Closes issue: https://github.com/kubespray/kargo/issues/479



Signed-off-by: Bogdan Dobrelya <bdobrelia@mirantis.com>

Multiple app deploy tools

Bootstrap

Reset replicacluster name of kube-dns-v19 back to kubedns

This broke upgraded clusters

Rename kube-dns back to kubedns

kubedns should stay named the same so that services which
depend on this name are not broken.

remove dependency on kpm for kubedns

Fix updating resolvconf

Move updating resolvconf to the network restart handler to
ensure changes applied to the /etc/resolv.conf.

Signed-off-by: Bogdan Dobrelya <bdobrelia@mirantis.com>

Issue 369

Remove search and nameserver entries from resolvconf base

Disable calicoctl from creating a default pool

Sometimes invoking calicoctl to create a pool also
creates a default pool, which causes errors in deploy.

Fix vagrant custom instance name prefix

These items conflict when they are provided also in head file
Fixes: #456

fix etcd checksum

Remove SecurityContextDeny API plugin

Add --no-sync to etcdctl member list

Rebase etcd to v3.0.6

Fixes #450

This is no longer recommended for use since K8s 1.2:
http://kubernetes.io/docs/admin/admission-controllers/#is-there-a-recommended-set-of-plug-ins-to-use

Fixes #447

Enable customization of calico-node docker image

upgrade weave version to 1.6.1

New vars: calico_node_image_repo and claico_node_image_tag
Defaults: calico/node and {{ calico_version }}, respectively

ensure bin dir for coreos before anything else