capi-demo

CAPI Demo on Rancher Manager using Turtles

Requirements

• Rancher >=v2.9.x or Rancher Prime >= v3.0 installed
• Turtles operator installed in Rancher Local Cluster
• clusterctl is installed
• clusterawsadm is installed

AWS infra with RKE2 k8s - Steps

Build an AMI

Original docs here: https://github.com/rancher/cluster-api-provider-rke2/blob/main/image-builder/README.md

Prerequisites

• Hashicorp packer

• AWS packer plugin

  packer plugins install github.com/hashicorp/amazon

Build AMI

1. Checkout the https://github.com/rancher/cluster-api-provider-rke2/tree/main project locally
2. cd into the image-builder directory
3. edit aws/opensuse-leap-156.json with valid existing AMI
4. build command uses 156 not 155 (which is in the docs)

---

Initial setup

1. Install Rancher Turtles

2. Source some bash functions for the next steps

   source providers/aws/ready-aws.sh

3. Prep env variables

   prep_env [your-aws-ssh-key-name]

4. Setup IAM profile

   clusterawsadm bootstrap iam create-cloudformation-stack

5. Install the capa-system namespace

   kubectl apply -f providers/aws/ns.yaml

6. Generate the secret

   generate_secret

7. Install the Infrastucture provider

   kubectl apply -f providers/aws/InfrastructureProviderAWS.yaml

Create a cluster (requires completion of Initial Setup)

1. Source some bash functions for the next steps

   source providers/aws/ready-aws.sh

2. Prep env variables

   • NOTE: Do not include .pem extension of the key name

   prep_env [your-aws-ssh-key-name]

3. Create a cluster yaml configuration

   • NOTE: These methods create the cluster configuration and apply it with the active kubeconfig.

   create_cluster [cluster-name]

   • OR Create a cluster with a namespace

   create_cluster_in_namespace [cluster-name] [namespace]

4. Import the cluster(s) into rancher

   import_clusters_in_namespace [namespace]

CAPI UI Extension

1. Add a repository in Rancher

• Select radio button: Target: https URL generated by helm
• Index URL: https://rancher.github.io/capi-ui-extension/
• Save

Debugging

If things don't go as expected, look at the capa-controller-manager pod logs. From there, hopefully you can work your way through other resources to figure out what is missing/misconfigured/etc.

Additional demo points not yet setup

• Alternate CAPI infrastructure provider. e.g. Azure, VMWare, etc.
  • value of instances is that the infra provider is agnostic to capi
  • using managed k8s means greater variation in configuration because it involves infra, bootstrap and control plane provider variations.
• Rancher projects vs Tanzu workspaces (in Rancher Manger or Downstream clusters?)
• Auto deployment of SUSE Security, RBAC configurations and SUSE Observability
• Modify downstream cluster volume sizes as part of the initial creation step
• Use ClusterClass for improved templating demo

Cleanup

1. for each cluster created kubectl delete -f <cluster_name>.yaml
2. Remove <cluster>.yaml files from the project.
3. Remove namespaces created with clusters
4. Delete the aws secret kubectl delete secret aws-variables -n capa-system