Building the Veloren Server for Raspberry Pi

The model of Raspberry Pi you have and the operating system your Pi is running can significantly impact your installation steps and Veloren's performance on your Pi. For the best performance we recommend using the Raspberry Pi model 4 and a 64-bit OS like Ubuntu Server, which is available through the Raspberry Pi Imaging utility. There is a 64-bit beta of Raspberry Pi OS as well.

Note: The amount of RAM on the Pi 4 does not matter. Even the official Veloren server uses well under 2 GiB of memory most of the time.

Cross-compiling or not

Cross-compilation means setting up a toolchain for the Raspberry Pi's instruction set on a separate computer in order to compile the Veloren server binary which you then copy onto the Pi.

Direct compiling is preferred if:

  • You don't want to install and set up as many things on your computer
  • You want to be up and running in fewer steps

Cross compiling is preferred if:

  • You have a computer setup for development
  • You want the compilation step to be faster

Note: Even with cross-compilation it is necessary to clone the assets folder from the code repository onto the Pi as these files are not compiled into the server binary.

Direct Compiling

Log in to your Pi using SSH.
ssh <username>@<ip> and enter password Install the Rust programming language on the Pi if it's not already installed. 

curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh

Install Git LFS in order to download the audio and visual assets along with the source code. 

sudo apt install git-lfs

Git clone the Veloren codebase. You may need to generate SSH keys on your Pi for your GitLab account. 

git clone https://gitlab.com/veloren/veloren.git

Compile the server binary from the code in the Veloren directory. 

cd veloren
cargo build --bin veloren-server-cli --release

Run the server binary, optionally with -h or --help to see the list of arguments you can supply. 

./target/release/veloren-server-cli

Note: Compilation on a Raspberry Pi 4 running Ubuntu Server 64-bit can take around 30 minutes for an optimized build.

Note: Remember to replace <username> by your own username and <ip> by the IP address of your Raspberry Pi!

Note: The process itself is resource heavy. Maybe still consider cross compiling. RAM usage can go over 8G, so you may consider creating a swap (although swap slows down the entire compilation process)

Note: If you experience an error when compiling wasmer-vm, you will need to disable plugin support by removing plugins = ["server/plugins"] from server-cli/Cargo.toml.

Cross Compiling

Installing Dependencies

Install the following dependencies on the cross compiling machine and the Raspberry Pi:

Cross Compiling PC:

  • git
  • git-lfs
  • cargo (which is installed along with Rust)
  • docker

Raspberry Pi:

  • git
  • git-lfs

Preparing the Raspberry Pi

First we need to clone the Veloren git repository.

git clone https://gitlab.com/veloren/veloren

Next we can reset the repository to the state of the last stable release. To do that we have to find out the git hash of that version. This can be found on https://gitlab.com/veloren/veloren/-/releases. The git hash is the sequence in the bottom left corner of the release. This step can be skipped if you want to play with the very latest updates.

cd veloren
git checkout <git hash>

Next we have to create some folders to later put the binary in.

mkdir target
mkdir target/release

Next, we need to add an environment variable to our .profile.

nano ~/.profile

In the last line of this file, add the line export VELOREN_USERDATA="$(pwd)/userdata

Lastly, we reload our environment to use the new variable

. ~/.profile

Cross Compiling the code

Again we have to clone the Veloren repository to our cross compiling machine, reset it to the state of the latest stable release and setup git-lfs.

git clone https://gitlab.com/veloren/veloren
cd veloren
git checkout <git hash>
git lfs install
cargo install cross

The install command will be different depending on which OS you have running on the Raspberry Pi.
If you are using a 64-bit OS like Ubuntu Server with an ARMv8 instruction set, run the command:

cross build --target aarch64-unknown-linux-gnu --bin veloren-server-cli --release

If you are using Raspberry Pi OS which is currently 32-bit and using ARMv7 instruction set for backwards compatibility reasons, run the command:

cross build --target armv7-unknown-linux-gnueabihf --bin veloren-server-cli --release

Note: You may need to register an account on docker hub and run the docker login command in the terminal to access the Docker image required for cross-compilation.

When the compilation process has finished, we can move the binary to the Raspberry Pi. If you have SSH enabled on your Raspberry Pi, you can use the scp command.

scp target/<instruction_set>/release/veloren-server-cli <username>@<ip>:~/veloren/target/release/veloren-server-cli

Note: <instruction_set> refers to either aarch64-unknown-linux-gnu or armv7-unknown-linux-gnueabihf, whichever one you used to compile. Remember to replace <username> by your own username and <ip> by the IP address of your Raspberry Pi!

Creating a systemd service

On the Raspberry Pi we can now create a systemd service for the server. To do that, create a service file with the content below.

sudo nano /etc/systemd/system/veloren-server.service

Note: If you cross-compiled the binary, add an environment variable to the Service section: Environment="VELOREN_USERDATA=/home/veloren/target/release/userdata"

[Unit]
Description=Veloren Server
After=network.target
StartLimitIntervalSec=0
[Service]
Type=simple
User=<username>
WorkingDirectory=/home/<username>/veloren
ExecStart=/home/<username>/veloren/target/release/./veloren-server-cli

[Install]
WantedBy=multi-user.target

Now our service is ready and can be started.

sudo systemctl start veloren-server.service

To watch how the server starts and to see debug info, you can use the following command.

sudo journalctl -f -u veloren-server.service

In case you want to start the server every time the Raspberry Pi boots, you can enable the service.

sudo systemctl enable veloren-server.service

Monitoring the server

When the server is running, you can watch your Pi's performance with tools like htop or set up a Prometheus server to scrape metrics from your Veloren server at the addresshttp://<ip>:14005/metrics.