- Creates and defaults an ansible variable for every configuration option in the `kubeproxy.config.k8s.io/v1alpha1` type spec
  - Fixes vars that were orphaned by removing non-kubeadm
  - Fixes previously harcoded kubeadm values
- Introduces a `main` directory for role default files per component (requires ansible 2.6.0+)
  - Split out just `kube-proxy.yml` in this first effort
- Removes the kube-proxy server field patch task

We should continue to pull out other components from `main.yml` into their own defaults files as I did here for `defaults/main/kube-proxy.yml`. I hope for and will need others to join me in this refactoring across the project until each component config template has a matching role defaults file, with shared defaults in `kubespray-defaults` or `downloads`

* Update the version of pypy used on CoreOS bootstrap-os

* update the pip installation process on CoreOS

* Add scale master features

* Add certificate management with kubeadm

* Add kubeadm kubeconfig

* Fix ymalroles error

* fix upgrade cluster fialed

* force update cert and keys when you reconfigure cluster

The containerd service and socket files have been dropped from the
openSUSE docker package so we should not require them in the docker
service anymore. This makes the docker service file look similar to
the one shipped by the openSUSE package.

Signed-off-by: Markos Chandras <mchandras@suse.de>

Add support install python-pip from local yum repository if local yum repository exist.

Added a line documenting where to find acceptable values for the
`docker_version` setting. If you use a value that is not used as
a key value by `docker_versioned_pkg` the container-engine/docker
playbook will throw a "Unexpected templating type error". (e.g.
If you use '18.06.1' or '18.06.1-ce', neither of which is used
as a key value of `docker_versioned_pkg`, rather than '18.06',
you'll get an error when installing on Ubuntu 18.04.)

Webhook token auth:
https://kubernetes.io/docs/reference/access-authn-authz/authentication/#webhook-token-authentication

Fixes #3063.

Fix GPU node Scheduling

* controlPlaneEndpoint set up through load balancer should be possible  even in single master setups

Enable load balancer for single-master setups
Fixes an issue where single-master setups are not reachable using the usual admin.conf from outside the cluster. 

controlPlaneEndpoint set up through load balancer should be possible  even in single master setups

* add fix to other api versions

* remove obsolete check completely

* remove check, pass 2

* removes checks in client configuration

* delete 'and'

download v1.12.3 kubernetes images failed

Delete unused controlPlane for join node

Fix installation using CRIO about download images failed

it is used for join master or use --experimental-control-plane argments

- NOTE: it assumes that the 'latest' version is newer than 17.05

Upgrade kubernetes to v1.13.1

* Add images downloader to download roles

* Use single jinja2 templates

* add kube_version to templates

Signed-off-by: Ganesh Maharaj Mahalingam <ganesh.mahalingam@intel.com>

* add vars for cilium init container

* make yamllint happy

* add var cilium_init in downloads

Signed-off-by: Ganesh Maharaj Mahalingam <ganesh.mahalingam@intel.com>

since bad886ca, dns_mode is set to coredns by default instead of kubedns

* Metrics server resizer addon should target metrics server deployment

* Target metrics server deployment without version