Iptables: Explanations, options & examples

Configuration file location and command commands

Configuration file

The iptables configuration file is located at: /etc/sysconfig/iptables. The contains all the tables, chains and rules. Additional configurations can be added directly to this file or via command line tools, for example:

[root@srv1 ~]# iptables -A INPUT -i eth0 -j ACCEPT

Common Commands

# start iptables
[root@srv1 ~]# service iptables start	

# get current status
[root@srv1 ~]# service iptables status

# stop iptables
[root@srv1 ~]# service iptables stop

# restart iptables
[root@srv1 ~]# service iptables restart

# save any newly added rules
[root@srv1 ~]# service iptables save

Structure of /etc/sysconfig/iptables

Tables

iptables includes 3 default tables:

• *filter - Default table for filtering packets
• *nat - Default table for Network Address Translation
• *mangle - Default table used for specific type of packet alteration

Chains

Each table has a group of built-in chains, corresponding to the actions to be performed on the packets. The chains for each section are as follows:

• The built-in chains for the filter table:
  • INPUT - Applies to packets targeted at the host (incoming traffic)
  • OUTPUT - Applies to locally-generated packets heading out of the system (outgoing traffic)
  • FORWARD - Applies to packets routed through the host (forwarded/routed traffic)
• The built-in chains for the nat table:
  • PREROUTING - Alters packets when they arrive
  • OUTPUT - Alters locally-general packets before they leave
  • POSTROUTING - Alters packets before they leave
• The built-in chains for the mangle table:
  • INPUT - Alters packets targeted for the host
  • OUTPUT - Alters locally-generated packets before they leave
  • FORWARD - Alters to packets routed through the host
  • PREROUTING - Alters incoming packets before they are routed
  • POSTROUTING- Alters packets before they leave

Explanation of an example iptables rule

[root@srv1 ~]# iptables -A INPUT -i virbr0 -p udp -m udp --dport 53 -j ACCEPT

• -A INPUT = Append the rule to the INPUT chain
• -i virbr0 = Interface = virbr0
• -p udp = Protocol = UDP
• -m udp = match = UDP - not sure what this means!
• --dport 53 = desintation port = 53
• -j ACCEPT = Jump to the target of ACCEPT. Basically, what to do if the packet matches the rule criteria. Option could be to ACCEPT, DENY or jump to another chain.

This rule will basically accept UDP traffic on port 53 across virbr0. In practice, this rule will exist in the *filter table and will allow DNS operations over the KVM virtual network interface known as virbr0.

Example operations with iptables

Open up specific ports (port 80 in this example)

[root@srv1 ~]# iptables -A INPUT -p tcp -m state --state NEW -m tcp --dport 80 -j ACCEPT

This rule needs to be in the *filter table. It will allow TCP traffic over port 80.

NB - The --state NEW signifies the packet is establishing a new connection. ESTABLISHSED would mean the packet is associated with an existing connected which has seen packet transfer in both directions. RELATED would mean a packet is establishing a new connection but is associated with an existing connections (useful for FTP).

Traffic forwarding over bridged interface

[root@srv1 ~]# iptables -I FORWARD -m physdev --physdev-is-bridged -j ACCEPT

This rule will be inserted into the FORWARD chain and will all packed to be forward across the bridged network adapter

NB - The -I FORWARD signified that this rule should be inserted into the specified chain, rather than appended.

Port forwarding (to a different IP address)

This can be used to forward incoming traffic (e.g. web traffic) to a different host or vm. For example, a host system could be running a web server as a virtual machine and incoming web traffic on TCP port 80 should be forwarded to that vm.

[root@srv1 ~]# iptables -A PREROUTING -i eth1 -p tcp -m tcp --dport 80 -j DNAT --to-destination 192.168.0.16:80

This rule will be appended to the PREROUTING chain within the *nat table.
As TCP traffic on port 80 data arrives across interface eth1, it is transferred to a host with IP address of 192.168.0.16 on port 80.

NB - DNAT is a virtual state, whereby the original destination differs from the reply source, which will be the case where NAT has or will take place.

Common rules for *filter table

Allow ICMP traffic

[root@srv1 ~]# iptables -A INPUT -p icmp -j ACCEPT

Allow any traffic from a given interface

[root@srv1 ~]# iptables -A INPUT -i lo -j ACCEPT
[root@srv1 ~]# iptables -A INPUT -i eth0 -j ACCEPT

Allow related and established traffic between two interfaces

[root@srv1 ~]# iptables -A FORWARD -i eth1 -o eth0 -m state --state RELATED,ESTABLISHED -j ACCEPT

Common rules for *nat table

Enable NAT for a specified interface

[root@srv1 ~]# iptables -A POSTROUTING -o eth1 -j MASQUERADE

Enable NAT for specific ports for a given IP subnet

[root@srv1 ~]# iptables -A POSTROUTING -s 192.168.122.0/24 ! -d 192.168.122.0/24 -p tcp -j MASQUERADE --to-ports 1024-65535

Practical examples

Enable NAT/Masquerading on a new system

Firstly, the Linux kernel needs to be told to entertain IP forwarding;

[root@srv1 ~]# echo 1 > /proc/sys/net/ipv4/ip_forward

Or for a permanent solution, edit /etc/sysctl.conf and change the line that says net.ipv4.ip_forward = 0 to net.ipv4.ip_forward = 1.