Connection Problems

This guide helps resolve network and connection-related issues with arb-assist.

RPC Connection Issues

Connection Refused

Error:

Error: Connection refused to RPC endpoint

Common Causes:

1. Incorrect URL format

2. Firewall blocking connection

3. RPC endpoint down

4. Invalid authentication

Troubleshooting Steps:

1. Verify URL format:

# Correct formats
rpc_url = "https://api.mainnet-beta.solana.com"
rpc_url = "http://localhost:8899"
rpc_url = "https://mainnet.helius-rpc.com/?api-key=YOUR_KEY"

# Common mistakes
rpc_url = "api.mainnet-beta.solana.com"  # Missing protocol
rpc_url = "https://api.mainnet-beta.solana.com/"  # Trailing slash

1. Test connection:

# Basic connectivity test
curl -I https://api.mainnet-beta.solana.com

# Test RPC method
curl -X POST -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"getHealth"}' \
  YOUR_RPC_URL

1. Check firewall:

# Check if port is open
nc -zv api.mainnet-beta.solana.com 443

# Check firewall rules
sudo ufw status

Rate Limiting

Error:

Error: 429 Too Many Requests

Solutions:

1. Use recommended RPC:

# Helius (cheaper option, fine for starting out)
rpc_url = "https://mainnet.helius-rpc.com/?api-key=YOUR_KEY"

# NFT-gated shared nodes (better performance)
rpc_url = "http://deez.top:80"
rpc_url = "http://shark.top:80"
rpc_url = "http://platinum.top:80"

1. Implement request distribution:

# Multiple RPCs for different purposes
rpc_url = "https://read-rpc.com"  # For data fetching
config_rpc_url = "https://bot-rpc.com"  # For bot config

sending_rpc_urls = [  # For transaction sending
  "https://send1.com",
  "https://send2.com",
  "https://send3.com",
]

1. Reduce request frequency:

update_interval = 30000  # 30 seconds instead of 10

SSL/TLS Errors

Error:

Error: SSL certificate problem

Solutions:

1. Update certificates:

sudo apt update
sudo apt install ca-certificates
sudo update-ca-certificates

1. Check system time:

# Incorrect time can cause SSL errors
timedatectl status
sudo timedatectl set-ntp true

1. Use HTTP if appropriate (local/private networks only):

rpc_url = "http://10.0.0.5:8899"  # Internal RPC

GRPC Connection Issues

Failed to Connect

Error:

Error: Failed to connect to GRPC endpoint

Debugging Steps:

1. Verify GRPC configuration:

# Yellowstone example
grpc_url = "http://grpc.example.com:10001"  # Note: often HTTP not HTTPS
grpc_token = ""  # May not need token
grpc_engine = "yellowstone"

# ThorStreamer example
grpc_url = "https://thor.example.com/stream"
grpc_token = "eyJ0eXAiOiJKV1QiLCJhbGc..."  # Required token
grpc_engine = "thor"

1. Test GRPC connectivity:

# For HTTP/2
curl -I http://grpc.example.com:10001

# Check if port is open
telnet grpc.example.com 10001

1. Verify authentication:

   • Check token hasn't expired

   • Ensure token is correctly formatted

   • Verify engine type matches provider

Stream Interruptions

Error:

Error: GRPC stream disconnected

Common Causes:

1. Network instability

2. Provider-side timeout

3. Resource limits

4. Authentication expiry

Solutions:

1. Implement reconnection logic (arb-assist handles this automatically)

2. Monitor connection stability:

# Continuous ping test
ping -i 1 grpc.example.com | ts

# Monitor packet loss
mtr grpc.example.com

1. Check provider status:

   • Visit provider status page

   • Check Discord/support channels

   • Try alternative endpoints

Authentication Failures

Error:

Error: GRPC authentication failed

Solutions:

1. Verify token format:

# JWT token (common)
grpc_token = "eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9..."

# API key
grpc_token = "sk_live_abcd1234..."

# Empty (some providers)
grpc_token = ""

1. Check token permissions:

   • Ensure token has streaming permissions

   • Verify IP whitelist if applicable

   • Check rate limits on token

2. Token rotation:

   • Some providers rotate tokens

   • Update token in config

   • Restart arb-assist

Network Configuration

Firewall Issues

Symptoms:

• All external connections fail

• Local connections work

• Timeout errors

Solutions:

1. Check outbound rules:

# UFW (Ubuntu Firewall)
sudo ufw status verbose

# Allow outbound HTTPS
sudo ufw allow out 443/tcp

# Allow outbound HTTP
sudo ufw allow out 80/tcp

# Allow GRPC ports (common range)
sudo ufw allow out 10000:10010/tcp

1. iptables rules:

# List current rules
sudo iptables -L -n -v

# Allow outbound connections
sudo iptables -A OUTPUT -p tcp --dport 443 -j ACCEPT
sudo iptables -A OUTPUT -p tcp --dport 80 -j ACCEPT

1. Cloud provider firewalls:

   • AWS: Check Security Groups

   • GCP: Check VPC firewall rules

   • Azure: Check Network Security Groups

DNS Resolution

Error:

Error: Failed to resolve hostname

Solutions:

1. Check DNS configuration:

# View current DNS servers
cat /etc/resolv.conf

# Test DNS resolution
nslookup api.mainnet-beta.solana.com
dig api.mainnet-beta.solana.com

1. Add reliable DNS servers:

# Edit resolv.conf
sudo nano /etc/resolv.conf

# Add these lines
nameserver 8.8.8.8      # Google DNS
nameserver 1.1.1.1      # Cloudflare DNS
nameserver 8.8.4.4      # Google DNS backup

1. Fix systemd-resolved (Ubuntu):

# Disable and reconfigure
sudo systemctl disable systemd-resolved
sudo systemctl stop systemd-resolved

# Create static resolv.conf
sudo rm /etc/resolv.conf
sudo bash -c 'echo "nameserver 8.8.8.8" > /etc/resolv.conf'

Proxy Configuration

If behind a corporate proxy:

1. System proxy settings:

export HTTP_PROXY=http://proxy.company.com:8080
export HTTPS_PROXY=http://proxy.company.com:8080
export NO_PROXY=localhost,127.0.0.1

1. Application-specific proxy:

# Some configurations may support
proxy_url = "http://proxy.company.com:8080"

Connection Pool Management

Too Many Connections

Symptoms:

• Connection refused after running for a while

• "Too many connections" errors

• Resource exhaustion

Solutions:

1. Reuse connections:

# For NotArb - use same RPC for multiple purposes
[notarb]
token_accounts_checker = { rpc_url = "" }  # Empty = use main RPC
blockhash_updater = { rpc_url = "" }

1. Connection limits:

# Check current limits
ulimit -n

# Increase if needed
ulimit -n 65536

1. Monitor connections:

# Count connections to RPC
netstat -an | grep :443 | grep ESTABLISHED | wc -l

# Watch connections over time
watch 'netstat -an | grep :443 | grep ESTABLISHED | wc -l'

Latency Optimization

High Latency Issues

Symptoms:

• Slow responses

• Timeouts despite connectivity

• Missed opportunities

Solutions:

1. Geographic optimization:

# Test latency to endpoints
ping -c 10 api.mainnet-beta.solana.com
ping -c 10 api.mainnet-beta.solana.com | grep avg

# Traceroute to identify slow hops
traceroute api.mainnet-beta.solana.com

1. Use regional endpoints:

# Choose endpoints near your server
# US East
rpc_url = "https://us-east.rpc.com"

# Europe
rpc_url = "https://eu.rpc.com"

# Asia
rpc_url = "https://asia.rpc.com"

1. Connection timing:

# Measure connection time
time curl -I https://api.mainnet-beta.solana.com

# Detailed timing
curl -w "@curl-format.txt" -o /dev/null -s https://api.mainnet-beta.solana.com

Create curl-format.txt:

time_namelookup:  %{time_namelookup}s\n
time_connect:  %{time_connect}s\n
time_appconnect:  %{time_appconnect}s\n
time_pretransfer:  %{time_pretransfer}s\n
time_redirect:  %{time_redirect}s\n
time_starttransfer:  %{time_starttransfer}s\n
time_total:  %{time_total}s\n

Connection Monitoring

Setup Monitoring

1. Create connection monitor script:

#!/bin/bash
# monitor-connections.sh

while true; do
    echo "=== $(date) ==="
    
    # Test RPC
    curl -s -o /dev/null -w "RPC Response: %{http_code} Time: %{time_total}s\n" \
        -X POST -H "Content-Type: application/json" \
        -d '{"jsonrpc":"2.0","id":1,"method":"getHealth"}' \
        YOUR_RPC_URL
    
    # Test GRPC
    timeout 5 nc -zv YOUR_GRPC_HOST YOUR_GRPC_PORT &>/dev/null
    if [ $? -eq 0 ]; then
        echo "GRPC: Connected"
    else
        echo "GRPC: Failed"
    fi
    
    # Count connections
    echo "Active connections: $(netstat -an | grep ESTABLISHED | wc -l)"
    
    echo ""
    sleep 60
done

1. Run monitor:

chmod +x monitor-connections.sh
./monitor-connections.sh > connection-monitor.log 2>&1 &

Alert on Failures

Setup alerts for connection issues:

#!/bin/bash
# alert-on-failure.sh

check_connection() {
    curl -s -f -X POST -H "Content-Type: application/json" \
        -d '{"jsonrpc":"2.0","id":1,"method":"getHealth"}' \
        YOUR_RPC_URL > /dev/null
    return $?
}

if ! check_connection; then
    echo "RPC connection failed at $(date)" | mail -s "arb-assist Alert" [e-mail]
    # Or send to Discord webhook, Telegram, etc.
fi

Best Practices

1. Use multiple endpoints: Always have fallbacks

2. Monitor continuously: Catch issues early

3. Document endpoints: Keep a list of working endpoints

4. Test before deploying: Verify all connections work

5. Have emergency contacts: Provider support, Discord channels

6. Keep credentials secure: Use environment variables or secure storage