Guides

Radicle Seeder's Guide

How to run a Radicle seed node


To seed is to give back. By seeding repositories on the Radicle network, you offer bandwidth, storage, and data availability to Radicle users.

In this guide, we’ll go through the various steps required to set up a Radicle seed node on a Linux system. Seeding only requires an internet connection, a public, static IP address, and modest hardware.

We’ll cover public seed nodes with an open seeding policy, as well as community seed nodes with more selective policies.

Introduction

The term seed node originally comes from BitTorrent. In the BitTorrent protocol, nodes that possess the data for a given torrent file and begin uploading it to peers are called seeders.

In Radicle, while all users contribute to the network by seeding data, the term seed node specifically refers to a server that is accessible through the internet and hosts content for peers. Seed nodes need to remain online and accessible to provide their service to the network.

The need for seed nodes

While a peer-to-peer network without seed nodes is feasible, it is impractical. This is because regular “user” nodes go online and offline all the time, so finding a user from which to download a certain piece of content can be challenging, or even impossible if all users with that content are offline.

Therefore, a healthy peer-to-peer network necessitates at least some highly available nodes that participate in the network like regular peers, but seldom go offline. These are called seed nodes.

Getting Started

For this guide, we’ll focus on setting up a seed node using systemd. If you’re running a different service manager, you should be able to follow along just fine, as we’ll be explaining the steps. Using a service manager is not required, but highly recommended.

What you will need

To run a Radicle seed node, you will need a server with:

Server

If you already have a VPS (Virtual Private Server) or home server running a Linux distribution, you should be all set. If not, any VPS will do. Between 1-2GB of RAM with a shared CPU and 10GB of disk space should be enough to get started. Once you’ve logged into your server, proceed with user setup and installation. For this setup process, you’ll need sudo capabilities.

Tip: If you have trouble choosing a VPS, we recommend getting a small instance on Hetzner or Digital Ocean. These are priced fairly and have good documentation and support.

User setup

If you wish to create a new user under which to run the seed node services, this is the time. In this guide, we’ve chosen to run our services under a user called seed. Radicle data and configuration will be stored under this user’s home folder. This ensures that if your Radicle services were to be compromised, the attacker would have very limited access to your system.

Here’s an example user setup with seed as the user and group name:

sudo groupadd --system seed
sudo useradd --system --gid seed --create-home seed

If you’re administering your server via a non-root account, you’ll want to add it to the seed group as well:

sudo usermod -a -G seed $(whoami)

Installation

For a seed node, you will need to install at minimum the Radicle CLI (rad), and network daemon (radicle-node).

To install these, head over to the download page and follow the instructions there. You will have to download, verify and extract the binaries and manuals to your preferred location.

For this guide, we recommend installing Radicle under /usr/local. This will require you to have write permissions to /usr/local/bin and /usr/local/man. You can give yourself these permission by changing ownership of these directories to the current user and group:

$ sudo chown $(whoami): /usr/local/{bin,man,man/man1}

You should then be able to extract the archive you downloaded and verified with:

$ tar -xvJf <archive> --strip-components=1 -C /usr/local/

Finally, login as the seed user and proceed with installation:

sudo su seed

To check your installation, run the rad command, and confirm where your Radicle home is with the rad path command.

To offer web browsing and HTTP access to your seed node, you will additionally need to run radicle-httpd, the HTTP daemon. We’ll cover this after we’ve set up our basic node.

Creating a profile

Once the Radicle binaries installed, we can create a Radicle profile. This consists of an Ed25519 key pair and directory under which Radicle stores user data.

rad auth --alias seed.radicle.example

The above command will create a profile with the given node alias in ~/.radicle, or $RAD_HOME if set. We recommend setting your alias to the domain name that points to your node’s host.

For seed nodes, it’s not generally necessary to set up a passphrase for your key, so you can simply skip that prompt by pressing enter.

Tip: If you wish to set a passphrase anyway, you will have to set the RAD_PASSPHRASE environment variable in your node’s systemd unit file.

Once your profile is initialized, a decentralized identifier (DID) will be output. The part after the did:key: prefix is your Node ID. This is how your seed node will be identified on the network. The Node ID is the public part of the Ed25519 key pair generated above, via rad auth.

You can view information about your Radicle profile by running rad self. The --nid and --alias flags can be used to return the Node ID and alias.

Configuring your node

There are a couple of things we need to set up a seed node.

First, we must set the node’s default seeding policy.

The seeding policy tells the node which repositories and forks to fetch and offer to the network. For public seed nodes, a permissive seeding policy is often set, such that all data on the network is stored and replicated.

Second, we must set an external address for the node to be reached on the network. This address will be advertised to peers, allowing them to connect to your seed node. Generally, this will be a DNS name with port 8776, for example seed.radicle.example:8776.

Here’s an example minimal configuration file with a permissive seeding policy and external address set:

{
  "node": {
    "alias": "seed.radicle.example",
    "externalAddresses": ["seed.radicle.example:8776"],
    "seedingPolicy": {
      "default": "allow",
      "scope": "all"
    }
  }
}

Your node is configured with a file named config.json in your Radicle home directory. You can get the full path of the config file with the rad self --config command. Additionally, you can output the current configuration with rad config. Attributes that aren’t set in config.json will take on default values. You can open your configuration file for editing with rad config edit.

Let’s start by configuring your seeding policy.

Seeding policies

The most important setting when it comes to seeding is your default seeding policy. This setting will determine what content your seed node fetches and replicates on the network when it encounters a repository it doesn’t know or hasn’t been given special instructions for.

The setting is configured under the node.seedingPolicy field in your configuration (~/.radicle/config.json). You can open your configuration file directly using the rad config edit command, or use your preferred editor. You can also enter the following command to display your current default policy:

rad config get node.seedingPolicy

When a seeding policy is not set for a specific repository or node, the default policy is applied, hence the importance of this setting.

Broadly, there are two options for the default policy.

A permissive seeding policy

A permissive or “open” policy is said to be fully-replicating, meaning your seed will try to have a full copy of all repository data available on the network.

An example of a node with this policy is seed.radicle.garden, a node operated by the Radicle team, for the Radicle community.

This is a good default for seeders who want to support the network without having to think about it too much.

Set node.seedingPolicy.default to allow and node.seedingPolicy.scope to all to configure your node this way.

{
  "node": {
    ...
    "seedingPolicy": {
      "default": "allow",
      "scope": "all"
    }
  }
}

A selective seeding policy

A selective or restricted policy requires you, the operator, to manually allow repositories to be seeded. This means that the node will ignore all repositories, except the ones that are pre-configured to allow seeding.

An example is the seed.radicle.xyz node, which only seeds core team repositories.

This is a good policy for communities, teams, companies and individuals who want to limit the data hosted by their seed, or node operators who want to require some form of authentication or payment for seeding.

Set node.seedingPolicy.default to block to enable this, and call rad seed to configure allow policies for specific repositories you want to seed. Your seed node won’t seed anything until you explicitly allow it to.

{
  "node": {
    ...
    "seedingPolicy": {
      "default": "block"
    }
  }
}

Setting a repository’s seeding policy

To override the default policy for a specific repository, the seed and block commands are used. For example, if the default policy is block, we can decide to allow a repository to be seeded by calling rad seed. For example,

rad seed rad:z3gqcJUoA1n9HaHKufZs5FCSGazv5

This will override the default block rule for this one specific repository. To remove this override, unseed can be used:

rad unseed rad:z3gqcJUoA1n9HaHKufZs5FCSGazv5

Without a seeding policy for this repository, the default policy will apply.

In the case of an open seeding policy, where the default is allow, we can explicitly block certain repositories from being seeded. For example,

rad block rad:z9DV738hJpCa6aQXqvQC4SjaZvsi

This will override the default policy and ensure that this repository is never replicated on your node.

Viewing policies

You can view your node’s default by entering rad config get node.seedingPolicy. To view the policy of a specific repository, use the rad inspect command. For example:

rad inspect rad:z3gqcJUoA1n9HaHKufZs5FCSGazv5 --policy

This will return the default policy if you haven’t set a specific policy for that repository.

Finally, to view all seeding policies that have been set on repositories, simply enter rad seed with no options. This will list all configured policies.

Your external address

Now that you’ve configured your seeding policy, it’s time to set your node’s external address. This is a public address, typically a DNS name that points to your node. Though a single address is sufficient, you are free to set up to 16 external addresses.

You’ll find this setting in your configuration file, under node.externalAddresses. External addresses are JSON strings of the form <host>:<port>, for example seed.radicle.example:8776, where <host> is a DNS name, and port is usually 8776.

{
  "node": {
    ...
    "externalAddresses": ["seed.radicle.example:8776"]
  }
}

Once at least one external address is set, you’re ready to start your node for the first time.

Your node address

For others to be able to connect to your node directly, they need your Node Address. This is a combination of your Node ID and your node’s external address.

If you’ve configured one or more external addresses, simply entering the following command from the seed user will output your node addresses:

rad node config --addresses

Share this with others, and they will be able to connect to your node using rad node connect <address>, or by adding your address to their configuration, under the node.connect field.

Running your node

Before setting up your node as a system service, it’s a good idea to run it once to make sure everything is working properly. Enter the following command to start your node in the foreground:

rad node start --foreground

You should see log output as your node starts to sync with the network. If there are any errors or issues connecting to the network, you should see errors in the output.

If the node started without problem, stop it with Ctrl-C, and exit the seed session by entering:

exit

This should throw you back into the original session you opened via SSH.

Configuring your node’s system service

Though it’s possible to simply run the Radicle node as a background process, it’s recommended to set up a service to ensure the node is started on boot.

In this guide, we will only cover setup using systemd, but the process is fairly similar for other service managers.

The first thing to do is to get a copy of the radicle-node.service unit file. Place it in /etc/systemd/system/radicle-node.service for it to be found by systemctl:

curl -sS https://seed.radicle.xyz/raw/rad:z3gqcJUoA1n9HaHKufZs5FCSGazv5/570a7eb141b6ba001713c46345d79b6fead1ca15/systemd/radicle-node.service -o /etc/systemd/system/radicle-node.service

Make sure it fits your needs by editing the file directly, or creating an override using systemctl edit.

The systemd unit should be configured to run your node process as the seed user, for security reasons. This is already the case in the above service file.

When you’re ready, you can enable and run the service:

systemctl enable --now radicle-node

Checking your node status

To check the service status of your node, run:

systemctl status radicle-node

Besides using systemctl status, you can also check your node’s status using rad node status. This will give you information on the peers connected to your node. You’ll have to run this as the seed user, like so:

sudo -u seed -- rad node status

To tail your node’s logs, use:

journalctl --unit radicle-node --follow

Changing your node’s configuration

If you change your node’s configuration while the node is running, you’ll have to restart your node for the changes to take effect.

First, verify that the new configuration is valid by running rad config, then restart your node with:

systemctl restart radicle-node

Securing the seed user

Now that your node is configured and working, you can secure the seed user by disabling shell access. This is optional, though recommended.

chsh -s /usr/sbin/nologin seed

This will prevent anyone from logging in as the seed user. Note that from this point onwards, if you chose to disable the seed user’s shell, you’ll have to run all commands via sudo, like we’ve been doing so far.

To facilitate running commands as the seed user, add this alias to your admin shell’s init scripts:

alias rad='sudo -u seed -- rad'

You can then run commands as the seed user by simply using rad as usual.

Firewalls

If you are running a firewall, ensure that port 8776 is open for TCP connections. This will allow inbound connections to your node.

It’s recommended to run a basic firewall to further lock down your server, using something like iptables, though this is out of scope for this guide.

Running the HTTP Daemon

In the sections above, we set up radicle-node, a background process that actively and continuously discovers and replicates repositories on the network, based on your seeding policy. This node allows users to collaborate, host, share and publish repositories on the network via the Radicle CLI or any compatible application. However, repositories on your seed node cannot be browsed or viewed without cloning them first, using Radicle. To enable web browsing of the content, the Radicle HTTP Daemon radicle-httpd needs to be deployed alongside radicle-node.

The HTTP Daemon is a background process that functions as a gateway between the Radicle protocol and the HTTP protocol. It is configured to have direct read-only access to the node’s storage and database and expose this data via an HTTP JSON API. For seed nodes, the HTTP Daemon is always configured as a read-only service over the node’s state.

Installation

Head over to the download page, and follow the instructions there. The process is the same as for the Radicle Node. You will have to download, verify and extract the binary (radicle-httpd) and manuals to your preferred location.

We recommend installing the daemon under /usr/local, just as we did for the node.

Configuring your HTTP daemon’s system service

As with radicle-node, we can start by downloading an example systemd unit file for the daemon:

curl -sS https://seed.radicle.xyz/raw/rad:z3gqcJUoA1n9HaHKufZs5FCSGazv5/570a7eb141b6ba001713c46345d79b6fead1ca15/systemd/radicle-httpd.service -o /etc/systemd/system/radicle-httpd.service

Make sure it fits your needs by editing the file directly, or creating an override using systemctl edit.

Then, enable and run the service:

systemctl enable --now radicle-httpd

You can check that your node service is running with:

systemctl status radicle-httpd

You can query the API with curl to ensure everything is working properly:

curl http://127.0.0.1:8080/api/v1

Adding support for HTTPS

For your HTTP Daemon to be accessible from a frontend such as https://app.radicle.xyz, it needs to respond to HTTPS requests. To do this, we recommend using Caddy.

Start by installing caddy; most linux distributions have a package you can install. If you are using Debian or Ubuntu, you can run:

apt-get install caddy

If you’re having trouble installing Caddy, check the installation guide. Once installed, run caddy version to ensure that everything was installed correctly.

caddy version
v2.6.4 h1:2hwYqiRwk1tf3VruhMpLcYTg+11fCdr8S3jhNAdnPy8=

Then, download the caddy unit file from Caddy’s GitHub repository:

curl https://raw.githubusercontent.com/caddyserver/dist/master/init/caddy.service -o /etc/systemd/system/caddy.service

Edit the file and change the User and Group attributes to seed, as we have for the other services. Also ensure that ExecStart and ExecReload are set to the correct path. You can find the path under which caddy is installed by running the which caddy command.

Finally, edit or create the Caddyfile at /etc/caddy/Caddyfile, and replace its contents with the following configuration, using the correct domain name for your seed:

seed.radicle.garden {
    reverse_proxy 127.0.0.1:8080
}

This will proxy all HTTPS requests from port 443 to your HTTP daemon running on port 8080. Make sure your firewall has port 443 open for incomning TCP connections.

Finally, enable and start the Caddy service:

systemctl enable --now caddy

If you encounter issues setting up Caddy, you can try following their guide instead.

If everything worked, you should now have HTTPS support for your daemon. To check, run the following command with your seed’s domain:

curl https://seed.radicle.garden/api/v1

You should now be able to visit your seed node via any Radicle web frontend as well. For example, https://app.radicle.xyz/nodes/seed.radicle.garden.

On some distributions, installing caddy will start the system service automatically. If you’re not able to connect to your HTTP daemon from the outside, try running systemctl reload caddy after you’ve updated the configuration.

You’re all set

If you got this far, congratulations, you now have a Radicle seed node up and running!

Come join us on our community chat and tell us about your seed node on the #seeds channel.

🎊👾