Skip to main content

Configure Rift Service

Once you have configured the machine for either [Docker], [VM] or Bare Metal rental - install and configure Rift system service.

Sign Up & Activate the Account

To start using CloudRift, please sign up at cloudrift.ai.

Configuring Rentable Machines

info

All steps below are included in our setup script, which can be found on GitHub.

Prerequisites

Before you begin, ensure your system has the following tools installed. These are required for downloading scripts, managing permissions, and installing packages.

  • curl: For downloading scripts and files.
  • sudo: For running commands as root.
sudo apt-get update
sudo apt-get install -y curl sudo

CloudRift System Service

Any Linux distribution is supported, but we typically test on Ubuntu 22.04 and 24.04.

curl -L https://cloudrift.ai/install-rift-service.sh | sudo sh

1. CloudRift CLI

Install CloudRift CLI to configure the system service.

curl -L https://cloudrift.ai/install-rift.sh | sh

2. Configure Credentials

The recommended way to authenticate a node is with an API key. You can create one in the CloudRift Console or with the CLI on a machine where you're logged in:

rift api-key add --name my-node-key

Set the CLOUDRIFT_API_KEY environment variable for the rift service:

sudo systemctl edit rift

Add the following, then save:

[Service]
Environment="CLOUDRIFT_API_KEY=<your-api-key>"

Alternatively, you can store the API key in the credentials file at /root/.config/cloudrift/credentials.yml:

api_key: <your-api-key>
danger

Node authentication with an account email/password, personal access token, or token is deprecated and will be removed in a future release. Nodes using a deprecated method still work but log a warning. Please migrate to API keys.

You can also configure email/password credentials interactively using the rift command line tool. Note that this authentication method is deprecated for nodes:

sudo rift configure
note

We're running the rift system service as root, so we need to configure it as root.

info

If editing credentials.yml manually, ensure you have the correct permissions and YAML formatting.

3. Restart System Service and Check Status

Restart CloudRift system service and check the status of the service and Docker Engine.

sudo systemctl restart rift
sudo systemctl status rift
sudo systemctl status docker
tip

You can use client_setup.sh to install and configure rift service:

sudo client_setup.sh --only=rift

Testing CloudRift Setup

1. Install CloudRift CLI

Install CloudRift CLI on your local machine. Installation instructions vary depending on your OS.

Loading...

2. Configure Credentials

Configure your credentials using the rift command line tool by running:

rift configure

3. Inspect the Cluster

Inspect the cluster. You should see all machines that you've installed CloudRift service on:

rift cluster info

4. Run Your First Job

If you have configured the machine for Docker - you can test it immediately. Otherwise, contact us and we'll help you to configure the console for VM rental.

Run your first job. For example, to run a simple echo command:

rift docker run --rm alpine echo "Hello CloudRift"

Troubleshooting

  • Docker or NVIDIA driver issues: Ensure you have rebooted after installation or removal of drivers. Check nvidia-smi and docker info.
  • Permission errors: Make sure you are running commands as root or with sudo.
  • Service not starting: Check logs with sudo journalctl -u rift or sudo systemctl status rift.

Next Steps

For more information on how to use the CloudRift platform to lease your computational resources, please see our documentation or send a request to dc@cloudrift.ai.

We support a variety of integrations, e.g., marketplace, white-label website, custom integration, and more.