环境：

master：master   10.51.98.41
node1: node1    10.25.134.181
nide2 : node2   10.51.55.208

准备：

    基于主机名通信

    时间同步

    关闭防火墙

安装步骤：

    1，etcd cluster,仅master节点

    2，flannel，集群的所有节点

    3，配置k8s的master：仅master节点

        kubernetes-master

        启动服务：

        kube-apiserver,kube-scheduler,kube-controller-manager

    4,配置k8s的各Node节点：

        kubernetes-node

        先设定启动docker服务；

        启动的k8s服务：

            kube-proxy,kubelet

kubeadm:

    1.master，nodes:安装kubelet  ,kubeadm,docker
    2,master：kubeadm init  完成集群初始化
    3,nodes：kubeadm join  加入集群
        https://github.com/kubernetes/kubeadm/blob/master/docs/design/design_v1.10.md

【master节点】

包下载：

配置yum源

    下载阿里云的docker-ce源

    下载阿里云的kubernetes源

[kubernetes]
name=Kubernetes Repo
baseurl=https://mirrors.aliyun.com/kubernetes/yum/repos/kubernetes-el7-x86_64/
gpgcheck=1
gpgkey=https://mirrors.aliyun.com/kubernetes/yum/doc/yum-key.gpg
enabeld=1
    [root@ll-sas01 /etc/yum.repos.d]# rpm --import rpm-package-key.gpg
    [root@ll-sas01 /etc/yum.repos.d]# rpm --import yum-key.gpg

    同步源到两个node节点

    yum install docker kubelet kubeadm kubectl -y


初始化docker 启动

修改docker镜像源地址

编辑vim /usr/lib/systemd/system/docker.service 文件，添加下边两行

    Environment="HTTPS_PROXY=http://www.ik8s.io:10080"
    Environment="NO_PROXY=127.0.0.0/8,172.20.0.0/16"

确保下边两个文件值为1

    [root@ll-sas01 ~]# cat /proc/sys/net/bridge/bridge-nf-call-ip6tables
    1
    [root@ll-sas01 ~]# cat /proc/sys/net/bridge/bridge-nf-call-iptables
    1

修改/etc/sysconfig/kubelet

    kubeadm init --kubernetes-version=v1.11.3 --pod-network-cidr=10.244.0.0/16 --service-cidr=10.96.0.0/12 --ignore-preflight-errors=Swap

查看初始化拖下来的docker images

    [root@ll-sas01 ~]# docker images    
    REPOSITORY                                 TAG                 IMAGE ID            CREATED             SIZE
    k8s.gcr.io/kube-proxy-amd64                v1.11.3             be5a6e1ecfa6        7 days ago          97.8MB
    k8s.gcr.io/kube-apiserver-amd64            v1.11.3             3de571b6587b        7 days ago          187MB
    k8s.gcr.io/kube-controller-manager-amd64   v1.11.3             a710d6a92519        7 days ago          155MB
    k8s.gcr.io/kube-scheduler-amd64            v1.11.3             ca1f38854f74        7 days ago          56.8MB
    k8s.gcr.io/coredns                         1.1.3               b3b94275d97c        3 months ago        45.6MB
    k8s.gcr.io/etcd-amd64                      3.2.18              b8df3b177be2        5 months ago        219MB
    k8s.gcr.io/pause                           3.1                 da86e6ba6ca1        9 months ago        742kB
    kubeadm join 123.57.224.60:6443 --token eza7wy.j0jonwz244k05yyi --discovery-token-ca-cert-hash sha256:e548d233d4b7674e5061923039f3d9d27c0fca20fe2a1499e49c42d9781142f0

初始晚推荐使用

    mkdir -p $HOME/.kube    
    sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
    sudo chown $(id -u):$(id -g)  $HOME/.kube/config

【node1】

    [root@xjd-sas01 /etc/yum.repos.d]# rpm --import rpm-package-key.gpg    
    [root@xjd-sas01 /etc/yum.repos.d]# rpm --import yum-key.gpg
    [root@xjd-sas01 ~]# yum install docker-ce kubelet kubeadm
    [root@xjd-sas01 ~]# systemctl enable docker
        Created symlink from /etc/systemd/system/multi-user.target.wants/docker.service to /usr/lib/systemd/system/docker.service.

【node2】

    [root@xjd-sas02 /etc/yum.repos.d]# rpm --import rpm-package-key.gpg    
    [root@xjd-sas02 /etc/yum.repos.d]# rpm --import yum-key.gpg
    [root@xjd-sas02 ~]# yum install docker-ce kubelet kubeadm
    [root@xjd-sas02 ~]# systemctl enable docker
        Created symlink from /etc/systemd/system/multi-user.target.wants/docker.service to /usr/lib/systemd/system/docker.service.

[master]

    状态信息

        [root@ll-sas01 ~]# kubectl get cs        
        NAME                 STATUS    MESSAGE              ERROR
        scheduler            Healthy   ok
        controller-manager   Healthy   ok
        etcd-0               Healthy   {"health": "true"}
        [root@ll-sas01 ~]# kubectl get componentstatus    #cs是componentstatus的简写
        NAME                 STATUS    MESSAGE              ERROR
        scheduler            Healthy   ok
        controller-manager   Healthy   ok
        etcd-0               Healthy   {"health": "true"}

查看节点信息

    [root@ll-sas01 ~]# kubectl get nodes
    NAME       STATUS     ROLES     AGE       VERSION
    ll-sas01   NotReady   master    26m       v1.11.3

属于未就绪，因为没有网络没有就绪，各节点无法通信

    flannel文档地址

        https://github.com/coreos/flannel

    1.7+以上的kubernetes安装flannel方法:

    kubectl apply -f     https://raw.githubusercontent.com/coreos/flannel/master/Documentation/kube-flannel.yml
    clusterrole.rbac.authorization.k8s.io/flannel created
    clusterrolebinding.rbac.authorization.k8s.io/flannel created
    serviceaccount/flannel created
    configmap/kube-flannel-cfg created
    daemonset.extensions/kube-flannel-ds-amd64 created
    daemonset.extensions/kube-flannel-ds-arm64 created
    daemonset.extensions/kube-flannel-ds-arm created
    daemonset.extensions/kube-flannel-ds-ppc64le created
    daemonset.extensions/kube-flannel-ds-s390x created


查看flannel镜像有没有下载完成

    [root@ll-sas01 ~]# docker images    
    quay.io/coreos/flannel                     v0.10.0-amd64       f0fad859c909        7 months ago        44.6MB
    [root@ll-sas01 ~]# kubectl get nodes
    NAME       STATUS    ROLES     AGE       VERSION
    ll-sas01   Ready     master    33m       v1.11.3

可以看到STATUS 已经为Ready状态了

查看名称空间：

    [root@ll-sas01 ~]# kubectl get ns    
    NAME          STATUS    AGE
    default       Active    45m
    kube-public   Active    45m
    kube-system   Active    45m

查看节点所有pod 指定kube-system名称空间

    [root@ll-sas01 ~]# kubectl get pods -n kube-system    
    NAME                               READY     STATUS    RESTARTS   AGE
    coredns-78fcdf6894-4rrzv           1/1       Running   0          34m
    coredns-78fcdf6894-94b75           1/1       Running   0          34m
    etcd-ll-sas01                      1/1       Running   0          33m
    kube-apiserver-ll-sas01            1/1       Running   0          33m
    kube-controller-manager-ll-sas01   1/1       Running   0          33m
    kube-flannel-ds-amd64-fzf4r        1/1       Running   0          2m
    kube-proxy-hc4ww                   1/1       Running   0          34m
    kube-scheduler-ll-sas01            1/1       Running   0          33m

复制配置master文件到node1，2

    [root@ll-sas01 ~]# scp /usr/lib/systemd/system/docker.service xjd-sas01:/usr/lib/systemd/system/docker.service    
    docker.service                                                                                                        100% 1240     1.7MB/s   00:00
    [root@ll-sas01 ~]# scp /usr/lib/systemd/system/docker.service xjd-sas02:/usr/lib/systemd/system/docker.service
    docker.service                                                                                                        100% 1240     1.6MB/s   00:00
    [root@ll-sas01 ~]# scp /etc/sysc
    sysconfig/   sysctl.conf  sysctl.d/
    [root@ll-sas01 ~]# scp /etc/sysconfig/kubelet xjd-sas01:/etc/sysconfig/
    kubelet                                                                                                               100%   42    78.8KB/s   00:00
    [root@ll-sas01 ~]# scp /etc/sysconfig/kubelet xjd-sas02:/etc/sysconfig/
    kubelet                                                                                                               100%   42    84.3KB/s   00:00
    [node1/2]
    [root@xjd-sas01 ~]# systemctl start docker
    Warning: docker.service changed on disk. Run 'systemctl daemon-reload' to reload units.
    [root@xjd-sas01 ~]# systemctl daemon-reload
    [root@xjd-sas01 ~]# systemctl start docker
    [root@xjd-sas01 ~]# systemctl enable docker kubelet
    Created symlink from /etc/systemd/system/multi-user.target.wants/kubelet.service to /etc/systemd/system/kubelet.service.
    [root@xjd-sas01 ~]# kubeadm join 123.57.224.60:6443 --token eza7wy.j0jonwz244k05yyi --discovery-token-ca-cert-hash sha256:e548d233d4b7674e5061923039f3d9d27c0fca20fe2a1499e49c42d9781142f0
    [master]

查看节点，发现两个node添加进来了

    [root@ll-sas01 ~]# kubectl get nodes    
    NAME        STATUS     ROLES     AGE       VERSION
    ll-sas01    Ready      master    1h        v1.11.3
    xjd-sas01   NotReady   <none>    2m        v1.11.3
    xjd-sas02   NotReady   <none>    8s        v1.11.3

等待两个node节点docker images同步结束

    [root@ll-sas01 ~]# kubectl get pods -n kube-system -o wide    
    NAME                               READY     STATUS    RESTARTS   AGE       IP              NODE        NOMINATED NODE
    coredns-78fcdf6894-4rrzv           1/1       Running   0          1h        10.244.0.2      ll-sas01    <none>
    coredns-78fcdf6894-94b75           1/1       Running   0          1h        10.244.0.3      ll-sas01    <none>
    etcd-ll-sas01                      1/1       Running   0          1h        10.51.98.41     ll-sas01    <none>
    kube-apiserver-ll-sas01            1/1       Running   0          1h        10.51.98.41     ll-sas01    <none>
    kube-controller-manager-ll-sas01   1/1       Running   0          1h        10.51.98.41     ll-sas01    <none>
    kube-flannel-ds-amd64-cxfsq        1/1       Running   0          12m       10.51.55.208    xjd-sas02   <none>
    kube-flannel-ds-amd64-fzf4r        1/1       Running   0          40m       10.51.98.41     ll-sas01    <none>
    kube-flannel-ds-amd64-gvs77        1/1       Running   0          13m       10.25.134.181   xjd-sas01   <none>
    kube-proxy-hc4ww                   1/1       Running   0          1h        10.51.98.41     ll-sas01    <none>
    kube-proxy-r54nq                   1/1       Running   0          12m       10.51.55.208    xjd-sas02   <none>
    kube-proxy-skfqg                   1/1       Running   0          13m       10.25.134.181   xjd-sas01   <none>
    kube-scheduler-ll-sas01            1/1       Running   0          1h        10.51.98.41     ll-sas01    <none>

【node1,2】

    查看iamges下载结束：

    [root@xjd-sas01 ~]# docker images    
    REPOSITORY                    TAG                 IMAGE ID            CREATED             SIZE
    k8s.gcr.io/kube-proxy-amd64   v1.11.3             be5a6e1ecfa6        7 days ago          97.8MB
    quay.io/coreos/flannel        v0.10.0-amd64       f0fad859c909        7 months ago        44.6MB
    k8s.gcr.io/pause              3.1                 da86e6ba6ca1        9 months ago        742kB