一、环境准备
我准备了三台Ubuntu 16.04虚拟机，各项参数如下：
| 节点 | IP地址 | CPU | 内存 |
|---|---|---|---|
| master | 192.168.0.158 | 4核 | 4GB |
| node1 | 192.168.0.159 | 1核 | 2GB |
| node2 | 192.168.0.160 | 1核 | 2GB |
Kubernetes官网上提到每台机器至少要有1GB内存，不然集群起来之后，留给运行在容器内的应用的内存就很少了。同时要保证所有机器之前的网络是互相连通的。
这里再说一下，一开始我只给master分配了2个核，但是等我把kubeadm跑起来后，docker在pull一些必需的镜像时会出现OutOfCPU的情况，大概是因为kubeadm把第一台机器做master以及第一个node，所以初始化的服务都跑在这台机器上？所以后来我把master加到了4核。
二、搭建步骤
（1/4）安装docker、kubelet、kubeadm和kubectl
以root用户ssh到每台机器上，运行：
curl -s https://packages.cloud.google.com/apt/doc/apt-key.gpg | apt-key add -
cat <<EOF > /etc/apt/sources.list.d/kubernetes.list
deb http://apt.kubernetes.io/ kubernetes-xenial main
EOF
apt-get update
# 先安装docker
apt-get install -y docker.io
apt-get install -y kubelet kubeadm kubectl kubernetes-cni（2/4）初始化master
装好kubelet和kubectl后，在master上以root用户运行：
kubeadm init --token=yiqian.123456注意： 官方文档上kubeadm init的时候是不带token的，但我这样运行之后，console也没给我一个默认的token(exo me???)。但其他节点在加入集群的时候是需要指定master的token的，所以我在kubeadm初始化的时候手动给master指定了一个token。
这里还有一个坑就是，如果初始化失败了，比如像Issue#33544，那么你停掉再重新init会出错，这是因为kubeadm在安装了相关包之后会生成/etc/kubernetes、/var/lib/kubelet等目录，所以我们先需要手动清理一下再重新初始化：
systemctl stop kubelet;
docker rm -f $(docker ps -q); mount | grep "/var/lib/kubelet/*" | awk '{print $3}' | xargs umount 1>/dev/null 2>/dev/null;
rm -rf /var/lib/kubelet /etc/kubernetes /var/lib/etcd /etc/cni;
systemctl start kubelet;
kubeadm init --token=<token>最后，init成功的运行输出如下：
root@xyq-k8s-master:/home/administrator# kubeadm init --token=yiqian.123456
[kubeadm] WARNING: kubeadm is in alpha, please do not use it for production clusters.
[preflight] Running pre-flight checks
[init] Using Kubernetes version: v1.5.1
[tokens] Validating provided token
[tokens] Accepted provided token
[certificates] Generated Certificate Authority key and certificate.
[certificates] Generated API Server key and certificate
[certificates] Generated Service Account signing keys
[certificates] Created keys and certificates in "/etc/kubernetes/pki"
[kubeconfig] Wrote KubeConfig file to disk: "/etc/kubernetes/kubelet.conf"
[kubeconfig] Wrote KubeConfig file to disk: "/etc/kubernetes/admin.conf"
[apiclient] Created API client, waiting for the control plane to become ready
[apiclient] All control plane components are healthy after 14.534854 seconds
[apiclient] Waiting for at least one node to register and become ready
[apiclient] First node is ready after 0.506558 seconds
[apiclient] Creating a test deployment
[apiclient] Test deployment succeeded
[token-discovery] Created the kube-discovery deployment, waiting for it to become ready
[token-discovery] kube-discovery is ready after 2.005108 seconds
[addons] Created essential addon: kube-proxy
[addons] Created essential addon: kube-dns
Your Kubernetes master has initialized successfully!
You should now deploy a pod network to the cluster.
Run "kubectl apply -f [podnetwork].yaml" with one of the options listed at:
http://kubernetes.io/docs/admin/addons/
You can now join any number of machines by running the following on each node:
kubeadm join --token=yiqian.123456 192.168.0.158这个时候呢，我们可以通过kubectl看看集群当前的状态：
root@xyq-k8s-master:/home/administrator# kubectl get nodes
NAME STATUS AGE
xyq-k8s-master Ready,master 1m
root@xyq-k8s-master:/home/administrator# kubectl get pods --all-namespaces
NAMESPACE NAME READY STATUS RESTARTS AGE
kube-system dummy-2088944543-gjmv4 1/1 Running 0 1m
kube-system etcd-xyq-k8s-master 0/1 Pending 0 2s
kube-system kube-apiserver-xyq-k8s-master 1/1 Running 0 20s
kube-system kube-controller-manager-xyq-k8s-master 1/1 Running 0 22s
kube-system kube-discovery-1769846148-9dpvt 1/1 Running 0 1m
kube-system kube-dns-2924299975-lc4rr 0/4 ContainerCreating 0 1m
kube-system kube-proxy-w5v9m 1/1 Running 0 1m
kube-system kube-scheduler-xyq-k8s-master 0/1 Pending 0 4s可以看出来，除了kube-dns之外，其他pods都跑起来了。是因为在没有部署集群网络时，dns是起不来的。So我们进入下一步！
（3/4）初始化master
官方一共给出了五种网络addon，我选了Weave Net（因为它的安装步骤最方便诶嘿嘿=。=），直接运行：
kubectl apply -f https://git.io/weave-kube等weave跑起来后，kube-dns自然就跑起来啦。
root@xyq-k8s-master:/home/administrator# kubectl apply -f https://git.io/weave-kube
daemonset "weave-net" created
root@xyq-k8s-master:/home/administrator# kubectl get pods --all-namespaces
NAMESPACE NAME READY STATUS RESTARTS AGE
kube-system dummy-2088944543-gjmv4 1/1 Running 0 3m
kube-system etcd-xyq-k8s-master 1/1 Running 3 1m
kube-system kube-apiserver-xyq-k8s-master 1/1 Running 0 2m
kube-system kube-controller-manager-xyq-k8s-master 1/1 Running 0 2m
kube-system kube-discovery-1769846148-9dpvt 1/1 Running 0 3m
kube-system kube-dns-2924299975-lc4rr 4/4 Running 0 3m
kube-system kube-proxy-w5v9m 1/1 Running 0 3m
kube-system kube-scheduler-xyq-k8s-master 1/1 Running 0 1m
kube-system weave-net-8j1zg 2/2 Running 0 1m（4/4）将其他节点加入集群
分别ssh到node1和node2上去，以root用户运行：
kubeadm join --token=yiqian.123456 192.168.0.158再回到我们的master节点查看nodes，就会发现集群已经搭好了。
root@xyq-k8s-master:/home/administrator# kubectl get nodes
NAME STATUS AGE
xyq-k8s-master Ready,master 10m
xyq-k8s-s1 Ready 5m
xyq-k8s-s2 Ready 5m关于翻墙
由于我是在远程服务器上搭建虚拟机然后ssh上去操作，所以不能设置全局代理，只能把http proxy加到docker的配置文件/etc/default/docker中：
export http_proxy=<这个不能告诉你>
export https_proxy=<这个也不能告诉你>先去kubernetes源码中找1.5版本中各个组件的版本：
| 镜像名称 | 版本号 |
|---|---|
| gcr.io/google_containers/kubedns-amd64 | 1.7 |
| gcr.io/google_containers/kube-dnsmasq-amd64 | 1.3 |
| gcr.io/google_containers/exechealthz-amd64 | 1.1 |
然后创建一个github项目，可以fork我的repo或者原作者的repo。
最后在Docker Hub上分别创建以上三个镜像的自动构建项目。以kube-dns为例：
创建好之后要手动Trigger一下编译
等待编译成功之后就可以在本地直接pull了。
docker pull yiqianx/kubedns-amd64以上。
有问题欢迎大家评论指正~