Verified and Tested 02/16/2015

Introduction

IPTables is a firewall that is either installed already or can be installed onto any of our Linux Distributions for our Cloud service. IPTables is used to configure packet filter rule chains and enforce the built-in or user-defined rule chains for your server. IPTables has main components to it involving Tables, Targets, and Options.

 

Tables

Tables in IPTables are used to separate packets into their matching category. The table sorting depends on modules that your system has loaded and where the packet matches. Right now, there are five distinct tables for use. They are filter, nat, mangle, raw, and security. For the most part, basic IPTables use will just involve the filter table. If you are doing internal routing or more complicated networking, you will focus on using the other tables as well. This page will be going over only filter type rules.

filter – This is the basic table as was mentioned and is typically the defaulted table. This table consists of the rule chains ACCEPT, FORWARD, OUTPUT.

nat – This table is used for packets that will be creating new connections. Like routing a public IP to a private IP. It consists of the rule chains PREROUTING, POSTROUTING, and OUTPUT.

mangle – This table is used for altering the way packets are handled. It consists of the rule chains INPUT, FORWARD, OUTPUT, PREROUTING, and POSTROUTING.

raw – This table is used for configuring exemptions from connection tracking with the NOTRACK target. It consists of PREROUTING and OUTPUT.

security – This table is for MAC (Mandatory Access Control) rules. It consists of INPUT, FORWARD, and OUTPUT

 

Targets

Targets are defined as the value to the rule chain you are creating. They can contain the values ACCEPT, DROP, QUEUE, or RETURN. These values tell the rule how to proceed with a packet that matches the rule.

ACCEPT – This is pretty straightforward and means to allow the packet through to your server.

DROP – Also pretty straightforward in that it means to deny the packet through to your server.

QUEUE – This target means to pass the packet into userspace so a user may define what to do with it.

RETURN – This target stops the processing of the current chain and tells it to resume processing at the previous chain’s next rule.

 

Options

There is a ton of options available for IPTables use. We will list some of the more common ones you will see.

-A, –append – Appends your rule to a chain.

-L, –list [chain] – This will list all the rules in the chain that you specify. If you don’t supply a chain, it will list all the rules.

-F, –flush [chain] – This will remove any chains and rules in the given chain. If you don’t provide a chain, it will remove all rules and chains currently running.

-h – This provides an output of all the options that you may do.

-p, –protocol protocol – This is the packets protocol. These can be TCP, UDP, UDPlite, ICMP, ESP, AH, SCTP, all, number equivalents to these, or begin with a ! to invert the protocol check.

-s, –source address[/mask][,…] – A source address. Can be the IP, IP range via netmask, network name, or hostname.

-d, –destination address[/mask][,…] – A destination address. Same format as –s.

-m, –match match – An extension module that tests for a property.

–dport[s] [port#][,…] – A port you are looking for. –dports is used to specify more than one which are separated by comma. When using –dports make sure to set -m as multiport. To specify a port range, use a : such as 1000:1100 which would be ports 1000-1100.

-j, –jump target – This is the option that allows you to specify the target for your rule.

-i, –in-interface name – Specifies an interface that the packet should be received on.

-o, –out-interface name – Specifies an interface that the packet will leave on.

-v, –verbose – It makes the list command show interface names, rule options, and any masks. Also shows packet and byte counters.

 

Usage and Examples

Here we will show some basic usage and examples. A typical IPTables rule will tend to follow these formats. Make sure all your rules are done above COMMIT as this is used to end a table. Rules for another table will follow their own end COMMIT.

For allow established traffic:

-A INPUT -m match –state ESTABLISHED,RELATED -j ACCEPT

In use this would be:

-A INPUT -m state –state ESTABLISHED,RELATED -j ACCEPT

For allowing specific port[s].

-A INPUT -p protocol -m match –dport port -j ACCEPT

In use, an example would be:

-A INPUT -p tcp -m tcp –dport 80 -j ACCEPT

Or

-A INPUT -p tcp -m multiport –dports 80,443 -j ACCEPT

This would allow web traffic through on port 80 (in the second example, port 443 too.) That’s not all you can do. You can also add the following like in the below examples.

-A INPUT -s IPADDR -p tcp -m tcp –dport 22 -j ACCEPT

This allows only the IP address (IPADDR) that you specify through on your port 22 (ssh.) You can use any form of IP addresses that the -s option allows.

-A INPUT -p tcp -m tcp –dport 22 -j DROP

Following in step with the above example, this will drop any and all TCP traffic traveling to you SSH port. It is a good rule of thumb to always have any DROP rules at the very bottom of your list of ACCEPT rules in the chain. For example:

-A INPUT … -j ACCEPT

-A INPUT … -j ACCEPT

-A INPUT … -j ACCEPT

-A INPUT … -j DROP

-A FORWARD … -j ACCEPT

-A FORWARD … -j ACCEPT

-A FORWARD … -j ACCEPT

-A FORWARD … -j DROP

-A OUTPUT … -j ACCEPT

-A OUTPUT … -j ACCEPT

-A OUTPUT … -j ACCEPT

-A OUTPUT … -j DROP

You can also define the targets (like ACCEPT or DROP) up at the top of your IPTables for default behavior. When it’s a new install of IPTables, you always see the default behavior (defined at the top where it lists INPUT, FORWARD, and OUTPUT) as ACCEPT. A good rule of thumb for security is to change it to DROP and only make rules for the ports or traffic that you want to allow. Only change it though if you are sure you have made the exceptions you need.

Also at the top of a table, you can define your own targets. Let’s say you have the following at the top of your IPTables:

*filter

:INPUT DROP [0:0]

:FORWARD DROP [0:0]

:OUTPUT DROP [0:0]

COMMIT

Now, as we have blocked everything, let’s start with adding our own targets like:

*filter

:INPUT DROP [0:0]

:FORWARD DROP [0:0]

:OUTPUT DROP [0:0]

:APP – [0:0]

:MyNewTarget – [0:0]

COMMIT

Now you can create a custom type rule chain like the below.

*filter

:INPUT DROP [0:0]

:FORWARD DROP [0:0]

:OUTPUT DROP [0:0]

:APP – [0:0]

:MyNewTarget – [0:0]

 

-A APP -p tcp -m multiport –dports 30:40,50:80 -j ACCEPT

-A INPUT -s IPADDR -j MyNewTarget

 

-A MyNewTarget -j APP

COMMIT

Now what we have done in the above is create new targets, APP and MyNewTarget and assign to them options and targets. What the last rule (above COMMIT) does, is if the connection comes from your source IP and is attempting to hit ports 30-40 or 50-80, allow it through! It’s all defined in the chain above. This level of customization isn’t truly needed unless you are using multiple hosts or using ports you find yourself repeating and want more of a shorthand or “shortcut” to writing out rules. For example, any new rule you need coming from that source IP, you can now just use MyNewTarget.

All in all, IPTables provides a lot of customization of rules and allows you to handle how your server handles incoming, outgoing, and internal traffic. Some rules can be confusing and hard to follow, but there is no doubt to the helpfulness they provide to server stability.