TODO(madhusdancs): 1. Make printSuccess prepend protocol/scheme to the IPs/hostnames. 1. Add a dry-run support. 2. Make all the API object names customizable.

Ex: federation-apiserver, federation-controller-manager, etc.

3. Make image name and tag customizable. 4. Separate etcd container from API server pod as a first step towards enabling HA. 5. Generate credentials of the following types for the API server:

i.  "known_tokens.csv"
ii. "basic_auth.csv"

6. Add the ability to customize DNS domain suffix. It should probably be derived

from cluster config.

7. Make etcd PVC size configurable. 8. Make API server and controller manager replicas customizable via the HA work.



View Source
const (
	APIServerCN                 = "federation-apiserver"
	ControllerManagerCN         = "federation-controller-manager"
	AdminCN                     = "admin"
	HostClusterLocalDNSZoneName = "cluster.local."
	APIServerNameSuffix         = "apiserver"
	CMNameSuffix                = "controller-manager"
	CredentialSuffix            = "credentials"
	KubeconfigNameSuffix        = "kubeconfig"

	// User name used by federation controller manager to make
	// calls to federation API server.
	ControllerManagerUser = "federation-controller-manager"

	// Name of the ServiceAccount used by the federation controller manager
	// to access the secrets in the host cluster.
	ControllerManagerSA = "federation-controller-manager"


This section is empty.


func NewCmdInit

func NewCmdInit(cmdOut io.Writer, config util.AdminConfig) *cobra.Command

NewCmdInit defines the `init` command that bootstraps a federation control plane inside a set of host clusters.


This section is empty.

Source Files