* Suport calico 3.4.0

Signed-off-by: wangxf1987 <xiaofeix.wang@gmail.com>

* Remove symlink + cni conflist template when 3.3.0+, handle Canal, addition of install-cni: sidecar(3.3.0) or initontainer(3.4.0), KUBECONFIG_FILEPATH, calico_cert_dir, advertise cluster ips

* scheduler.alpha.kubernetes.io/critical-pod deprecated since 1.12

Addressing the discussion started in #4064, this PR moves kubeadm and
hyperkube binaries to /usr/local/bin before running them on the master
nodes.

It is to address the case where local_release_dir points to /tmp
(kubespray default) and /tmp is mounted with noexec mode, preventing
any binaries to be run in that partition.

In role "node", we still move kubeadm to bin_dir only on the worker
nodes.

* Mount host /run/xtables.lock in nodelocaldns container

* fix typo in nodelocaldns daemonset manifest yml

* Add prometheus scrape annotation, updateStrategy and reduce termination grace period

* fix indentation

* actually fix it..

* Bump k8s-dns-node-cache tag to 1.15.1 (fixes https://github.com/kubernetes/dns/issues/282)

Supported by kubeadm v1.11

So that binary can be sourced from anywhere - not only github.

Looks like `epel_enabled` was not configured for the epel install in `bootstrap-centos.yml`. Also, there were no conditionals that would trigger bootstrap for RHEL.

The `docker-ce` 18.09.1 packaging missed an `After` dependency on containerd in the systemd service. Upstream PR: https://github.com/docker/docker-ce-packaging/pull/290

* Use external LB IP for external api endpoint

Use loadbalancer_apiserver.address instead of apiserver_loadbalancer_domain_name for kudadm init --apiserver-advertise-address argument

https://kubernetes.io/docs/reference/setup-tools/kubeadm/kubeadm-init/#options states apiserver-advertise-address needs to be a IPv4 or IPv6 address

* only use loadbalancer IP if it is defined

Helm v2.12.2 has fixes for a security vuln, and there have been several improvements since our last update.

* Generate external admin.conf with kubeadm

* Fix apiserver sans

Docker 18.09.1 is out and it includes some fixes that are quite critical for RHEL distros, details here: https://docs.docker.com/engine/release-notes/#18091

I found a potential use case where `writable` could be null and therfore
not treated like a boolean, so this adds an extra default statement to
avoid negating a non-boolean as boolean which would lead to undefined. refs #4020

Looks like the template is removing the trailing space between storage
class entries, and since CI only has one storage class we never hit this
issue. This change will prevent the yaml from printing on a single line
when multiple storage classes are defined.

--address deprecated

In v1beta1 of `ClusterConfiguration` the extraVolumes `writable` field was changed to `readOnly` and its boolean value must be negated.

Also, the json field for `useHyperKubeImage` was incorrectly capitalized.

Right now we're consistently getting warnings about kubelet not found in
path during `kubeadm init`. We fixed this for `kubeadm join` in #3342, and this brings the change to init
as well.

- Fixed an issue where storage class host directories were looped
through excessive target hosts
- Fixes examples in the LVP `README.md` to use nested dicts instead of a
list of dicts

* Makes local volume provisioner more dynamic

* Correct variable name in local storage provisioner defaults

* Updates external-provisioner readme

* Updates variable naming to be more clear, more documentation, fixes sample inventory

* Variable refactor, untangled some jinja2 loops

* Corrects variable name

* No variable substitution in dict keys, replaced with anchor

* Fixes default storage_classes dict, inline docs

* Fixes spelling in inline docs

* Addresses comments in review

* Updates all the defaults

* Fix failing CI task

* Fixes external provisioner daemonset

* allows to override the bind addresses for controller-manager and scheduler

Useful for Prometheus metrics monitoring

* Add bind addr override support in kubeadm/v1beta1

Adds support for override of bind addresses for controller-manager
and scheduler in kubeadm/v1beta1

* Move location of bind address vars

* Remove double declaration of schedulerExtraArgs

use api server loadbalancer ip if external loadbalancer is used - this fixes the broken kube-router deployment

* remove os_family variable from bootstrap-os

* quote the conditions another time to fix the syntax error

* Use master node taint same as kubeadm configuration v1alpha3 or before.

* Fix mixup http/https in bootstrap-debian.yml

* Update bootstrap-debian.yml

The change implemented in #3908 remove line breaks for supplementary
addresses in kubeadm SANs, causing errors in the config file and
failure to bring cluster up. This commit reimplement line breaks in
between supplementary addresses.

- 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