Iptables is a standard firewall included in most Linux distributions by default (a modern variant called nftables will begin to replace it). It is actually a front end to the kernel-level netfilter hooks that can manipulate the Linux network stack. It works by matching each packet that crosses the networking interface against a set of rules to decide what to do.
In this article, we'll move on to a practical example to demonstrate how to create a basic rule set for an Ubuntu 14.04 server.
Basic iptables CommandsNow that you have a good understanding of iptables concepts, we should cover the basic commands that will be used to form complex rule sets and to manage the iptables interface in general.
First, you should be aware that iptables commands must be run with root privileges. This means you need to log in as root, use
sudo -ito gain a root shell, or precede all commands with
sudo. We are going to use
sudoin this guide since that is the preferred method on an Ubuntu system.
A good starting point is to list the current rules that are configured for iptables. You can do that with the
sudo iptables -L
Chain INPUT (policy ACCEPT) target prot opt source destination Chain FORWARD (policy ACCEPT) target prot opt source destination Chain OUTPUT (policy ACCEPT) target prot opt source destination
As you can see, we have our three default chains (INPUT,OUTPUT, and FORWARD). We also can see each chain's default policy (each chain has ACCEPT as its default policy). We also see some column headers, but we don't see any actual rules. This is because Ubuntu doesn't ship with a default rule set.
We can see the output in a format that reflects the commands necessary to enable each rule and policy by instead using the
sudo iptables -S
-P INPUT ACCEPT -P FORWARD ACCEPT -P OUTPUT ACCEPT
To replicate the configuration, we'd just need to type
sudo iptablesfollowed by each of the lines in the output. (Depending on the configuration, it may actually slightly more complicated if we are connected remotely so that we don't institute a default drop policy before the rules are in place to catch and allow our current connection.)
If you do have rules in place and wish to scrap them and start over, you can flush the current rules by typing:
sudo iptables -F
Once again, the default policy is important here, because, while all of the rules are deleted from your chains, the default policy will not change with this command. That means that if you are connected remotely, you should ensure that the default policy on your INPUT and OUTPUT chains are set to ACCEPT prior to flushing your rules. You can do this by typing:
sudo iptables -P INPUT ACCEPT sudo iptables -P OUTPUT ACCEPT sudo iptables -F
You can then change the default drop policy back to DROP after you've established rules that explicitly allow your connection. We'll go over how to do that in a moment.
Make your First RuleWe're going to start to build our firewall policies. As we said above, we're going to be working with the INPUT chain since that is the funnel that incoming traffic will be sent through. We are going to start with the rule that we've talked about a bit above: the rule that explicitly accepts your current SSH connection.
The full rule we need is this:
sudo iptables -A INPUT -m conntrack --ctstate ESTABLISHED,RELATED -j ACCEPT
This may look incredibly complicated, but most of it will make sense when we go over the components:
- -A INPUT: The
-Aflag appends a rule to the end of a chain. This is the portion of the command that tells iptables that we wish to add a new rule, that we want that rule added to the end of the chain, and that the chain we want to operate on is the INPUT chain.
- -m conntrack: iptables has a set of core functionality, but also has a set of extensions or modules that provide extra capabilities.
In this portion of the command, we're stating that we wish to have access to the functionality provided by the
conntrackmodule. This module gives access to commands that can be used to make decisions based on the packet's relationship to previous connections.
- --ctstate: This is one of the commands made available by calling the
conntrackmodule. This command allows us to match packets based on how they are related to packets we've seen before.
We pass it the value of
ESTABLISHEDto allow packets that are part of an existing connection. We pass it the value of
RELATEDto allow packets that are associated with an established connection. This is the portion of the rule that matches our current SSH session.
- -j ACCEPT: This specifies the target of matching packets. Here, we tell iptables that packets that match the preceding criteria should be accepted and allowed through.
We can see the changes if we list the rules:
sudo iptables -L
Chain INPUT (policy ACCEPT) target prot opt source destination ACCEPT all -- anywhere anywhere ctstate RELATED,ESTABLISHED Chain FORWARD (policy ACCEPT) target prot opt source destination Chain OUTPUT (policy ACCEPT) target prot opt source destination
Now that you know the general syntax, let's continue by adding some more cases where we want to accept the connection.
Accept Other Necessary ConnectionsWe have told iptables to keep open any connections that are already open and to allow new connections related to those connections. However, we need to create some rules to establish when we want to accept new connections that don't meet those criteria.
We want to keep two ports open specifically. We want to keep our SSH port open (we're going to assume in this guide that this is the default 22. If you've changed this in your SSH configuration, modify your value here). We are also going to assume that this computer is running a web server on the default port 80. If this is not the case for you, you don't have to add that rule.
The two lines we're going to use to add these rules are:
sudo iptables -A INPUT -p tcp --dport 22 -j ACCEPT sudo iptables -A INPUT -p tcp --dport 80 -j ACCEPT
As you can see, these are very similar to our first rule, but perhaps more simple. The new options are:
- -p tcp: This option matches packets if the protocol being used is TCP. This is a connection-based protocol that will be used by most applications because it allows for reliable communication.
- --dport: This option is available if the
-p tcpflag is given. It gives a further requirement of matching the destination port for the matching packet. The first rule matches for TCP packets destined for port 22, while the second rule matches TCP traffic pointed towards port 80.
loopback device, which directs traffic back to itself rather than to other computers.
So if one service wants to communicate with another service that is listening for connections on port 4555, it can send a packet to port 4555 of the loopback device. We want this type of behavior to be allowed, because it is essential for the correct operation of many programs.
The rule we need to add is this:
sudo iptables -I INPUT 1 -i lo -j ACCEPT
This looks a bit different than our other commands. Let's go over what it is doing:
- -I INPUT 1: The
-Iflag tells iptables to insert a rule. This is different than the
-Aflag which appends a rule to the end. The
-Iflag takes a chain and the rule position where you want to insert the new rule.
In this case, we're adding this rule as the very first rule of the INPUT chain. This will bump the rest of the rules down. We want this at the top because it is fundamental and should not be affected by subsequent rules.
- -i lo: This component of the rule matches if the interface that the packet is using is the "lo" interface. The "lo" interface is another name for the loopback device. This means that any packet using that interface to communicate (packets generated on our server, for our server) should be accepted.
-Sflag. This is because the
-Lflag doesn't include some information, like the interface that a rule is tied to, which is an important part of the rule we just added:
sudo iptables -S
-P INPUT ACCEPT -P FORWARD ACCEPT -P OUTPUT ACCEPT -A INPUT -i lo -j ACCEPT -A INPUT -m conntrack --ctstate RELATED,ESTABLISHED -j ACCEPT -A INPUT -p tcp -m tcp --dport 22 -j ACCEPT -A INPUT -p tcp -m tcp --dport 80 -j ACCEPT
Implementing a Drop RuleWe now have four separate rules that explicitly accept packets based on certain criteria. However, our firewall currently is not blocking anything.
If a packet enters the INPUT chain and doesn't match one of the four rules that we made, it is being passed to our default policy, which is to accept the packet anyways. We need to change this.
There are two different ways that we can do this, with some pretty important differences.
The first way we could do this is to modify the default policy of our INPUT chain. We can do this by typing:
sudo iptables -P INPUT DROP
This will catch any packets that fall through our INPUT chain, and drop them. This is what we call a default drop policy. One of the implications of this type of a design is that it falls back on dropping packets if the rules are flushed.
The alternative approach is to keep the default policy for the chain as accept and add a rule that drops every remaining packet to the bottom of the chain itself.
If you changed the default policy for the INPUT chain above, you can set it back to follow along by typing:
sudo iptables -P INPUT ACCEPT
Now, you can add a rule to the bottom of the chain that will drop any remaining packets:
sudo iptables -A INPUT -j DROP
The result under normal operating conditions is exactly the same as a default drop policy. This rule works by matching every remaining packet that reaches it. This prevents a packet from ever dropping all of the way through the chain to reach the default policy.
Basically, this is used to keep the default policy to accept traffic. That way, if there are any problems and the rules are flushed, you will still be able to access the machine over the network. This is a way of implementing a default action without altering the policy that will be applied to an empty chain.
Of course, this also means that any rule that any additional rule that you wish to add to the end of the chain will have to be added before the drop rule. You can do this either by temporarily removing the drop rule:
sudo iptables -D INPUT -j DROP sudo iptables -A INPUT new_rule_here sudo iptables -A INPUT -j DROP
Or, you can insert rules that you need at the end of the chain (but prior to the drop) by specifying the line number. To insert a rule at line number 4, you could type:
sudo iptables -I INPUT 4 new_rule_here
If you are having trouble knowing which line number each rule is, you can tell iptables to number the rules by typing:
sudo iptables -L --line-numbers
Chain INPUT (policy DROP) num target prot opt source destination 1 ACCEPT all -- anywhere anywhere 2 ACCEPT all -- anywhere anywhere ctstate RELATED,ESTABLISHED 3 ACCEPT tcp -- anywhere anywhere tcp dpt:ssh 4 ACCEPT tcp -- anywhere anywhere tcp dpt:http Chain FORWARD (policy ACCEPT) num target prot opt source destination Chain OUTPUT (policy ACCEPT) num target prot opt source destination
This can be helpful to make sure you are adding your rule at the appropriate position.
Saving your iptables ConfigurationBy default, the rules that you add to iptables are ephemeral. This means that when you restart your server, your iptables rules will be gone.
This is actually a feature for some user because it gives them an avenue to get back in if they have accidentally locked themselves out of the server. However, most users will want a way to automatically save the rules you have created and to load them when the server starts.
There are a few ways to do this, but the easiest way is with the
iptables-persistentpackage. You can download this from Ubuntu's default repositories:
sudo apt-get update sudo apt-get install iptables-persistent
During the installation, you will be asked if you would like to save your current rules to be automatically loaded. If you are happy with your current configuration (and you have tested your ability to create independent SSH connections, you can select to save your current rules.
It will also ask you if you want to save the IPv6 rules that you have configured. These are configured through a separate utility called
ip6tableswhich controls the flow of IPv6 packets in almost the same way.
Once the installation is complete, you will have a new service called
iptables-persistentthat is configured to run at boot. This service will load in your rules and apply them when the server is started.
ConclusionYou should now have a good starting point to developing a firewall that addresses your needs. There are many other firewall utilities and some that may be easier, but
iptablesis a good learning tool, if only because it exposes some of the underlying netfilter structure and because it is present in so many systems.