Cluster API bootstrap provider kubeadm

What is the Cluster API bootstrap provider kubeadm?

Cluster API bootstrap provider Kubeadm (CABPK) is a component of Cluster API that is responsible of generating a cloud-init script to turn a Machine into a Kubernetes Node; this implementation uses kubeadm for kubernetes bootstrap.

Resources

• design doc
• cluster-api.sigs.k8s.io
• The Kubebuilder Book

How to build, deploy and test CABPK

CABPK is built using kubebuilder. Please refer to the The Kubebuilder Book for build and deploy in a cluster.

CABPK is only part of a whole and does not function in isolation. It requires coordination with the core provider, Cluster API (CAPI), and an infrastructure provider; see the CAEP document for more information. For the remainder of this document we will be referring to cluster-api-provider-docker (CAPD) as our infrastructure provider.

A convenient way to set up infrastructure for testing is to use kind as the platform cluster to install the controllers and CRDs into. Then use a tool like tilt or skaffold to manage your dev environment. A minimal example of a Tiltfile looks like this:

allow_k8s_contexts('kubernetes-admin@kubernetes')

controllers = {
    'capi': {
        'path': '../cluster-api',
        'image': 'gcr.io/k8s-staging-cluster-api/cluster-api-controller:dev',
    },
    'cabpk': {
        'path': './',
        'image': 'gcr.io/k8s-staging-cluster-api/cluster-api-bootstrap-provider-kubeadm:dev',
    },
    'capd': {
        'path': '../cluster-api-provider-docker',
        'image': 'gcr.io/k8s-staging-cluster-api/cluster-api-provider-docker:dev',
    },
}

for name, controller in controllers.items():
    command = '''sed -i'' -e 's@image: .*@image: '"{}"'@' ./{}/config/default/manager_image_patch.yaml'''.format(controller['image'], controller['path'])
    local(command)

    k8s_yaml(local('kustomize build ' + controller['path'] + '/config/default'))

    docker_build(controller['image'], controller['path'])

See capi-dev for an example of a more complex developemt environment using tilt.

How does CABPK work?

Once your test environment is in place, create a Cluster object and its corresponding DockerCluster infrastructure object.

kind: DockerCluster
apiVersion: infrastructure.cluster.x-k8s.io/v1alpha2
metadata:
  name: my-cluster-docker
---
kind: Cluster
apiVersion: cluster.x-k8s.io/v1alpha2
metadata:
  name: my-cluster
spec:
  infrastructureRef:
    kind: DockerCluster
    apiVersion: infrastructure.cluster.x-k8s.io/v1alpha2
    name: my-cluster-docker

Now you can start creating machines by defining a Machine, its corresponding DockerMachine object, and the KubeadmConfig bootstrap object.

kind: KubeadmConfig
apiVersion: bootstrap.cluster.x-k8s.io/v1alpha2
metadata:
  name: my-control-plane1-config
spec:
  initConfiguration:
    nodeRegistration:
      kubeletExtraArgs:
        eviction-hard: nodefs.available<0%,nodefs.inodesFree<0%,imagefs.available<0%
  clusterConfiguration:
    controllerManager:
      extraArgs:
        enable-hostpath-provisioner: "true"
---
kind: DockerMachine
apiVersion: infrastructure.cluster.x-k8s.io/v1alpha2
metadata:
  name: my-control-plane1-docker
---
kind: Machine
apiVersion: cluster.x-k8s.io/v1alpha2
metadata:
  name: my-control-plane1
  labels:
    cluster.x-k8s.io/cluster-name: my-cluster
    cluster.x-k8s.io/control-plane: "true"
    set: controlplane
spec:
  bootstrap:
    configRef:
      kind: KubeadmConfig
      apiVersion: bootstrap.cluster.x-k8s.io/v1alpha2
      name: my-control-plane1-config
  infrastructureRef:
    kind: DockerMachine
    apiVersion: infrastructure.cluster.x-k8s.io/v1alpha2
    name: my-control-plane1-docker
  version: "v1.14.2"

CABPK's main responsibility is to convert a KubeadmConfig bootstrap object into a cloud-init script that is going to turn a Machine into a Kubernetes Node using kubeadm.

The cloud-init script will be saved into the KubeadmConfig.Status.BootstrapData and then the infrastructure provider (CAPD in this example) will pick up this value and proceed with the machine creation and the actual bootstrap.

KubeadmConfig objects

The KubeadmConfig object allows full control of Kubeadm init/join operations by exposing raw InitConfiguration, ClusterConfiguration and JoinConfiguration objects.

CABPK will fill in some values if they are left empty with sensible defaults:

KubeadmConfig field	Default
clusterConfiguration.KubernetesVersion	Machine.Spec.Version [1]
clusterConfiguration.clusterName	Cluster.metadata.name
clusterConfiguration.controlPlaneEndpoint	Cluster.status.apiEndpoints[0]
clusterConfiguration.networking.dnsDomain	Cluster.spec.clusterNetwork.serviceDomain
clusterConfiguration.networking.serviceSubnet	Cluster.spec.clusterNetwork.service.cidrBlocks[0]
clusterConfiguration.networking.podSubnet	Cluster.spec.clusterNetwork.pods.cidrBlocks[0]
joinConfiguration.discovery	a short lived BootstrapToken generated by CABPK

IMPORTANT! overriding above defaults could lead to broken Clusters.

[1] if both clusterConfiguration.KubernetesVersion and Machine.Spec.Version are empty, the latest Kubernetes version will be installed (as defined by the default kubeadm behavior).

Examples

Valid combinations of configuration objects are:

• at least one of InitConfiguration and ClusterConfiguration for the first control plane node only
• JoinConfiguration for worker nodes and additional control plane nodes

Bootstrap control plane node:

kind: KubeadmConfig
apiVersion: bootstrap.cluster.x-k8s.io/v1alpha2
metadata:
  name: my-control-plane1-config
spec:
  initConfiguration:
    nodeRegistration:
      kubeletExtraArgs:
        eviction-hard: nodefs.available<0%,nodefs.inodesFree<0%,imagefs.available<0%
  clusterConfiguration:
    controllerManager:
      extraArgs:
        enable-hostpath-provisioner: "true"

Additional control plane nodes:

kind: KubeadmConfig
apiVersion: bootstrap.cluster.x-k8s.io/v1alpha2
metadata:
  name: my-control-plane2-config
spec:
  joinConfiguration:
    nodeRegistration:
      kubeletExtraArgs:
        eviction-hard: nodefs.available<0%,nodefs.inodesFree<0%,imagefs.available<0%
    controlPlane: {}

worker nodes:

kind: KubeadmConfig
apiVersion: bootstrap.cluster.x-k8s.io/v1alpha2
metadata:
  name: my-worker1-config
spec:
  joinConfiguration:
    nodeRegistration:
      kubeletExtraArgs:
        eviction-hard: nodefs.available<0%,nodefs.inodesFree<0%,imagefs.available<0%

Bootstrap Orchestration

CABPK supports multiple control plane machines initing at the same time. The generation of cloud-init scripts of different machines is orchestrated in order to ensure a cluster bootstrap process that will be compliant with the correct Kubeadm init/join sequence. More in detail:

1. cloud-config-data generation starts only after Cluster.InfrastructureReady flag is set to true.
2. at this stage, cloud-config-data will be generated for the first control plane machine even if multiple control plane machines are ready (kubeadm init).
3. after Cluster.metadata.Annotations[cluster.x-k8s.io/control-plane-ready] is set to true, the cloud-config-data for all the other machines are generated (kubeadm join/join —control-plane).

Certificate Management

The user can choose two approaches for certificate management:

1. provide required certificate authorities (CAs) to use for kubeadm init/kubeadm join --control-plane; such CAs should be provided as a Secrets objects in the management cluster.
2. let CABPK to generate the necessary Secrets objects with a self-signed certificate authority for kubeadm

TODO: Add more info about certificate secrets

Additional Features

The KubeadmConfig object supports customizing the content of the config-data:

• KubeadmConfig.Files specifies additional files to be created on the machine
• KubeadmConfig.PreKubeadmCommands specifies a list of commands to be executed before kubeadm init/join
• KubeadmConfig.PostKubeadmCommands same as above, but after kubeadm init/join
• KubeadmConfig.Users specifies a list of users to be created on the machine
• KubeadmConfig.NTP specifies NPT settings for the machine

Versioning, Maintenance, and Compatibility

• We follow Semantic Versioning (semver).
• Cluster API bootstrap provider kubeadm versioning is syncronized with Cluster API.
• The master branch is where development happens, this might include breaking changes.
• The release-X branches contain stable, backward compatible code. A new release-X branch is created at every major (X) release.

Get involved!

• Join the Cluster API discuss forum.

• Join the sig-cluster-lifecycle Google Group for access to documents and calendars.

• Join our Cluster API working group sessions

  • Weekly on Wednesdays @ 10:00 PT on Zoom
  • Previous meetings: [ notes | recordings ]

• Provider implementer office hours

  • Weekly on Tuesdays @ 12:00 PT (Zoom) and Wednesdays @ 15:00 CET (Zoom)
  • Previous meetings: [ notes ]

• Chat with us on Slack: #cluster-api