test-voxpupuli-puppet-k8s

• .gitignore: from a Bolt project created following https://www.puppet.com/docs/bolt/latest/bolt_installing_modules.html.
• Provision 2 or more CentOS 9 machines.
• Configure local.yaml for your environment.
• Install modules by running this command:

  bolt module install --force
  

  On Windows, you can run:

  .\InstallModules.ps1
  

• Provision controller on desired machine by running this command:

  bolt plan run test_voxpupuli_puppet_k8s::myplan --targets <ControllerFqdn> --user <ControllerUsername> --password <ControllerPassword> --inventory inventory.yaml node_type=controller control_plane_url=https://<ControllerFqdn>:6443 etcd_servers=https://<ControllerFqdn>:2379 secret_16_char=<Secret16Char>
  

  On Windows, you can run:

  .\Build.ps1 -Target <ControllerFqdn> -User <ControllerUsername> -Pass <ControllerPassword> -NodeType controller -ControlPlaneUrl https://<ControllerFqdn>:6443 -EtcdServer https://<ControllerFqdn>:2379 -Secret16Char <Secret16Char>
  
  

  It should fails with this error:

  Err: /Stage[main]/K8s::Server::Resources::Bootstrap/Kubectl_apply[puppet:cluster-info:reader Role]: Could not evaluate: Execution of '/bin/kubectl --namespace kube-system --kubeconfig /root/.kube/config get Role puppet:cluster-info:reader --output json' returned 1: error: the server doesn't have a resource type "Role"
  

  Run the same command a second time, and it should succeed.
• On the controller machine, copy /root/.kube/config to ~/.kube/config. Eg:

  mkdir --parents $HOME/.kube
  sudo cp --interactive /srv/kubernetes/kube-controller-manager.kubeconf $HOME/.kube/config
  sudo chown $(id -u):$(id -g) $HOME/.kube/config
  

  • Check Kubernetes cluster by running this command:

    kubectl cluster-info
    

    It should list something like this:

    Kubernetes control plane is running at https://<ControllerFqdn>:6443
    CoreDNS is running at https://<ControllerFqdn>:6443/api/v1/namespaces/kube-system/services/kube-dns:dns/proxy
    
    To further debug and diagnose cluster problems, use 'kubectl cluster-info dump'.
    

  • Check Kubernetes namespaces by running this commands:

    kubectl get namespaces
    

    It should list something like this:

    NAME              STATUS   AGE
    default           Active   20m
    kube-node-lease   Active   20m
    kube-public       Active   20m
    kube-system       Active   20m
    

• Provision worker on desired machines by running this command:

  bolt plan run test_voxpupuli_puppet_k8s::myplan --targets <WorkerFqdn> --user <WorkerUsername> --password <WorkerPassword> --inventory inventory.yaml node_type=worker control_plane_url=https://<ControllerFqdn>:6443 etcd_servers=https://<ControllerFqdn>:2379 secret_16_char=<Secret16Char>
  

  On Windows, you can run:

  .\Build.ps1 -Target <WorkerFqdn> -User <WorkerUsername> -Pass <WorkerPassword> -NodeType worker -ControlPlaneUrl https://<ControllerFqdn>:6443 -EtcdServer https://<ControllerFqdn>:2379 -Secret16Char <Secret16Char>
  

• On the same controller machine as above:
  • Check Kubernetes nodes by running this commands:

    kubectl get nodes
    

    It should list something like this:

    NAME             STATUS   ROLES    AGE     VERSION
    <workerFqdn>     Ready    <none>   5h29m   v1.26.1
    

  • List the certificate signing request from the workers:

    kubectl get csr --sort-by=.metadata.creationTimestamp
    

    and approve them:

    kubectl certificate approve '<csr-id>'
    

    TODO: is there a way to automate that step?