本地环境要求

vagrant 1.7
virtualbox 5.0

部署支持的kubernetes集群

一键部署,集成dns服务

结构：

主机	作用	服务
e1	etcd服务	etcd2
c1	kubernetes控制机器	kubelet、（proxy、apiserver、kube-scheduler、podMaster）容器
w1	kubernetes节点	kubelet、（proxy）容器

本地主机kubectl配置

下载kubernetes客户端

ARCH=darwin; wget https://storage.googleapis.com/kubernetes-release/release/v1.0.6/bin/$ARCH/amd64/kubectl

环境变量ARCH基于当前机器的操作系统，值有”linux”或者”darwin”

下载完成后执行以下命令：

$ chmod +x kubectl  
$ mv kubectl /usr/local/bin/kubectl

配置kubectl

$ kubectl config set-cluster vagrant --server=https://172.17.4.101:443 --certificate-authority=${PWD}/ssl/ca.pem
$ kubectl config set-credentials vagrant-admin --certificate-authority=${PWD}/ssl/ca.pem --client-key=${PWD}/ssl/admin-key.pem --client-certificate=${PWD}/ssl/admin.pem
$kubectl config set-context vagrant --cluster=vagrant --user=vagrant-admin
$ kubectl config use-context vagrant

客户端准备就绪，kubernetes集群启动完成之后可在当前宿主机上使用kubectl客户端了。

---

下载kubernetes集群安装程序

从github上下载基于vagrant的配置
coreos官方原版（原版需要本地机器能够直接访问gcr.io）

$ git clone https://github.com/coreos/coreos-kubernetes.git
$ cd coreos-kubernetes/multi-node/

中国本土改良版(解决了各种镜像翻墙问题，直接从hub.docker.com上获取):

$ git clone https://github.com/shenshouer/coreos-kubernetes.git
$ cd coreos-kubernetes/multi-node/

新增kubernetes v1.2.0-alpha.2版本支持

启动机器

复制config.rb.sample到config.rb并修改如下值：

\#$update_channel="alpha"
\#$controller_count=1
\#$controller_vm_memory=512
\#$worker_count=1
\#$worker_vm_memory=512
\#$etcd_count=1
\#$etcd_vm_memory=512

执行vagrant up启动集群

ging machine 'c1' up with 'virtualbox' provider...
Bringing machine 'w1' up with 'virtualbox' provider...
==> e1: Box 'coreos-alpha' could not be found. Attempting to find and install...
    e1: Box Provider: virtualbox
    e1: Box Version: >= 766.0.0
==> e1: Loading metadata for box 'http://alpha.release.core-os.net/amd64-usr/current/coreos_production_vagrant.json'
    e1: URL: http://alpha.release.core-os.net/amd64-usr/current/coreos_production_vagrant.json
==> e1: Adding box 'coreos-alpha' (v808.0.0) for provider: virtualbox
    e1: Downloading: http://alpha.release.core-os.net/amd64-usr/808.0.0/coreos_production_vagrant.box

注意： *启动过程中会去coreos官网下载core的coreos_production_vagrant.box，如果速度比较慢，请试用其他工具下载，下载完成之后再执行
vagrant box add coreos-alpha coreos_production_vagrant.box
将box添加到vagrant box列表中，然后在当前目录下执行vagrant up即可启动集群。*
默认配置启动完成之后状态如下：

sope:vagrant goyoo$ vagrant status
Current machine states:
e1                        running (virtualbox)
c1                        running (virtualbox)
w1                        running (virtualbox)
This environment represents multiple VMs. The VMs are all listed
above with their current state. For more information about a specific
VM, run `vagrant status NAME`.
sope:vagrant goyoo$

检查服务

使用vagrant ssh c1登陆到kubernetes集群的控制机器上。

使用系统日志查看命令journalctl -f --system可以看到系统正在龟速下载coreos官方的flannel镜像，以及一些其他的依赖镜像。等待吧!

Sep 23 04:10:49 c1 sdnotify-proxy[1370]: Unable to find image 'quay.io/coreos/flannel:0.5.3' locally
Sep 23 04:10:49 c1 dockerd[1225]: time="2015-09-23T04:10:49.419376085Z" level=info msg="POST /v1.20/images/create?fromImage=quay.io%2Fcoreos%2Fflannel&tag=0.5.3"
Sep 23 04:10:58 c1 sdnotify-proxy[1370]: Pulling repository quay.io/coreos/flannel
Sep 23 04:11:02 c1 systemd[1]: Started OpenSSH per-connection server daemon (10.0.2.2:65423).
Sep 23 04:11:02 c1 sshd[1384]: Accepted publickey for core from 10.0.2.2 port 65423 ssh2: RSA SHA256:1M4RzhMyWuFS/86uPY/ce2prh/dVTHW7iD2RhpquOZA
Sep 23 04:11:04 c1 sdnotify-proxy[1370]: ac385fc755d9: Layer already being pulled by another client. Waiting.
Sep 23 04:11:07 c1 systemd[1]: Starting Generate /run/coreos/motd...
Sep 23 04:11:07 c1 systemd[1]: Started Generate /run/coreos/motd.
Sep 23 04:12:07 c1 systemd[1]: Starting Generate /run/coreos/motd...
Sep 23 04:12:07 c1 systemd[1]: Started Generate /run/coreos/motd.
Sep 23 04:12:19 c1 systemd[1]: flanneld.service: Start operation timed out. Terminating.

等待所有需要的容器自动下载完成之后，集群将自动启动所有所需的基础服务。在当前宿主机上即可使用kubectl工具进行集群的操作了，如下：

sope:~ goyoo$ kubectl --namespace=kube-system get po -o wide
NAME                                   READY     STATUS    RESTARTS   AGE       NODE
kube-apiserver-172.17.4.101            1/1       Running   0          2h        172.17.4.101
kube-controller-manager-172.17.4.101   1/1       Running   0          2h        172.17.4.101
kube-dns-v9-xr2h8                      4/4       Running   1          2h        172.17.4.201
kube-podmaster-172.17.4.101            2/2       Running   2          2h        172.17.4.101
kube-proxy-172.17.4.101                1/1       Running   0          2h        172.17.4.101
kube-proxy-172.17.4.201                1/1       Running   0          2h        172.17.4.201
kube-scheduler-172.17.4.101            1/1       Running   0          2h        172.17.4.101

下一步部署Tectonic服务到集群