Support ipvs mode for kube-proxy

Fix can not use devicemapper driver

--etcd-quorum-read is depricated in kube >= 1.9

Upgrade to Kubernetes v1.9.2

Add missing group var default values to kubespray-defaults

Updated vSphere cloud provider config for Kubernetes >= v1.9.2 and added resource pool deployment variable

Vault should use cert auth for etcd

Simplify and update OpenStack cloud provider

Fix can not be used devicemapper driver

Make spliting system_search_domains more robust

Cloud resolvers are mandatory for hosts on GCE and OpenStack
clouds. The 8.8.8.8 alternative resolver was dropped because
there is already a default nameserver. The new var name
reflects the purpose better.

Also restart apiserver when modifying dns settings.

Bump kube-dns to 1.14.8

DNS entries generated from 'etcd_cert_alt_names' variable in etcd's
openssl.conf are not terminated by a newline.

This fixes issue #2207.

Support ipvs mode for kube-proxy

Add lib/modules to kube-proxy to enable LVS

Fix: always only one container got synced after download

import_tasks will consume far less memory, so it should be
used whenever it is compatible.

Network size large deploy documentation

Raise drain timeout to 5m

Enable etcd secure client to prevent etcdctl access without cert and key

Enable flannel hairpin mode

* iptables: flush nat table as well as filter table upon reset

* Indentation fix

Update dashboard version to v1.8.1

Support for disabling apiserver insecure port (the sequel)

Dnsmasq autoscaler image should be a variable

Fix for Issue #2141