So, with credit to the Factorio wiki and cbednarski's helpful gist, I have been able to successfully setup a Factorio headless server on AWS EC2 using the Amazon Linux 2 AMI (HVM), SSD Volume Type AMI.
This guide was done using the Amazon Linux 2 AMI image for EC2, and Factorio 1.19.0. However it is possible for this guide to also work on other linux AMI images.
- AWS Account
- Licensed copy of factorio
- That's it!
Sign into your AWS account and go to your dashboard. Search for 'EC2' and select that service to bring up the EC2 dashboard / console page. Click on the orange 'Launch Instance' button near the bottom of the page to bring up the EC2 configuration page.
Select the Amazon Linux 2 AMI (HVM), SSD Volume Type AMI for x86 (AMI Id: ami-047a51fa27710816e (64-bit x86))
Make sure you select x86 and NOT arm.
This will vary depending on your server needs. For starting a new game I recommend t2.large which comes with 8GB of RAM and 2 vCPU cores - you can always swap out the instance type later on if performance is becoming an issue.
Leave the default settings here, unless there is something specific you want to change.
The default setup comes with 8GB of storage, which is plenty for storing save files, however if you plan to run mods, you will need to figure out how much storage all mods would require on the server.
Click on 'Review and Launch'
You may receive a warning about Your security group, your-security-group-name, is open to the world.
Click on Edit security groups, and you will see an inbound rule for SSH already configured. Under the 'source' column, change the 'Custom' to 'My IP' so that only your computer is allowed to SSH into the server. Note that you may need to update this rule multiple times if your IP address is not static. If you know exactly what your IP address range is going to be, you can configure it using the Custom rule instead. For simplicity reasons I chose 'My IP'.
Keep advancing to the next page until you reach the Select an existing key pair or create a new key pair window.
In the key pair window, select Create a new key-pair, and give the key-pair a name, for example: factorio-ec2-instance. Then hit Download Key Pair to download the factorio-ec2-instance.pem file. This is your private key that allows you to SSH into the EC2 instance. Without it, you will not be able to SSH in, and you will have to regenerate a new key pair for yourself.
Save this .pem file somewhere so you don't forget it. You will need to reference it each time you wish to SSH in to your EC2 instance. For the purposes of this guide, my .pem file is stored on ~/documents on OSX, or C:/.../My Documents in Windows.
Once you've downloaded the .pem file and configured everything, go ahead and launch your instance.
Once you have launched your instance, go to the Instances page and you will see your instance appear in the list. Click on your 'instance-id' and then click on the Connect button.
On the Connect to instance page, click on the SSH Client tab and follow the 4 steps laid out on that page to ssh into your EC2 instance. You will need to make sure that you provide the path to the .pem file, or ensure you launch your terminal in the same directory as the .pem file. You will get a command similar to the one below:
ssh -i "factorio-ec2-instance.pem" ec2-user@ec2-xxx-xx-xxx-xxx.compute-1.amazonaws.com
For the purposes of this guide, we will be using port 62754 for hosting the game.
Go to your EC2 instance summary page -> Security tab, and click on the blue link under Security Groups to bring up the rules page.
The only thing we need to do here is add Inbound Rules exceptions for ports 62754 and 34197 for both TCP and UDP.
Click Edit Inbound Rules
Click Add Rule choose Custom TCP for the Type, and for Port Range type in 62754, for Source, choose Anywhere. Repeat these process until you have the following rules on your EC2 instance.
Type Port-Range Source
SSH 22 Custom
Custom TCP 62754 Anywhere
Custom UDP 62754 Anywhere
Custom TCP 34197 Anywhere
Custom UDP 34197 Anywhere
Note that if you wish to be able to ping your EC2 instance, you will also need to enable inbound rules for Custom ICMP - IPv6 and Custom ICMP - IPv4 otherwise the AWS firewall will block any ping requests.
On your EC2 instance, run the following command to download the latest headless server
wget -O factorio_headless.tar.gz https://factorio.com/get-download/stable/headless/linux64
Also, as they redirect download requests, wget will not name the file correctly. The easiest way to rectify it is to manually set the file name using the -O flag and manually providing the file name as used in the example above.
Wait for the file to download before proceeding.
Factorio runs out of the /opt directory, a directory resevered in UNIX for non-default software installation.
Run the following command to extract and install on your EC2 instance:
sudo tar xvf factorio_headless.tar.gz -C /opt
Just for the sake of keeping things tidy, I'd also delete the downloaded archive rm ~/factorio_headless.tar.gz, as we are now finished with it.
If you wish to add any mods or saves, you can do so by first creating the mods and saves directories within /opt/factorio. sudo mkdir /opt/factorio/saves and sudo mkdir /opt/factorio/mods should achieve this, then you may copy the save & mod archives into the relevant folders.
As you don't want Factorio running as root, let's setup a user that it can run under. The user won't need a lot of functionality, as it's only going to be running Factorio and won't need to be logged into, so let's provide some flags to create a really basic user:
sudo adduser --disabled-login --no-create-home --gecos factorio factorio
The above command will add a user, not setting a password --disabled-login, without creating a home directory in /home --no-create-home, without asking for user information --gecos, create user factorio and add them/create the group factorio.
Now that the new user is created, we need to make it the owner of the Factorio directory so that it can access and perform operations within it, sudo chown -R factorio:factorio /opt/factorio. The -R flag being recursive.
Factorio headless requires a JSON formatted config file. They provide an example one to get you started, in which can be copied and then edited to suit your requirements.
First, lets change directory cd /opt/factorio/data where the server-settings.example.json config file is located, it's not necessary, but just the way I do things. Now to make a copy of the config file under its working name sudo cp server-settings.example.json server-settings.json and then edit it with a text editor of your choice, I like nano as its easy sudo nano server-settings.json.
The configuration is fairly straight forward and mirrors that of the GUI when setting up a server via the standard edition of Factorio. Here is the configuration reference. Configure it as you wish, save, exit and proceed onto the next step.
Here is an example of the server settings config file:
{
"name": "my factorio server",
"description": "a sample server",
"tags": ["game", "tags"],
"_comment_max_players": "Maximum number of players allowed, admins can join even a full server. 0 means unlimited.",
"max_players": 0,
"_comment_visibility": ["public: Game will be published on the official Factorio matching server",
"lan: Game will be broadcast on LAN"],
"visibility":
{
"public": true,
"lan": true
},
"_comment_credentials": "Your factorio.com login credentials. Required for games with visibility public",
"username": "your-username-here",
"password": "",
"_comment_token": "Authentication token. May be used instead of 'password' above.",
"token": "",
"game_password": "",
"_comment_require_user_verification": "When set to true, the server will only allow clients that have a valid Factorio.com account",
"require_user_verification": true,
"_comment_max_upload_in_kilobytes_per_second" : "optional, default value is 0. 0 means unlimited.",
"max_upload_in_kilobytes_per_second": 0,
"_comment_max_upload_slots" : "optional, default value is 5. 0 means unlimited.",
"max_upload_slots": 5,
"_comment_minimum_latency_in_ticks": "optional one tick is 16ms in default speed, default value is 0. 0 means no minimum.",
"minimum_latency_in_ticks": 0,
"_comment_ignore_player_limit_for_returning_players": "Players that played on this map already can join even when the max player limit was reached.",
"ignore_player_limit_for_returning_players": false,
"_comment_allow_commands": "possible values are, true, false and admins-only",
"allow_commands": "admins-only",
"_comment_autosave_interval": "Autosave interval in minutes",
"autosave_interval": 10,
"_comment_autosave_slots": "server autosave slots, it is cycled through when the server autosaves.",
"autosave_slots": 10,
"_comment_afk_autokick_interval": "How many minutes until someone is kicked when doing nothing, 0 for never.",
"afk_autokick_interval": 0,
"_comment_auto_pause": "Whether should the server be paused when no players are present.",
"auto_pause": true,
"only_admins_can_pause_the_game": true,
"_comment_autosave_only_on_server": "Whether autosaves should be saved only on server or also on all connected clients. Default is true.",
"autosave_only_on_server": true,
"_comment_non_blocking_saving": "Highly experimental feature, enable only at your own risk of losing your saves. On UNIX systems, server will fork itself to create an autosave. Autosaving on connected Will",
"non_blocking_saving": false,
"_comment_segment_sizes": "Long network messages are split into segments that are sent over multiple ticks. Their size depends on the number of peers currently connected. Increasing the segment size will...",
"minimum_segment_size": 25,
"minimum_segment_size_peer_count": 20,
"maximum_segment_size": 100,
"maximum_segment_size_peer_count": 10
}
Copy your save into the same location as your .pem file, or specify the full path when you do the command. Be sure to reference the same ec2-user@instance-id as was provided by AWS when you were going through the Connect window.
scp -i "factorio-ec2-instance.pem" your_factorio_save.zip ec2-user@ec2-xxx-xx-xxx-xxx.compute-1.amazonaws.com:.
Afterwards you can sudo cp your_factorio_save.zip /opt/factorio/saves to copy your save into the factorio installation directory.
So, we want Factorio to run as a service. So lets create a new service for Factorio within the systemd service scheduler. Again, I like nano as its easy, but any text editor will do sudo nano /etc/systemd/system/factorio.service. The service should contain the following:
[Unit]
Description=Factorio Headless Server
[Service]
Type=simple
User=factorio
ExecStart=/opt/factorio/bin/x64/factorio --server-settings /opt/factorio/data/server-settings.json --start-server-load-latest –Port 62754
Change the {save_file} for the file name of your save. You may also want to swap out --start-server /opt/factorio/saves/{save_file}.zip for --start-server-load-latest, which will automatically pull through the latest version of your world across manual and auto saves. If you wish to look at more parameters, there is more information on the Factorio multiplayer wiki.
Save the file and exit. Time to start the service and see if our setup has worked!
Firstly, reload the available service daemons systemctl daemon-reload, now start the newly created service systemctl start factorio and see if it's running systemctl status factorio.service. That is the output from the Factorio server, any errors or configuration issues will present themselves here if you ever need to debug.
If you do make any configuration changes, you'll need to restart the service sudo systemctl restart factorio.service.
That's it! You should now have a working Factorio server. Head over to your client and under "Multiplayer" > "Connect to a server", provide your address that you set it up against (and port if you changed it from the default) and play spend hours labouring over tiny details to maximize efficiency!
- How to upgrade
With a quick shout out to dillinger.io for making MD editing a more pleasant experience...
Out of curiosity, how much to you estimate running this on AWS costs?