Migrate to the KubeOneCluster API

[xmudrii]

What this PR does / why we need it:

Migrate to the KubeOneCluster API implemented in #353 and proposed in the following proposal.

This PR brings breaking changes as some fields are removed or renamed. The new API is a Kubernetes-style API and supports API versioning, which allows us to make changes to the API without affecting the end users and breaking the backwards compatibility.

The new API is called KubeOneCluster, the API group is kubeone.io and the initial version is v1alpha1.

Once this PR is merged it will not be possible to create KubeOne clusters using the old configuration file and API. The following changes must be made in order to migrate to the new API:

• apiVersion and kind must be added at the top of the configuration file:

apiVersion: kubeone.io/v1alpha1
kind: KubeOneCluster

• The following fields at the root level are renamed, so they follow the Kubernetes-style of naming fields and structures:
  • apiserver -> apiEndpoints
  • provider -> cloudProvider
  • network -> clusterNetwork
  • machine_controller -> machineController
• All hosts fields must be changed to the camel case:
  • public_address -> publicAddress
  • private_address -> privateAddress
  • ssh_port -> sshPort
  • ssh_username -> sshUsername
  • ssh_private_key_file -> sshPrivateKeyFile
  • ssh_agent_socket -> sshAgentSocket
• The apiserver structure is replaced with the apiEndpoint structure which contains the following fields:
  • host - address or hostname of the API endpoint (by default load balancer DNS)
  • port - port of the API (by default 6443)
• All cloudProvider (prev. provider) fields must be changed to the camel case:
  • cloud_config -> cloudConfig
• The network structure is renamed to clusterNetwork, all existing fields are renamed to follow the camel case and a new field is added:
  • pod_subnet -> podSubnet
  • service_subnet -> serviceSubnet
  • node_port_range -> nodePortRange
  • [NEW] serviceDomainName -> isn't used currently
• http_proxy and https_proxy fields in the proxy structure are renamed to drop the _proxy suffix and no_proxy is renamed to follow the camel case:
  • http_proxy -> http
  • https_proxy -> https
  • no_proxy -> noProxy
• The credentials field from machineController structure (prev. machine_controller) is moved to the root level.
• All fields in the features structure are renamed to follow the camel case:
  • pod_security_policy -> podSecurityPolicy
  • dynamic_audit_log -> dynamicAuditLog
  • metrics_server -> metricsServer
  • openid_connect -> openidConnect
• All openidConnect fields are renamed to follow the camel case.
• workers.Config is renamed to workers.providerSpec.

Changes affecting using KubeOne as a Go library:

• Structures are renamed to follow the camel case and some structures are changed or removed (see above points for more details),
• WorkerConfig.Config.CloudProviderSpec and WorkerConfig.Config.OperatingSystemSpec are taking json.RawMessage instead of map[string]interface{},
• All fields in the Features structure are now pointers. All fields for enabling the feature are called Enable and are type of bool (prev. pointer on bool),
• MachineController is now a pointer of MachineControllerConfig (prev. non-pointer MachineControllerConfig),
• MachineController.Deploy is now bool instead of pointer on bool.

The following changes should be considered before merging the PR:

• Make serviceDomainName working.
• Support more than one API endpoint (a follow-up).
• Is openidConnect a correct JSON tag or should it be openIdConnect?
• Fix a typo in apiEndpoint (should be apiEndpoints).
• Make sure that the cluster is updated after the hostname or operating system is changed.
• Remove the old API (pkg/config) (a follow-up).
• Build a migration tool for migrating from the old API to the new KubeOneCluster API (a follow-up).
• Update documentation and examples (a follow-up).

Which issue(s) this PR fixes (optional, in fixes #<issue number>(, fixes #<issue_number>, ...) format, will close the issue(s) when PR gets merged):
xref #259
The issue will be closed once the old API is removed.

Documentation:

Documentation is up to be updated.

Release note:

[BREAKING CHANGES] Migrate to the new KubeOneCluster API. Disable the old API.

/assign @kron4eg
cc @scheeles

/hold
until we don't agree on the final API structure

[xmudrii]

@kron4eg Removed the hack and committed the appropriate fix. ServiceDomainName is now usable as well.

[xmudrii]

/hold cancel
Issues will be created for other tasks once the PR is merged

[xmudrii]

/test pull-kubeone-build

[xmudrii]

/retest

[xmudrii]

One DNS flake and one DO API flake.
/retest

[kron4eg]

/lgtm
/approve

[kubermatic-bot]

LGTM label has been added.

Git tree hash: f6e7aa01f1b20bb2fa626e1d315cf33c0de46bc7

[kubermatic-bot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: kron4eg

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [kron4eg]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment