kubeadm Configuration (v1beta3)
Package v1beta3 defines the v1beta3 version of the kubeadm configuration file format. This version improves on the v1beta2 format by fixing some minor issues and adding a few new fields.
A list of changes since v1beta2:
- The deprecated "ClusterConfiguration.useHyperKubeImage" field has been removed. Kubeadm no longer supports the hyperkube image.
- The "ClusterConfiguration.DNS.Type" field has been removed since CoreDNS is the only supported DNS server type by kubeadm.
- Include "datapolicy" tags on the fields that hold secrets. This would result in the field values to be omitted when API structures are printed with klog.
- Add "InitConfiguration.SkipPhases", "JoinConfiguration.SkipPhases" to allow skipping a list of phases during kubeadm init/join command execution.
- Add "InitConfiguration.NodeRegistration.ImagePullPolicy" and "JoinConfiguration.NodeRegistration.ImagePullPolicy" to allow specifying the images pull policy during kubeadm "init" and "join". The value must be one of "Always", "Never" or "IfNotPresent". "IfNotPresent" is the default, which has been the existing behavior prior to this addition.
- Add "InitConfiguration.Patches.Directory", "JoinConfiguration.Patches.Directory" to allow the user to configure a directory from which to take patches for components deployed by kubeadm.
- Move the BootstrapToken* API and related utilities out of the "kubeadm" API group to a new group "bootstraptoken". The kubeadm API version v1beta3 no longer contains the BootstrapToken* structures.
Migration from old kubeadm config versions
- kubeadm v1.15.x and newer can be used to migrate from v1beta1 to v1beta2.
- kubeadm v1.22.x and newer no longer support v1beta1 and older APIs, but can be used to migrate v1beta2 to v1beta3.
- kubeadm v1.27.x and newer no longer support v1beta2 and older APIs,
The preferred way to configure kubeadm is to pass an YAML configuration file with the
--config option. Some of the
configuration options defined in the kubeadm config file are also available as command line flags, but only
the most common/simple use case are supported with this approach.
A kubeadm config file could contain multiple configuration types separated using three dashes (
kubeadm supports the following configuration types:
apiVersion: kubeadm.k8s.io/v1beta3 kind: InitConfiguration apiVersion: kubeadm.k8s.io/v1beta3 kind: ClusterConfiguration apiVersion: kubelet.config.k8s.io/v1beta1 kind: KubeletConfiguration apiVersion: kubeproxy.config.k8s.io/v1alpha1 kind: KubeProxyConfiguration apiVersion: kubeadm.k8s.io/v1beta3 kind: JoinConfiguration
To print the defaults for "init" and "join" actions use the following commands:
kubeadm config print init-defaults kubeadm config print join-defaults
The list of configuration types that must be included in a configuration file depends by the action you are
join) and by the configuration options you are going to use (defaults or advanced
If some configuration types are not provided, or provided only partially, kubeadm will use default values; defaults
provided by kubeadm includes also enforcing consistency of values across components when required (e.g.
--cluster-cidr flag on controller manager and
clusterCIDR on kube-proxy).
Users are always allowed to override default values, with the only exception of a small subset of setting with relevance for security (e.g. enforce authorization-mode Node and RBAC on api server).
If the user provides a configuration types that is not expected for the action you are performing, kubeadm will ignore those types and print a warning.
Kubeadm init configuration types
When executing kubeadm init with the
--config option, the following configuration types could be used:
InitConfiguration, ClusterConfiguration, KubeProxyConfiguration, KubeletConfiguration, but only one
between InitConfiguration and ClusterConfiguration is mandatory.
apiVersion: kubeadm.k8s.io/v1beta3 kind: InitConfiguration bootstrapTokens: ... nodeRegistration: ...
The InitConfiguration type should be used to configure runtime settings, that in case of kubeadm init are the configuration of the bootstrap token and all the setting which are specific to the node where kubeadm is executed, including:
NodeRegistration, that holds fields that relate to registering the new node to the cluster; use it to customize the node name, the CRI socket to use or any other settings that should apply to this node only (e.g. the node ip).
LocalAPIEndpoint, that represents the endpoint of the instance of the API server to be deployed on this node; use it e.g. to customize the API server advertise address.
apiVersion: kubeadm.k8s.io/v1beta3 kind: ClusterConfiguration networking: ... etcd: ... apiServer: extraArgs: ... extraVolumes: ... ...
The ClusterConfiguration type should be used to configure cluster-wide settings, including settings for:
networkingthat holds configuration for the networking topology of the cluster; use it e.g. to customize Pod subnet or services subnet.
etcd: use it e.g. to customize the local etcd or to configure the API server for using an external etcd cluster.
kube-apiserver, kube-scheduler, kube-controller-manager configurations; use it to customize control-plane components by adding customized setting or overriding kubeadm default settings.
apiVersion: kubeproxy.config.k8s.io/v1alpha1 kind: KubeProxyConfiguration ...
The KubeProxyConfiguration type should be used to change the configuration passed to kube-proxy instances deployed in the cluster. If this object is not provided or provided only partially, kubeadm applies defaults.
See https://kubernetes.io/docs/reference/command-line-tools-reference/kube-proxy/ or https://pkg.go.dev/k8s.io/kube-proxy/config/v1alpha1#KubeProxyConfiguration for kube-proxy official documentation.
apiVersion: kubelet.config.k8s.io/v1beta1 kind: KubeletConfiguration ...
The KubeletConfiguration type should be used to change the configurations that will be passed to all kubelet instances deployed in the cluster. If this object is not provided or provided only partially, kubeadm applies defaults.
See https://kubernetes.io/docs/reference/command-line-tools-reference/kubelet/ or https://pkg.go.dev/k8s.io/kubelet/config/v1beta1#KubeletConfiguration for kubelet official documentation.
Here is a fully populated example of a single YAML file containing multiple
configuration types to be used during a
kubeadm init run.
apiVersion: kubeadm.k8s.io/v1beta3 kind: InitConfiguration bootstrapTokens: - token: "9a08jv.c0izixklcxtmnze7" description: "kubeadm bootstrap token" ttl: "24h" - token: "783bde.3f89s0fje9f38fhf" description: "another bootstrap token" usages: - authentication - signing groups: - system:bootstrappers:kubeadm:default-node-token nodeRegistration: name: "ec2-10-100-0-1" criSocket: "/var/run/dockershim.sock" taints: - key: "kubeadmNode" value: "someValue" effect: "NoSchedule" kubeletExtraArgs: v: 4 ignorePreflightErrors: - IsPrivilegedUser imagePullPolicy: "IfNotPresent" localAPIEndpoint: advertiseAddress: "10.100.0.1" bindPort: 6443 certificateKey: "e6a2eb8581237ab72a4f494f30285ec12a9694d750b9785706a83bfcbbbd2204" skipPhases: - addon/kube-proxy --- apiVersion: kubeadm.k8s.io/v1beta3 kind: ClusterConfiguration etcd: # one of local or external local: imageRepository: "registry.k8s.io" imageTag: "3.2.24" dataDir: "/var/lib/etcd" extraArgs: listen-client-urls: "http://10.100.0.1:2379" serverCertSANs: - "ec2-10-100-0-1.compute-1.amazonaws.com" peerCertSANs: - "10.100.0.1" # external: # endpoints: # - "10.100.0.1:2379" # - "10.100.0.2:2379" # caFile: "/etcd/kubernetes/pki/etcd/etcd-ca.crt" # certFile: "/etcd/kubernetes/pki/etcd/etcd.crt" # keyFile: "/etcd/kubernetes/pki/etcd/etcd.key" networking: serviceSubnet: "10.96.0.0/16" podSubnet: "10.244.0.0/24" dnsDomain: "cluster.local" kubernetesVersion: "v1.21.0" controlPlaneEndpoint: "10.100.0.1:6443" apiServer: extraArgs: authorization-mode: "Node,RBAC" extraVolumes: - name: "some-volume" hostPath: "/etc/some-path" mountPath: "/etc/some-pod-path" readOnly: false pathType: File certSANs: - "10.100.1.1" - "ec2-10-100-0-1.compute-1.amazonaws.com" timeoutForControlPlane: 4m0s controllerManager: extraArgs: "node-cidr-mask-size": "20" extraVolumes: - name: "some-volume" hostPath: "/etc/some-path" mountPath: "/etc/some-pod-path" readOnly: false pathType: File scheduler: extraArgs: bind-address: "10.100.0.1" extraVolumes: - name: "some-volume" hostPath: "/etc/some-path" mountPath: "/etc/some-pod-path" readOnly: false pathType: File certificatesDir: "/etc/kubernetes/pki" imageRepository: "registry.k8s.io" clusterName: "example-cluster" --- apiVersion: kubelet.config.k8s.io/v1beta1 kind: KubeletConfiguration # kubelet specific options here --- apiVersion: kubeproxy.config.k8s.io/v1alpha1 kind: KubeProxyConfiguration # kube-proxy specific options here
Kubeadm join configuration types
kubeadm join with the
--config option, the JoinConfiguration type should be provided.
apiVersion: kubeadm.k8s.io/v1beta3 kind: JoinConfiguration ...
The JoinConfiguration type should be used to configure runtime settings, that in case of
are the discovery method used for accessing the cluster info and all the setting which are specific
to the node where kubeadm is executed, including:
nodeRegistration, that holds fields that relate to registering the new node to the cluster; use it to customize the node name, the CRI socket to use or any other settings that should apply to this node only (e.g. the node ip).
apiEndpoint, that represents the endpoint of the instance of the API server to be eventually deployed on this node.
BootstrapToken describes one bootstrap token, stored as a Secret in the cluster
BootstrapTokenString is a token of the format
abcdef.abcdef0123456789 that is used
for both validation of the practically of the API server from a joining node's point
of view and as an authentication method for the node in the bootstrap phase of
"kubeadm join". This token is and should be short-lived.
|No description provided.|
|No description provided.|
ClusterConfiguration contains cluster-wide configuration for a kubeadm cluster
The cluster name.
InitConfiguration contains a list of elements that is specific "kubeadm init"-only runtime
kubeadm init-only information. These fields are solely used the first time
kubeadm init runs.
After that, the information in the fields IS NOT uploaded to the
that is used by
kubeadm upgrade for instance. These fields must be omitempty.
JoinConfiguration contains elements describing a particular node.
APIEndpoint struct contains elements of API server instance deployed on a node.
APIServer holds settings necessary for API server deployments in the cluster
|(Members of |
BootstrapTokenDiscovery is used to set the options for bootstrap token based discovery
ControlPlaneComponent holds settings common to control plane component of the cluster
DNS defines the DNS addon that should be used in the cluster
|(Members of |
Discovery specifies the options for the kubelet to use during the TLS Bootstrap process.
Etcd contains elements describing Etcd configuration.
ExternalEtcd describes an external etcd cluster. Kubeadm has no knowledge of where certificate files live and they must be supplied.
FileDiscovery is used to specify a file or URL to a kubeconfig file from which to load cluster information.
HostPathMount contains elements describing volumes that are mounted from the host.
ImageMeta allows to customize the image used for components that are not originated from the Kubernetes/Kubernetes release process
JoinControlPlane contains elements describing an additional control plane instance to be deployed on the joining node.
LocalEtcd describes that kubeadm should run an etcd cluster locally
|(Members of |
ImageMeta allows to customize the container used for etcd.
Networking contains elements describing cluster's networking configuration
NodeRegistrationOptions holds fields that relate to registering a new control-plane or node to the cluster, either via "kubeadm init" or "kubeadm join"
Patches contains options related to applying patches to components deployed by kubeadm.
This page is automatically generated.
If you plan to report an issue with this page, mention that the page is auto-generated in your issue description. The fix may need to happen elsewhere in the Kubernetes project.