🔗 Network Bonding and Teaming in AlmaLinux: High-Speed Redundancy

Ever had your network go down and lost everything? 😱 Or wished you could combine two network cards for double the speed? That’s where network bonding comes in! I learned this the hard way when a single network cable failure took down our entire web service for 3 hours. Never again! Today I’m showing you how to set up network bonding and teaming in AlmaLinux - combine multiple network interfaces for redundancy AND performance. Let’s make your network bulletproof! 🛡️

🤔 Why Network Bonding Matters

Bonding isn’t just for backup - it’s about performance too! Here’s why you need it:

• 🔄 Automatic Failover - One link dies? Traffic continues!
• ⚡ Increased Bandwidth - Combine multiple NICs for speed
• 🎯 Load Balancing - Distribute traffic across links
• 💰 Cost Effective - Cheaper than 10Gb Ethernet
• 🏢 Enterprise Ready - Standard in data centers
• 🚀 Zero Downtime - Maintenance without interruption

True story: We bonded four 1Gb NICs and got nearly 4Gb throughput for a fraction of the cost of 10Gb equipment! 💪

🎯 What You Need

Before we start bonding networks, ensure you have:

• ✅ AlmaLinux system with 2+ network interfaces
• ✅ Root or sudo access
• ✅ Switch that supports LACP (for some modes)
• ✅ 20 minutes to bulletproof your network
• ✅ Cables to test failover (unplug them! 🔌)

📝 Step 1: Understanding Bonding vs Teaming

Let’s clarify the options first!

Network Bonding (Traditional)

# Check if bonding module is loaded
lsmod | grep bonding

# Load bonding module
sudo modprobe bonding

# Make permanent
echo "bonding" | sudo tee /etc/modules-load.d/bonding.conf

# View bonding info
cat /sys/class/net/bond0/bonding/mode
cat /proc/net/bonding/bond0

Network Teaming (Modern Alternative)

# Install teamd
sudo dnf install -y teamd NetworkManager-team

# Check team daemon
rpm -qa | grep team

# Teaming advantages:
# - Better performance
# - More features
# - Active development
# - NetworkManager integration
# - D-Bus control interface

Bonding Modes Explained

# Mode 0 (balance-rr) - Round Robin
# Packets sent in sequence across all slaves
# Provides: Load balancing + fault tolerance
# Requires: Switch configuration

# Mode 1 (active-backup) - Active/Standby
# Only one slave active at a time
# Provides: Fault tolerance only
# Requires: No special switch config

# Mode 2 (balance-xor) - XOR Hash
# Source/Dest MAC address determines slave
# Provides: Load balancing + fault tolerance
# Requires: Switch configuration

# Mode 3 (broadcast) - Broadcast
# Everything sent on all slaves
# Provides: Fault tolerance
# Requires: Very specific use cases

# Mode 4 (802.3ad) - Dynamic Link Aggregation (LACP)
# IEEE 802.3ad standard
# Provides: Best performance + fault tolerance
# Requires: LACP-capable switch

# Mode 5 (balance-tlb) - Adaptive Transmit Load Balancing
# Outgoing traffic distributed by load
# Provides: TX load balancing + fault tolerance
# Requires: No special switch config

# Mode 6 (balance-alb) - Adaptive Load Balancing
# Both TX and RX load balancing
# Provides: Full load balancing + fault tolerance
# Requires: No special switch config

🔧 Step 2: Configure Network Bonding

Let’s create a bonded interface!

Method 1: Using nmcli (NetworkManager)

# List network interfaces
nmcli device status

# Create bond interface
sudo nmcli connection add type bond \
    con-name bond0 \
    ifname bond0 \
    mode active-backup

# Or with LACP (mode 4)
sudo nmcli connection add type bond \
    con-name bond0 \
    ifname bond0 \
    mode 802.3ad

# Configure bond options
sudo nmcli connection modify bond0 \
    bond.options "mode=active-backup,miimon=100,fail_over_mac=1"

# Add slave interfaces
sudo nmcli connection add type ethernet \
    slave-type bond \
    con-name bond0-slave1 \
    ifname enp0s3 \
    master bond0

sudo nmcli connection add type ethernet \
    slave-type bond \
    con-name bond0-slave2 \
    ifname enp0s8 \
    master bond0

# Set IP address
sudo nmcli connection modify bond0 \
    ipv4.addresses 192.168.1.100/24 \
    ipv4.gateway 192.168.1.1 \
    ipv4.dns 8.8.8.8 \
    ipv4.method manual

# Activate bond
sudo nmcli connection up bond0

# Verify
nmcli connection show bond0
cat /proc/net/bonding/bond0

Method 2: Configuration Files

# Create bond interface config
sudo nano /etc/sysconfig/network-scripts/ifcfg-bond0

DEVICE=bond0
TYPE=Bond
BONDING_MASTER=yes
BOOTPROTO=none
ONBOOT=yes
IPADDR=192.168.1.100
NETMASK=255.255.255.0
GATEWAY=192.168.1.1
BONDING_OPTS="mode=1 miimon=100 fail_over_mac=1"

# Configure first slave
sudo nano /etc/sysconfig/network-scripts/ifcfg-enp0s3

DEVICE=enp0s3
TYPE=Ethernet
BOOTPROTO=none
ONBOOT=yes
MASTER=bond0
SLAVE=yes

# Configure second slave
sudo nano /etc/sysconfig/network-scripts/ifcfg-enp0s8

DEVICE=enp0s8
TYPE=Ethernet
BOOTPROTO=none
ONBOOT=yes
MASTER=bond0
SLAVE=yes

# Restart network
sudo systemctl restart NetworkManager

Advanced Bonding Options

# Configure bonding parameters
sudo nano /etc/sysconfig/network-scripts/ifcfg-bond0

# Add these options:
BONDING_OPTS="mode=802.3ad miimon=100 lacp_rate=fast xmit_hash_policy=layer3+4"

# Options explained:
# miimon=100        - Link monitoring every 100ms
# lacp_rate=fast    - Fast LACP negotiation (1s vs 30s)
# xmit_hash_policy  - How to distribute packets:
#   layer2          - MAC addresses only
#   layer2+3        - MAC + IP addresses
#   layer3+4        - IP + Port (best for servers)
# primary=enp0s3    - Preferred interface
# primary_reselect  - When to switch back to primary
# fail_over_mac=1   - Use same MAC on failover

🌟 Step 3: Configure Network Teaming

The modern approach with more features!

Create Team Interface

# Method 1: Using nmcli
sudo nmcli connection add type team \
    con-name team0 \
    ifname team0 \
    config '{"runner": {"name": "activebackup"}}'

# With load balancing
sudo nmcli connection add type team \
    con-name team0 \
    ifname team0 \
    config '{"runner": {"name": "loadbalance"}}'

# With LACP
sudo nmcli connection add type team \
    con-name team0 \
    ifname team0 \
    config '{"runner": {"name": "lacp", "active": true, "fast_rate": true}}'

# Add team slaves
sudo nmcli connection add type ethernet \
    slave-type team \
    con-name team0-slave1 \
    ifname enp0s3 \
    master team0

sudo nmcli connection add type ethernet \
    slave-type team \
    con-name team0-slave2 \
    ifname enp0s8 \
    master team0

# Configure IP
sudo nmcli connection modify team0 \
    ipv4.addresses 192.168.1.100/24 \
    ipv4.gateway 192.168.1.1 \
    ipv4.dns 8.8.8.8 \
    ipv4.method manual

# Activate team
sudo nmcli connection up team0

Advanced Team Configuration

# Create detailed team config
cat > /tmp/team-config.json << 'EOF'
{
    "device": "team0",
    "runner": {
        "name": "lacp",
        "active": true,
        "fast_rate": true,
        "tx_hash": ["eth", "vlan", "ipv4", "ipv6", "tcp", "udp"]
    },
    "link_watch": {
        "name": "ethtool"
    },
    "ports": {
        "enp0s3": {
            "prio": 100,
            "sticky": true
        },
        "enp0s8": {
            "prio": 50
        }
    }
}
EOF

# Apply config
sudo nmcli connection modify team0 \
    team.config "$(cat /tmp/team-config.json)"

# Monitor team status
sudo teamdctl team0 state
sudo teamdctl team0 config dump
sudo teamnl team0 ports

✅ Step 4: Testing and Monitoring

Let’s make sure it really works!

Test Failover

#!/bin/bash
# Test network failover

echo "🧪 Network Failover Test"
echo "========================"

# Continuous ping in background
ping -i 0.2 google.com > /tmp/ping.log &
PING_PID=$!

# Show current active interface
echo "Current active interface:"
if [ -f /proc/net/bonding/bond0 ]; then
    grep "Currently Active Slave" /proc/net/bonding/bond0
else
    sudo teamdctl team0 state | grep -A2 "active port"
fi

# Disconnect primary interface
echo ""
echo "⚠️ Disconnecting primary interface in 3 seconds..."
sleep 3
sudo ip link set enp0s3 down

# Check failover
sleep 2
echo ""
echo "After failover:"
if [ -f /proc/net/bonding/bond0 ]; then
    grep "Currently Active Slave" /proc/net/bonding/bond0
else
    sudo teamdctl team0 state | grep -A2 "active port"
fi

# Check packet loss
sleep 3
kill $PING_PID
echo ""
echo "Packet loss during failover:"
grep -c "time=" /tmp/ping.log
grep -c "Destination Host Unreachable" /tmp/ping.log

# Restore interface
sudo ip link set enp0s3 up
echo "✅ Interface restored"

Performance Testing

#!/bin/bash
# Test bonded network performance

test_bandwidth() {
    echo "📊 Bandwidth Test"
    
    # Install iperf3 if needed
    which iperf3 || sudo dnf install -y iperf3
    
    # Single interface baseline
    echo "Single interface test:"
    sudo ip link set enp0s8 down
    iperf3 -c 192.168.1.200 -t 10
    
    # Both interfaces bonded
    echo ""
    echo "Bonded interfaces test:"
    sudo ip link set enp0s8 up
    sleep 2
    iperf3 -c 192.168.1.200 -t 10 -P 4
    
    # Show interface statistics
    echo ""
    echo "Interface statistics:"
    cat /proc/net/bonding/bond0 | grep -E "Slave Interface|MII Status|Speed"
}

test_bandwidth

🎮 Quick Examples

Example 1: HA Web Server Setup 🌐

#!/bin/bash
# High availability network for web server

setup_ha_network() {
    echo "🌐 Setting up HA Network for Web Server"
    
    # Create active-backup bond
    sudo nmcli connection add type bond \
        con-name ha-bond \
        ifname bond0 \
        mode active-backup \
        bond.options "miimon=100,fail_over_mac=active,primary=enp0s3"
    
    # Add interfaces
    for iface in enp0s3 enp0s8; do
        sudo nmcli connection add type ethernet \
            slave-type bond \
            con-name ha-bond-$iface \
            ifname $iface \
            master bond0
    done
    
    # Configure static IP
    sudo nmcli connection modify ha-bond \
        ipv4.addresses 192.168.1.10/24 \
        ipv4.gateway 192.168.1.1 \
        ipv4.dns "8.8.8.8,8.8.4.4" \
        ipv4.method manual \
        connection.autoconnect yes
    
    # Activate
    sudo nmcli connection up ha-bond
    
    # Configure monitoring
    cat > /usr/local/bin/monitor-bond.sh << 'SCRIPT'
#!/bin/bash
while true; do
    if ! grep -q "up" /sys/class/net/bond0/operstate; then
        echo "ALERT: Bond interface down!" | \
            mail -s "Network Alert on $(hostname)" [email protected]
    fi
    sleep 10
done
SCRIPT
    
    chmod +x /usr/local/bin/monitor-bond.sh
    
    # Create systemd service
    cat > /etc/systemd/system/bond-monitor.service << 'SERVICE'
[Unit]
Description=Bond Network Monitor
After=network-online.target

[Service]
Type=simple
ExecStart=/usr/local/bin/monitor-bond.sh
Restart=always

[Install]
WantedBy=multi-user.target
SERVICE
    
    sudo systemctl enable --now bond-monitor
    
    echo "✅ HA Network configured!"
    echo "📊 Status: $(cat /proc/net/bonding/bond0 | grep 'Currently Active')"
}

setup_ha_network

Example 2: Load-Balanced Database Server 💾

#!/bin/bash
# LACP bond for database server

setup_db_network() {
    echo "💾 Configuring LACP Bond for Database"
    
    # Create LACP team (better than bonding for LACP)
    cat > /tmp/lacp-team.json << 'EOF'
{
    "runner": {
        "name": "lacp",
        "active": true,
        "fast_rate": true,
        "tx_hash": ["ipv4", "tcp", "udp"],
        "agg_select_policy": "bandwidth"
    },
    "link_watch": {
        "name": "ethtool",
        "delay_up": 5,
        "delay_down": 1
    }
}
EOF
    
    # Create team interface
    sudo nmcli connection add type team \
        con-name db-team \
        ifname team0 \
        team.config "$(cat /tmp/lacp-team.json)"
    
    # Add 4 interfaces for 4Gb aggregate
    for i in {3..6}; do
        sudo nmcli connection add type ethernet \
            slave-type team \
            con-name db-team-enp0s$i \
            ifname enp0s$i \
            master team0
    done
    
    # Configure IP and Jumbo frames
    sudo nmcli connection modify db-team \
        ipv4.addresses 10.0.1.50/24 \
        ipv4.method manual \
        802-3-ethernet.mtu 9000
    
    # Apply MTU to slaves
    for i in {3..6}; do
        sudo nmcli connection modify db-team-enp0s$i \
            802-3-ethernet.mtu 9000
    done
    
    # Activate
    sudo nmcli connection up db-team
    
    # Performance tuning
    cat > /etc/sysctl.d/99-db-network.conf << 'EOF'
# Database network optimization
net.core.rmem_max = 134217728
net.core.wmem_max = 134217728
net.ipv4.tcp_rmem = 4096 87380 134217728
net.ipv4.tcp_wmem = 4096 65536 134217728
net.core.netdev_max_backlog = 30000
net.ipv4.tcp_congestion_control = bbr
net.ipv4.tcp_mtu_probing = 1
EOF
    
    sudo sysctl -p /etc/sysctl.d/99-db-network.conf
    
    echo "✅ Database network optimized!"
    sudo teamdctl team0 state
}

setup_db_network

Example 3: Dynamic Failover Manager 🔄

#!/bin/bash
# Intelligent network failover manager

cat > /usr/local/bin/smart-failover.py << 'EOF'
#!/usr/bin/env python3
import subprocess
import time
import json
import smtplib
from email.mime.text import MIMEText

class NetworkManager:
    def __init__(self, interface="bond0"):
        self.interface = interface
        self.history = []
        
    def get_bond_status(self):
        """Get current bond status"""
        try:
            with open(f"/proc/net/bonding/{self.interface}") as f:
                return f.read()
        except:
            return None
    
    def get_active_slave(self):
        """Get currently active slave"""
        status = self.get_bond_status()
        if status:
            for line in status.split('\n'):
                if 'Currently Active Slave' in line:
                    return line.split(':')[1].strip()
        return None
    
    def get_slave_stats(self, slave):
        """Get statistics for a slave interface"""
        try:
            stats = {}
            with open(f"/sys/class/net/{slave}/statistics/rx_errors") as f:
                stats['rx_errors'] = int(f.read().strip())
            with open(f"/sys/class/net/{slave}/statistics/tx_errors") as f:
                stats['tx_errors'] = int(f.read().strip())
            
            # Get speed and duplex
            result = subprocess.run(['ethtool', slave], 
                                  capture_output=True, text=True)
            for line in result.stdout.split('\n'):
                if 'Speed:' in line:
                    stats['speed'] = line.split(':')[1].strip()
                if 'Duplex:' in line:
                    stats['duplex'] = line.split(':')[1].strip()
            
            return stats
        except:
            return None
    
    def check_health(self):
        """Check network health and make decisions"""
        slaves = self.get_all_slaves()
        active = self.get_active_slave()
        
        issues = []
        
        for slave in slaves:
            stats = self.get_slave_stats(slave)
            if stats:
                # Check for errors
                if stats.get('rx_errors', 0) > 100:
                    issues.append(f"{slave}: High RX errors ({stats['rx_errors']})")
                
                if stats.get('tx_errors', 0) > 100:
                    issues.append(f"{slave}: High TX errors ({stats['tx_errors']})")
                
                # Check link speed
                if 'speed' in stats and '1000Mb' not in stats['speed']:
                    issues.append(f"{slave}: Degraded speed ({stats['speed']})")
        
        if issues:
            self.alert(f"Network issues detected:\n" + "\n".join(issues))
            
            # Force failover if active slave has issues
            if active and any(active in issue for issue in issues):
                self.force_failover(active)
    
    def force_failover(self, current_slave):
        """Force failover to backup interface"""
        print(f"⚠️ Forcing failover from {current_slave}")
        subprocess.run(['sudo', 'ip', 'link', 'set', current_slave, 'down'])
        time.sleep(2)
        subprocess.run(['sudo', 'ip', 'link', 'set', current_slave, 'up'])
        
        new_active = self.get_active_slave()
        print(f"✅ Failover complete. New active: {new_active}")
        self.alert(f"Forced failover from {current_slave} to {new_active}")
    
    def get_all_slaves(self):
        """Get all slave interfaces"""
        slaves = []
        status = self.get_bond_status()
        if status:
            for line in status.split('\n'):
                if 'Slave Interface:' in line:
                    slaves.append(line.split(':')[1].strip())
        return slaves
    
    def alert(self, message):
        """Send alert notification"""
        print(f"🚨 ALERT: {message}")
        # Add email/Slack/monitoring integration here
    
    def monitor_loop(self):
        """Main monitoring loop"""
        print(f"👀 Monitoring {self.interface}...")
        
        while True:
            try:
                self.check_health()
                
                # Log current status
                active = self.get_active_slave()
                if active:
                    stats = self.get_slave_stats(active)
                    print(f"✅ Active: {active} | Speed: {stats.get('speed', 'Unknown')}")
                
                time.sleep(10)
                
            except KeyboardInterrupt:
                print("\n👋 Monitoring stopped")
                break
            except Exception as e:
                print(f"❌ Error: {e}")
                time.sleep(30)

if __name__ == "__main__":
    manager = NetworkManager("bond0")
    manager.monitor_loop()
EOF

chmod +x /usr/local/bin/smart-failover.py
echo "✅ Smart failover manager installed"

🚨 Fix Common Problems

Problem 1: Bond Not Working ❌

Interfaces not bonding?

# Check module loaded
lsmod | grep bonding

# Load if missing
sudo modprobe bonding

# Check slave status
cat /proc/net/bonding/bond0

# Fix slave configuration
sudo ip link set enp0s3 down
sudo ip link set enp0s3 master bond0
sudo ip link set enp0s3 up

# Restart NetworkManager
sudo systemctl restart NetworkManager

Problem 2: No Increased Speed ❌

Not getting aggregate bandwidth?

# Check bonding mode
cat /sys/class/net/bond0/bonding/mode

# Mode must support aggregation (0,2,4,5,6)
# Active-backup (mode 1) does NOT increase bandwidth

# For LACP, check switch configuration
# Switch must have LACP/802.3ad enabled

# Test with multiple streams
iperf3 -c server -P 4  # 4 parallel streams

Problem 3: Failover Not Working ❌

Traffic stops on link failure?

# Check monitoring interval
cat /sys/class/net/bond0/bonding/miimon

# Should be 100 (100ms)
echo 100 | sudo tee /sys/class/net/bond0/bonding/miimon

# Check fail_over_mac setting
cat /sys/class/net/bond0/bonding/fail_over_mac

# Test failover
sudo ip link set enp0s3 down
ping -c 10 google.com  # Should continue working

Problem 4: Switch Compatibility ❌

LACP not negotiating?

# Check LACP status
cat /proc/net/bonding/bond0 | grep -A10 "802.3ad"

# Switch-side config needed (Cisco example):
# interface Port-channel1
#  switchport mode access
#  switchport access vlan 100
# interface GigabitEthernet0/1
#  channel-group 1 mode active
# interface GigabitEthernet0/2
#  channel-group 1 mode active

# Try simpler mode
sudo nmcli connection modify bond0 \
    bond.options "mode=balance-alb"

📋 Simple Commands Summary

💡 Tips for Success

1. Test Failover 🧪 - Actually unplug cables
2. Match Switch Config 🔌 - LACP needs both sides
3. Use Right Mode 🎯 - active-backup for simple HA
4. Monitor Continuously 📊 - Set up alerts
5. Document Setup 📝 - For future troubleshooting
6. Benchmark First 📈 - Know your baseline

Funny story: We once spent hours troubleshooting “broken” LACP… turns out the network team configured the wrong switch ports. Always verify both ends! 😅

🏆 What You Learned

You’re now a network bonding expert! You can:

• ✅ Configure network bonding and teaming
• ✅ Set up different bonding modes
• ✅ Implement automatic failover
• ✅ Achieve load balancing
• ✅ Test and verify redundancy
• ✅ Troubleshoot bonding issues
• ✅ Monitor bond health

🎯 Why This Matters

Network bonding provides:

• 🛡️ Bulletproof connectivity
• ⚡ Increased bandwidth
• 💰 Cost-effective redundancy
• 🔄 Zero-downtime maintenance
• 🏢 Enterprise reliability
• 😴 Peace of mind

Last month, a client’s primary network port died during Black Friday sales. The bonded backup took over instantly - zero downtime, zero lost sales. They made $50K that day and didn’t even know there was a problem until Monday! 💪

Remember: In networking, redundancy isn’t optional - it’s essential. One cable failure shouldn’t take down your service! 🔗

Happy bonding! May your networks be redundant and your failovers seamless! 🌐✨