KubeROS Platform (Preview)

This repository contains the main components of the KubeROS platform, which is built on Django, with a strong community support. This platform can be easily deployed in any self-managed Kubernetes cluster, which acts as the main cluster in KubeROS.

KubeROS API server: handling API requests and interacting among various system elements.
Controllers: Monitors the system state, making necessary adjustments to maintain the desired state.
Integrated hardware resource manager: manages the available resources and provides the system state for the schedulers and controllers

To use the platform with the kuberos-cli tools, please refer to the CLI repository: kuberos-cli

Basic exmaples and explanation of the KubeROS deployment yaml file can be found here: ROS2 Basic Examples

Preliminary

Before installing the platform, you will need to create a Kubernetes cluster and do some configurations:

There are different distributions of Kubernetes and also many tools are available. For users who are not very familiar with Kubernetes, we provide the following step-by-step guide, you can use any of them: - Use Kubeadm manually setup the cluster: Creating K8s with Kubeadm - Use Kubespray to automatically initialize a new cluster: - Create a lightweight K3s cluster (Only work with FastDDS): Creating a K3s cluster

Due to the complexity and the diversity of both ROS2 and Kubernetes, we recommend using calico as the network plugin, to support both FastDDS and CycloneDDS as the communication middleware.

Deploy a KubeROS platform

Since the backend of KubeROS is built on Django, you can easily deploy it like a typical web backend in a Kubernetes cluster (Main cluster in your edge side).

The deployment yaml file can be found in the deployment directory and the installation guide is here: Deploy KubeROS backend into a Kubernetes cluster

Development

For development, we recommend using the DevContainer as your development environment. All required componenents such as PostgreSQL, Redis, Django are preconfigured and can be started out of box. For more details, see docker-compose.yaml

After starting the containers for the first time, you need to create a superuser in Django and initialize the database with the following commands:

python3 manage.py createsuperuser
python3 manage.py migrate

In the DevContainer, you should open two terminals and start the django dev server and celery workers for testing and debugging.

# Start dev server
./run_dev_server.sh
# Start the celery workers
cd /workspace/kuberos
celery -A settings worker -l info
# (Optional) Start beat services
celery -A settings beat -l info

For more details about the using of celery in KubeROS, refer to Using Celery as Task Executor

Conventions

To make the interface more clear between different components, we suggest using the following conventions throughout the development process: Conventions in KubeROS

Acknowledgments