Skip to main content

Prerequisites

Prerequisites before install arcadia.

1. Prepare a Kubernetes Cluster​

info

The Kubernetes version must equal to or greater than v1.20.0.

To check the version: kubectl version --short.

If you do not have a cluster yet, create one by using one of the following provision tools:

To utilize GPUs in kubernetes cluster,you must instlal GPU-Operator.

Next we show you how to prepare a kubernetes cluster with kind

Create a dev cluster with Kind​

info
  1. Configure docker to use nvidia runtime

Edit /etc/docker/daemon.json to configure default docker runtime

{
    "default-runtime": "nvidia",
    "runtimes": {
        "nvidia": {
            "args": [],
            "path": "nvidia-container-runtime"
        }
    }
}

Restart docker when docker runtime updated

sudo systemctl restart docker
  1. Configure nvidia container runtime

set accept-nvidia-visible-devices-as-volume-mounts = true in /etc/nvidia-container-runtime/config.toml

  1. Install Kind CLI

Follow document to install kind to your local machine

  1. Create kind cluster kubeagi

At the root directory of kubeagi/arcadia.

Run command:

make kind

When kind created,link ldconfig to ldconfig.real in kind container.

docker exec -ti kubeagi-control-plane ln -s /sbin/ldconfig /sbin/ldconfig.real
  1. Install Nvidia GPU-Operator
helm repo add nvidia https://helm.ngc.nvidia.com/nvidia
helm repo update
helm install --generate-name \
     -n gpu-operator --create-namespace \
     nvidia/gpu-operator --set driver.enabled=false

Check the installation status:

kubectl get pods -ngpu-operator --watch

2. Install Kubebb​

See install kubebb for more detals.