution Failed​

Error: DNS lookup failed for example.com

Solutions:

  1. Check DNS servers are configured:

    exit:
      dns:
        servers:
          - "8.8.8.8:53"

  2. Check DNS servers are reachable from exit host:

    dig @8.8.8.8 example.com

  3. Increase DNS timeout:

    exit:
      dns:
        timeout: 10s

Destination Not Allowed​

Error: destination not in allowed routes

Solution: Add route to exit config:

exit:
  routes:
    - "0.0.0.0/0"  # Or specific CIDR matching destination

Shell Issues​

Command Rejected​

Error: command not in whitelist

Solution: Add command to whitelist:

shell:
  whitelist:
    - bash
    - your-command  # Add needed command

Shell Authentication Failed​

Error: shell authentication failed

Solutions:

  1. Check password:

    muti-metroo shell -p correct-password agent-id bash

  2. Verify password hash in config matches

File Transfer Issues​

Path Not Allowed​

Error: path not in allowed paths

Solution: Add path prefix to allowed_paths:

file_transfer:
  allowed_paths:
    - /tmp
    - /path/to/your/files

File Too Large​

Error: file exceeds maximum size

Solution: Increase max_file_size or use 0 for unlimited:

file_transfer:
  max_file_size: 0  # Unlimited

Performance Issues​

See Performance Troubleshooting for:

  • Slow connections
  • High latency
  • Memory usage
  • CPU usage

Windows-Specific Issues​

Port 8080 Already in Use​

Windows commonly has services using port 8080. You'll see:

Error: listen tcp :8080: bind: Only one usage of each socket address

Solution: Use a different port:

http:
  address: ":8083"  # Or another free port

Check what's using the port:

netstat -anb | findstr ":8080"

Running Agent in Background on Windows​

Unlike Linux, Windows doesn't support nohup. Use PowerShell:

# Start agent in background
Start-Process -FilePath "C:\muti-metroo\muti-metroo.exe" `
  -ArgumentList "run","-c","C:\muti-metroo\config.yaml" `
  -WorkingDirectory "C:\muti-metroo" `
  -WindowStyle Hidden

# Verify it's running
tasklist | findstr muti-metroo

# Stop the agent
taskkill /IM muti-metroo.exe /F

For persistent background operation, install as a Windows service:

muti-metroo.exe service install -c C:\muti-metroo\config.yaml

Firewall Blocking Connections​

Windows Firewall may block inbound or outbound connections.

# Check if Windows Firewall is blocking
netsh advfirewall firewall show rule name=all | findstr muti

# Allow inbound on specific port
netsh advfirewall firewall add rule name="Muti Metroo" `
  dir=in action=allow protocol=tcp localport=3000

# Or allow the executable
netsh advfirewall firewall add rule name="Muti Metroo" `
  dir=in action=allow program="C:\muti-metroo\muti-metroo.exe"

Path Issues with Backslashes​

Windows paths use backslashes. In YAML, either escape or use forward slashes:

# Option 1: Escaped backslashes
data_dir: "C:\\muti-metroo\\data"
tls:
  cert: "C:\\muti-metroo\\certs\\agent.crt"

# Option 2: Forward slashes (also works)
data_dir: "C:/muti-metroo/data"

Multi-Agent Deployment​

Start Order Matters​

When deploying multiple agents, start them in the correct order:

  1. Listeners first - Agents that accept connections (have listeners: configured)
  2. Dialers second - Agents that connect outbound (have peers: configured)

Example for a chain A -> B -> C:

  1. Start Agent C (has listener, exit role)
  2. Start Agent B (has listener, connects to C)
  3. Start Agent A (connects to B, SOCKS5 ingress)

Peer Not Connecting After Restart​

After restarting a listener agent, dialers may take time to reconnect due to exponential backoff.

Speed up reconnection:

# Restart the dialer agent
pkill -f muti-metroo
./muti-metroo run -c config.yaml

# Or trigger faster reconnect by reducing max_delay
connections:
  reconnect:
    initial_delay: 500ms
    max_delay: 10s

Routes Not Propagating​

After starting agents, routes may take time to propagate (up to advertise_interval).

Trigger immediate advertisement:

# On exit agent
curl -X POST http://localhost:8080/routes/advertise

# Wait a few seconds for propagation
sleep 5

# Check routes on ingress agent
curl http://ingress-agent:8080/healthz | jq '.routes'

Verifying Multi-Agent Topology​

Check the full topology from the dashboard API:

# List all known nodes
curl -s http://localhost:8080/api/nodes | jq '.nodes[] | {id: .short_id, name: .display_name, connected: .is_connected}'

# Check route paths
curl -s http://localhost:8080/api/dashboard | jq '.routes[] | {network, origin, hops: .hop_count, path: .path_display}'

Upgrading Agents​

When deploying new binary versions:

# 1. Stop the agent
pkill -f muti-metroo  # Linux/macOS
taskkill /IM muti-metroo.exe /F  # Windows

# 2. Replace binary
cp muti-metroo-linux-amd64 /opt/muti-metroo/muti-metroo
chmod +x /opt/muti-metroo/muti-metroo

# 3. Start agent
cd /opt/muti-metroo && ./muti-metroo run -c config.yaml

# 4. Verify
curl http://localhost:8080/healthz

For remote agents via SSH:

# Upload new binary
scp muti-metroo-linux-amd64 user@remote:/opt/muti-metroo/muti-metroo

# Stop, set permissions, restart
ssh user@remote "pkill -f muti-metroo; chmod +x /opt/muti-metroo/muti-metroo"
ssh user@remote "cd /opt/muti-metroo && nohup ./muti-metroo run -c config.yaml > agent.log 2>&1 &"

Deployment Verification Checklist​

After deploying agents, verify each component:

# 1. Check each agent is running
curl http://agent-a:8080/healthz  # Should show running: true

# 2. Check peer connections
curl http://agent-a:8080/healthz | jq '.peer_count'  # Should be > 0

# 3. Check routes exist
curl http://agent-a:8080/healthz | jq '.route_count'  # Should match expected routes

# 4. Test SOCKS5 proxy (if applicable)
curl --socks5 127.0.0.1:1080 https://ifconfig.me/ip

# 5. Check all nodes visible
curl http://agent-a:8080/api/nodes | jq '.nodes | length'

Quick health check script:

#!/bin/bash
for agent in "localhost:8080" "remote1:8080" "remote2:8082"; do
  echo "=== $agent ==="
  curl -s "http://$agent/healthz" | jq '{running, peers: .peer_count, routes: .route_count}'
done

Debug Mode​

Enable debug logging for detailed diagnostics:

muti-metroo run -c config.yaml --log-level debug

Or in config:

agent:
  log_level: "debug"

Getting Help​

If these solutions don't work:

  1. Check logs with debug level
  2. Review the specific troubleshooting guides
  3. Check the protocol and limits documentation
  4. Search existing issues
  5. Open a new issue with:
    • Configuration (redacted)
    • Logs (debug level)
    • Steps to reproduce

Next Steps​