Create a cluster using Fleet

This section will guide you through creating a cluster that utilizes ClusterClass using a GitOps workflow with Fleet.

This guide uses the examples repository.

Prerequisites

Rancher Manager cluster with Rancher Turtles installed
Cluster API providers installed for your scenario - we’ll be using the Docker infrastructure and RKE2 bootstrap/control plane providers in these instructions - see Initialization for common providers using Turtles' CAPIProvider

Configure Rancher Manager

Configure your Fleet repository

To simplify the process of cluster provisioning, we will be using a series of pre-configured examples that you can find in the repository https://github.com/rancher/turtles/tree/main/examples. By inspecting the contents of this repository, you will find:

An applications folder for additional apps that can be deployed using Fleet (CNI, CSI, CPI).
A clusterclasses folder with a bunch of ClusterClass examples.
A clusters/docker/rke2 folder with an example of Docker cluster definition.

If you prefer, you can create your own Fleet repository using the same base structure.

The cluster definitions will be imported into the Rancher Manager cluster (which is also acting as a Cluster API management cluster) using the Continuous Delivery feature (which uses Fleet).

There are 2 options to provide the configuration. The first is using the Rancher Manager UI and the second is by applying some YAML to your cluster. Both are covered below.

Import Cluster and ClusterClass Definitions

Using the Rancher Manager UI
Manually using kubectl

Go to Rancher Manager
Select Continuous Delivery from the menu:
Select fleet-local as the namespace from the top right
Select Git Repos from the sidebar
Click Add Repository
Enter clusters as the name
Get the HTTPS clone URL from your git repo
Add the URL into the Repository URL field
Change the branch name to main

Add the following paths to specify which directories to deploy:

Fleet paths configuration

examples/clusterclasses/docker/rke2
examples/applications/cni/calico
examples/applications/lb/docker
examples/clusters/docker/rke2

Click Next
Click Create
Click on the clusters name
Watch the resources become ready
Select Cluster Management from the menu
Check your cluster has been imported

Get the HTTPS clone URL from your git repo
Create a new file called repo.yaml

Add the following contents to the new file:

apiVersion: fleet.cattle.io/v1alpha1
kind: GitRepo
metadata:
  name: clusters
  namespace: fleet-local
spec:
  branch: main
  repo: https://github.com/rancher/turtles.git
  paths:
    - examples/clusterclasses/docker/rke2
    - examples/applications/cni/calico
    - examples/applications/lb/docker
    - examples/clusters/docker/rke2
  targets: []

Apply the file to the Rancher Manager cluster using kubectl:
```
kubectl apply -f repo.yaml
```
Go to Rancher Manager
Select Continuous Delivery from the sidebar
Select fleet-local as the namespace from the top right
Select Git Repos from the sidebar
Click on the clusters name
Watch the resources become ready
Select Cluster Management from the menu
Check your cluster has been imported