22

✏️ Editing: load_balancing.nft

#!/usr/sbin/nft -f

# This example file shows how to implement load balancing using the nftables
# framework.
# This script is meant to be loaded with `nft -f <file>`
# You require linux kernel >= 4.12 and nft >= 0.7
# For up-to-date information please visit https://wiki.nftables.org

flush ruleset

table ip nat {
	chain prerouting {
		type nat hook prerouting priority -300;
		# round-robing load balancing between the 2 IPv4 addresses:
		dnat to numgen inc mod 2 map {
				0 : 192.168.10.100, \
				1 : 192.168.20.200 }
		# emulate flow distribution with different backend weights using intervals:
		dnat to numgen inc mod 10 map {
				0-5 : 192.168.10.100, \
				6-9 : 192.168.20.200 }
		# tcp port based distribution is also possible:
		ip protocol tcp dnat to 192.168.1.100 : numgen inc mod 2 map {
				0 : 4040 ,\
				1 : 4050 }
		# consistent hash-based distribution:
		dnat to jhash ip saddr . tcp dport mod 2 map {
				0 : 192.168.20.100, \
				1 : 192.168.30.100 }
	}
}

table ip raw {
	chain prerouting {
		type filter hook prerouting priority -300;
		# using stateless NAT, round-robing distribution (you could use hashing too):
		tcp dport 80 notrack ip daddr set numgen inc mod 2 map { 0 : 192.168.1.100, 1 : 192.168.1.101 }
	}
}

table netdev mytable {
	chain ingress {
		# mind the NIC devices, they must exist in the system
		type filter hook ingress device eth0 priority 0;
		# using Direct Server Return (DSR), connectionless approach:
		udp dport 53 ether saddr set aa:bb:cc:dd:ff:ee ether daddr set numgen inc mod 2 map {
					0 : aa:aa:aa:aa:aa:aa,
					1 : bb:bb:bb:bb:bb:bb } fwd to eth1
		# using Direct Server Return (DSR), connection-oriented flows:
		 tcp dport 80 ether saddr set aa:bb:cc:dd:ff:ee ether daddr set jhash ip saddr . tcp sport mod 2 map {
					0 : aa:aa:aa:aa:aa:aa,
					1 : bb:bb:bb:bb:bb:bb } fwd to eth1
	}
}

MR SKB