二进制部署Kubernetes集群参考文档（V1.15.0）

一、基础概念

创新互联-专业网站定制、快速模板网站建设、高性价比蒲县网站开发、企业建站全套包干低至880元,成熟完善的模板库,直接使用。一站式蒲县网站制作公司更省心,省钱,快速模板网站建设找我们，业务覆盖蒲县地区。费用合理售后完善，十载实体公司更值得信赖。

1、概念

Kubernetes（通常写成“k8s”）Kubernetes是Google开源的容器集群管理系统。其设计目标是在主机集群之间提供一个能够自动化部署、可拓展、应用容器可运营的平台。

Kubernetes通常结合docker容器工具工作，并且整合多个运行着docker容器的主机集群。

2、功能特性

a、自动化容器部署
b、自动化扩/缩容器规模
c、提供容器间的负载均衡
d、快速更新和快速回滚

3、相关组件说明

3.1、Master节点组件

master节点上主要运行四个组件：api-server、scheduler、controller-manager、etcd。

APIServer：APIServer负责对外提供RESTful的Kubernetes API服务，它是系统管理指令的统一入口，任何对资源进行增删改查的操作都给APIServer处理后再提交给etcd。

schedule：scheduler的职责很明确，就是负责调度pod到合适的Node上。如果把scheduler看成一个黑匣子，那么它的输入是pod和由多个Node组成的列表，输出是Pod和一

个Node的绑定，即将这个pod部署到这个Node上。Kubernetes目前提供了调度算法，但是同样也保了接口，用户可以根据自己的需求定义自己的调度算法。

controller-manager：如果说APIServer做的是“前台”的工作的话，那controller manager就是负责“后台”的。每个资源一般都对一个控制器，而controller manager就是

负责管理这些控制器的。比如我们通过APIServer创建一个pod，当这个pod创建成功后，APIServer的任务就算完成了。而后面保证Pod的状态始终和我们预期的一样的重任

就由controller manager去保证了。

etcd：etcd是一个高可用的键值存储系统，Kubernetes使用它来存储各个资源的状态，从而实现了Restful的API。

3.2、Node节点组件

每个Node节点主要由三个模块组成：kubelet、kube-proxy、runtime。

runtime：指的是容器运行环境，目前Kubernetes支持docker和rkt两种容器。

kube-proxy：该模块实现了Kubernetes中的服务发现和反向代理功能。反向代理方面：kube-proxy支持TCP和UDP连接转发，默认基于Round Robin算法将客户端流量转发到

与service对应的一组后端pod。服务发现方面，kube-proxy使用etcd的watch机制，监控集群中service和endpoint对象数据的动态变化，并且维护一个service到endpoint

的映射关系，从而保证了后端pod的IP变化不会对访问者造成影响。另外kube-proxy还支持session affinity。

kubelet：Kubelet是Master在每个Node节点上面的agent，是Node节点上面最重要的模块，它负责维护和管理该Node上面的所有容器但是如果容器不是通过Kubernetes创建

的，它并不会管理。本质上，它负责使Pod得运行状态与期望的状态一致。

3.3、pod

Pod是k8s进行资源调度的最小单位，每个Pod中运行着一个或多个密切相关的业务容器，这些业务容器共享这个Pause容器的IP和Volume，我们以这个不易死亡的Pause容器

作为Pod的根容器，以它的状态表示整个容器组的状态。一个Pod一旦被创建就会放到Etcd中存储，然后由Master调度到一个Node绑定，由这个Node上的Kubelet进行实例化。

每个Pod会被分配一个单独的Pod IP，Pod IP + ContainerPort 组成了一个Endpoint。

3.4、Service

Service其功能使应用暴露，Pods 是有生命周期的，也有独立的 IP 地址，随着 Pods 的创建与销毁，一个必不可少的工作就是保证各个应用能够感知这种变化。这就要提

到 Service 了，Service 是 YAML 或 JSON 定义的由 Pods 通过某种策略的逻辑组合。更重要的是，Pods 独立 IP 需要通过 Service 暴露到网络中。

二、安装部署

部署方式有多中，此篇文章我们采用二进制方式部署。

1、环境介绍

主机名	IP	安装软件包	系统版本
k8s-master	192.168.248.65	kube-apiserver,kube-controller-manager,kube-scheduler	Red Hat Enterprise Linux Server release 7.3
k8s-node1	192.168.248.66	etcd,kubelet,kube-proxy,flannel,docker	Red Hat Enterprise Linux Server release 7.3
k8s-node2	192.168.248.67	etcd,kubelet,kube-proxy,flannel,docker	Red Hat Enterprise Linux Server release 7.3
k8s-node3	192.168.248.68	etcd,kubelet,kube-proxy,flannel,docker	Red Hat Enterprise Linux Server release 7.3

软件部署版本及下载链接

版本

kubenetes version v1.15.0

etcd version v3.3.10

flannel version v0.11.0

下载链接

kubernetes网址：https://github.com/kubernetes/kubernetes/blob/master/CHANGELOG-1.15.md#v1150

server端二进制文件：https://dl.k8s.io/v1.15.0/kubernetes-server-linux-amd64.tar.gz

node端二进制文件：https://dl.k8s.io/v1.15.0/kubernetes-node-linux-amd64.tar.gz

etcd网址：https://github.com/etcd-io/etcd/releases

flannel网址：https://github.com/coreos/flannel/releases

2、服务器初始化环境准备

同步系统时间

# ntpdate time1.aliyun.com
# echo "*/5 * * * * /usr/sbin/ntpdate -s time1.aliyun.com" > /var/spool/cron/root

修改主机名

# hostnamectl --static set-hostname k8s-master
# hostnamectl --static set-hostname k8s-node1
# hostnamectl --static set-hostname k8s-node2
# hostnamectl --static set-hostname k8s-node3

添加hosts解析

[root@k8s-master ~]# cat /etc/hosts
192.168.248.65 k8s-master
192.168.248.66 k8s-node1
192.168.248.67 k8s-node2
192.168.248.68 k8s-node3

关闭并禁用firewalld及selinux

# systemctl stop firewalld
# systemctl disable firewalld
# setenforce 0
# vim /etc/sysconfig/selinux
  SELINUX=disabled

关闭swap

# swapoff -a && sysctl -w vm.swappiness=0
# sed -i '/ swap / s/^\(.*\)$/#\1/g' /etc/fstab

设置系统参数

# cat /etc/sysctl.d/kubernetes.conf
net.ipv4.ip_forward=1
net.ipv4.tcp_tw_recycle=0
vm.swappiness=0
vm.overcommit_memory=1
vm.panic_on_oom=0
fs.inotify.max_user_watches=89100
fs.file-max=52706963
fs.nr_open=52706963
net.ipv6.conf.all.disable_ipv6=1

3、kubernetes集群安装部署

所有node节点安装docker-ce

# wget -P /etc/yum.repos.d/ https://mirrors.aliyun.com/docker-ce/linux/centos/docker-ce.repo
# yum makecache
# yum install docker-ce-18.06.2.ce-3.el7 -y
# systemctl start docker && systemctl enable docker

创建安装目录

# mkdir /data/{install,ssl_config} -pv
# mkdir /data/ssl_config/{etcd,kubernetes} -pv
# mkdir /cloud/k8s/etcd/{bin,cfg,ssl} -pv
# mkdir /cloud/k8s/kubernetes/{bin,cfg,ssl} -pv

添加环境变量

vim /etc/profile
######Kubernetes########
export PATH=$PATH:/cloud/k8s/etcd/bin/:/cloud/k8s/kubernetes/bin/

4、创建ssl证书

下载证书生成工具

[root@k8s-master ~]# wget -P /usr/local/bin/ https://pkg.cfssl.org/R1.2/cfssl_linux-amd64
[root@k8s-master ~]# wget -P /usr/local/bin/ https://pkg.cfssl.org/R1.2/cfssljson_linux-amd64
[root@k8s-master ~]# wget -P /usr/local/bin/ https://pkg.cfssl.org/R1.2/cfssl-certinfo_linux-amd64 
[root@k8s-master ~]# mv /usr/local/bin/cfssl_linux-amd64 /usr/local/bin/cfssl
[root@k8s-master ~]# mv /usr/local/bin/cfssljson_linux-amd64 /usr/local/bin/cfssljson
[root@k8s-master ~]# mv /usr/local/bin/cfssl-certinfo_linux-amd64 /usr/local/bin/cfssl-certinfo
[root@k8s-master ~]# chmod +x /usr/local/bin/*

创建etcd相关证书

# etcd证书ca配置
[root@k8s-master etcd]# pwd
/data/ssl_config/etcd
[root@k8s-master etcd]# cat ca-config.json 
{
  "signing": {
    "default": {
      "expiry": "87600h"
    },
    "profiles": {
      "www": {
         "expiry": "87600h",
         "usages": [
            "signing",
            "key encipherment",
            "server auth",
            "client auth"
        ]
      }
    }
  }
}

# etcd ca配置文件
[root@k8s-master etcd]# cat ca-csr.json 
{
    "CN": "etcd CA",
    "key": {
        "algo": "rsa",
        "size": 2048
    },
    "names": [
        {
            "C": "CN",
            "L": "Beijing",
            "ST": "Beijing"
        }
    ]
}

# etcd server 证书
[root@k8s-master etcd]# cat server-csr.json 
{
    "CN": "etcd",
    "hosts": [
    "k8s-node3",
    "k8s-node2",
    "k8s-node1",
    "192.168.248.66",
    "192.168.248.67",
    "192.168.248.68"
    ],
    "key": {
        "algo": "rsa",
        "size": 2048
    },
    "names": [
        {
            "C": "CN",
            "L": "Beijing",
            "ST": "Beijing"
        }
    ]
}

# 生成etcd ca证书和私钥
# cfssl gencert -initca ca-csr.json | cfssljson -bare ca -
# cfssl gencert -ca=ca.pem -ca-key=ca-key.pem -config=ca-config.json -profile=www server-csr.json | cfssljson -bare server

创建kubernetes相关证书

# kubernetes 证书ca配置
[root@k8s-master kubernetes]# pwd
/data/ssl_config/kubernetes
[root@k8s-master kubernetes]# cat ca-config.json
{
  "signing": {
    "default": {
      "expiry": "87600h"
    },
    "profiles": {
      "kubernetes": {
         "expiry": "87600h",
         "usages": [
            "signing",
            "key encipherment",
            "server auth",
            "client auth"
        ]
      }
    }
  }
}

# 创建ca证书配置
[root@k8s-master kubernetes]# cat ca-csr.json 
{
    "CN": "kubernetes",
    "key": {
        "algo": "rsa",
        "size": 2048
    },
    "names": [
        {
            "C": "CN",
            "L": "Beijing",
            "ST": "Beijing",
            "O": "k8s",
            "OU": "System"
        }
    ]
}

# 生成API_SERVER证书
[root@k8s-master kubernetes]# cat server-csr.json 
{
    "CN": "kubernetes",
    "hosts": [
      "10.0.0.1",
      "127.0.0.1",
      "192.168.248.65",
      "k8s-master",
      "kubernetes",
      "kubernetes.default",
      "kubernetes.default.svc",
      "kubernetes.default.svc.cluster",
      "kubernetes.default.svc.cluster.local"
    ],
    "key": {
        "algo": "rsa",
        "size": 2048
    },
    "names": [
        {
            "C": "CN",
            "L": "Beijing",
            "ST": "Beijing",
            "O": "k8s",
            "OU": "System"
        }
    ]
}

# 创建 Kubernetes Proxy 证书
[root@k8s-master kubernetes]# cat kube-proxy-csr.json
{
  "CN": "system:kube-proxy",
  "hosts": [],
  "key": {
    "algo": "rsa",
    "size": 2048
  },
  "names": [
    {
      "C": "CN",
      "L": "Beijing",
      "ST": "Beijing",
      "O": "k8s",
      "OU": "System"
    }
  ]
}

# 生成ca证书
# cfssl gencert -initca ca-csr.json | cfssljson -bare ca -
# 生成 api-server 证书
# cfssl gencert -ca=ca.pem -ca-key=ca-key.pem -config=ca-config.json -profile=kubernetes server-csr.json | cfssljson -bare server
# 生成 kube-proxy 证书
# cfssl gencert -ca=ca.pem -ca-key=ca-key.pem -config=ca-config.json -profile=kubernetes kube-proxy-csr.json | cfssljson -bare kube-proxy

5、部署etcd集群(在所有node节点操作)

解压并配置etcd软件包

# tar -xvf etcd-v3.3.10-linux-amd64.tar.gz
# cp etcd-v3.3.10-linux-amd64/{etcd,etcdctl} /cloud/k8s/etcd/bin/

编写etcd配置文件

[root@k8s-node1 ~]# cat /cloud/k8s/etcd/cfg/etcd 
#[Member]
ETCD_NAME="etcd01"
ETCD_DATA_DIR="/var/lib/etcd/default.etcd"
ETCD_LISTEN_PEER_URLS="https://192.168.248.66:2380"
ETCD_LISTEN_CLIENT_URLS="https://192.168.248.66:2379"

#[Clustering]
ETCD_INITIAL_ADVERTISE_PEER_URLS="https://192.168.248.66:2380"
ETCD_ADVERTISE_CLIENT_URLS="https://192.168.248.66:2379"
ETCD_INITIAL_CLUSTER="etcd01=https://192.168.248.66:2380,etcd02=https://192.168.248.67:2380,etcd03=https://192.168.248.68:2380"
ETCD_INITIAL_CLUSTER_TOKEN="etcd-cluster"
ETCD_INITIAL_CLUSTER_STATE="new"

[root@k8s-node2 ~]# cat /cloud/k8s/etcd/cfg/etcd 
#[Member]
ETCD_NAME="etcd01"
ETCD_DATA_DIR="/var/lib/etcd/default.etcd"
ETCD_LISTEN_PEER_URLS="https://192.168.248.67:2380"
ETCD_LISTEN_CLIENT_URLS="https://192.168.248.67:2379"

#[Clustering]
ETCD_INITIAL_ADVERTISE_PEER_URLS="https://192.168.248.67:2380"
ETCD_ADVERTISE_CLIENT_URLS="https://192.168.248.67:2379"
ETCD_INITIAL_CLUSTER="etcd01=https://192.168.248.66:2380,etcd02=https://192.168.248.67:2380,etcd03=https://192.168.248.68:2380"
ETCD_INITIAL_CLUSTER_TOKEN="etcd-cluster"
ETCD_INITIAL_CLUSTER_STATE="new"

[root@k8s-node3 ~]# cat /cloud/k8s/etcd/cfg/etcd 
#[Member]
ETCD_NAME="etcd01"
ETCD_DATA_DIR="/var/lib/etcd/default.etcd"
ETCD_LISTEN_PEER_URLS="https://192.168.248.68:2380"
ETCD_LISTEN_CLIENT_URLS="https://192.168.248.68:2379"

#[Clustering]
ETCD_INITIAL_ADVERTISE_PEER_URLS="https://192.168.248.68:2380"
ETCD_ADVERTISE_CLIENT_URLS="https://192.168.248.68:2379"
ETCD_INITIAL_CLUSTER="etcd01=https://192.168.248.66:2380,etcd02=https://192.168.248.67:2380,etcd03=https://192.168.248.68:2380"
ETCD_INITIAL_CLUSTER_TOKEN="etcd-cluster"
ETCD_INITIAL_CLUSTER_STATE="new"

创建etcd启动文件

[root@k8s-node1 ~]# cat /usr/lib/systemd/system/etcd.service 
[Unit]
Description=Etcd Server
After=network.target
After=network-online.target
Wants=network-online.target

[Service]
Type=notify
EnvironmentFile=/cloud/k8s/etcd/cfg/etcd
ExecStart=/cloud/k8s/etcd/bin/etcd \
--name=${ETCD_NAME} \
--data-dir=${ETCD_DATA_DIR} \
--listen-peer-urls=${ETCD_LISTEN_PEER_URLS} \
--listen-client-urls=${ETCD_LISTEN_CLIENT_URLS},http://127.0.0.1:2379 \
--advertise-client-urls=${ETCD_ADVERTISE_CLIENT_URLS} \
--initial-advertise-peer-urls=${ETCD_INITIAL_ADVERTISE_PEER_URLS} \
--initial-cluster=${ETCD_INITIAL_CLUSTER} \
--initial-cluster-token=${ETCD_INITIAL_CLUSTER_TOKEN} \
--initial-cluster-state=new \
--cert-file=/cloud/k8s/etcd/ssl/server.pem \
--key-file=/cloud/k8s/etcd/ssl/server-key.pem \
--peer-cert-file=/cloud/k8s/etcd/ssl/server.pem \
--peer-key-file=/cloud/k8s/etcd/ssl/server-key.pem \
--trusted-ca-file=/cloud/k8s/etcd/ssl/ca.pem \
--peer-trusted-ca-file=/cloud/k8s/etcd/ssl/ca.pem
Restart=on-failure
LimitNOFILE=65536

[Install]
WantedBy=multi-user.target

将生成的etcd证书文件拷贝到所有node节点

[root@k8s-master etcd]# pwd
/data/ssl_config/etcd
[root@k8s-master etcd]# scp *.pem k8s-node1:/cloud/k8s/etcd/ssl/
[root@k8s-master etcd]# scp *.pem k8s-node2:/cloud/k8s/etcd/ssl/
[root@k8s-master etcd]# scp *.pem k8s-node3:/cloud/k8s/etcd/ssl/

启动etcd集群服务

systemctl daemon-reload
systemctl enable etcd
systemctl start etcd

查看启动状态(任意一个node节点上执行即可)

[root@k8s-node1 ssl]# etcdctl --ca-file=/cloud/k8s/etcd/ssl/ca.pem --cert-file=/cloud/k8s/etcd/ssl/server.pem --key-file=/cloud/k8s/etcd/ssl/server-key.pem --endpoints="https://192.168.248.66:2379,https://192.168.248.67:2379,https://192.168.248.68:2379"  cluster-health
member 2830381866015ef6 is healthy: got healthy result from https://192.168.248.67:2379
member 355a96308320dc2a is healthy: got healthy result from https://192.168.248.66:2379
member a9a44d5d05a31ce0 is healthy: got healthy result from https://192.168.248.68:2379
cluster is healthy

6、部署flannel网络(所有node节点)

向etcd集群中写入pod网段信息(任意一台node节点上执行)

etcdctl --ca-file=/cloud/k8s/etcd/ssl/ca.pem \
--cert-file=/cloud/k8s/etcd/ssl/server.pem \
--key-file=/cloud/k8s/etcd/ssl/server-key.pem  \
--endpoints="https://192.168.248.66:2379,https://192.168.248.67:2379,https://192.168.248.68:2379"  \
set /coreos.com/network/config '{ "Network": "172.18.0.0/16", "Backend": {"Type": "vxlan"}}'

查看写入etcd集群中的网段信息

# etcdctl --ca-file=/cloud/k8s/etcd/ssl/ca.pem \
--cert-file=/cloud/k8s/etcd/ssl/server.pem \
--key-file=/cloud/k8s/etcd/ssl/server-key.pem  \
--endpoints="https://192.168.248.66:2379,https://192.168.248.67:2379,https://192.168.248.68:2379"  \
get /coreos.com/network/config

# [root@k8s-node1 ssl]# etcdctl --ca-file=/cloud/k8s/etcd/ssl/ca.pem \
> --cert-file=/cloud/k8s/etcd/ssl/server.pem \
> --key-file=/cloud/k8s/etcd/ssl/server-key.pem  \
> --endpoints="https://192.168.248.66:2379,https://192.168.248.67:2379,https://192.168.248.68:2379" \
> ls /coreos.com/network/subnets
/coreos.com/network/subnets/172.18.95.0-24
/coreos.com/network/subnets/172.18.22.0-24
/coreos.com/network/subnets/172.18.54.0-24

解压并配置flannel网络插件

# tar xf  flannel-v0.11.0-linux-amd64.tar.gz
# mv flanneld mk-docker-opts.sh /cloud/k8s/kubernetes/bin/

配置flannel

[root@k8s-node1 cfg]# cat /cloud/k8s/kubernetes/cfg/flanneld 
FLANNEL_OPTIONS="--etcd-endpoints=https://192.168.248.66:2379,https://192.168.248.67:2379,https://192.168.248.68:2379 -etcd-cafile=/cloud/k8s/etcd/ssl/ca.pem -etcd-certfile=/cloud/k8s/etcd/ssl/server.pem -etcd-keyfile=/cloud/k8s/etcd/ssl/server-key.pem"

配置flanneld启动文件

[root@k8s-node1 cfg]# cat /usr/lib/systemd/system/flanneld.service 
[Unit]
Description=Flanneld overlay address etcd agent
After=network-online.target network.target
Before=docker.service

[Service]
Type=notify
EnvironmentFile=/cloud/k8s/kubernetes/cfg/flanneld
ExecStart=/cloud/k8s/kubernetes/bin/flanneld --ip-masq $FLANNEL_OPTIONS
ExecStartPost=/cloud/k8s/kubernetes/bin/mk-docker-opts.sh -k DOCKER_NETWORK_OPTIONS -d /run/flannel/subnet.env
Restart=on-failure

[Install]
WantedBy=multi-user.target

配置Docker启动指定子网段

[root@k8s-node1 cfg]# cat /usr/lib/systemd/system/docker.service 
[Unit]
Description=Docker Application Container Engine
Documentation=https://docs.docker.com
After=network-online.target firewalld.service
Wants=network-online.target

[Service]
Type=notify
EnvironmentFile=/run/flannel/subnet.env
# the default is not to use systemd for cgroups because the delegate issues still
# exists and systemd currently does not support the cgroup feature set required
# for containers run by docker
ExecStart=/usr/bin/dockerd $DOCKER_NETWORK_OPTIONS
ExecReload=/bin/kill -s HUP $MAINPID
# Having non-zero Limit*s causes performance problems due to accounting overhead
# in the kernel. We recommend using cgroups to do container-local accounting.
LimitNOFILE=infinity
LimitNPROC=infinity
LimitCORE=infinity
# Uncomment TasksMax if your systemd version supports it.
# Only systemd 226 and above support this version.
#TasksMax=infinity
TimeoutStartSec=0
# set delegate yes so that systemd does not reset the cgroups of docker containers
Delegate=yes
# kill only the docker process, not all processes in the cgroup
KillMode=process
# restart the docker process if it exits prematurely
Restart=on-failure
StartLimitBurst=3
StartLimitInterval=60s

[Install]
WantedBy=multi-user.target

启动服务

systemctl daemon-reload
systemctl start flanneld
systemctl enable flanneld
systemctl restart docker

验证fiannel网络配置

node节点之间相互ping测docker0网卡的ip地址，能ping通说明flanneld网络插件部署成功。

7、部署master节点组件

解压master节点安装包

# tar xf kubernetes-server-linux-amd64.tar.gz
# cp kubernetes//server/bin/{kube-scheduler,kube-apiserver,kube-controller-manager,kubectl}  /cloud/k8s/kubernetes/bin/

配置kubernetes相关证书

# cp /data/ssl_config/kubernetes/*.pem /cloud/k8s/kubernetes/ssl/

部署 kube-apiserver 组件

创建 TLS Bootstrapping Token

[root@k8s-master cfg]# head -c 16 /dev/urandom | od -An -t x | tr -d ' '   #生成随机字符串
[root@k8s-master cfg]# pwd
/cloud/k8s/kubernetes/cfg
[root@k8s-master cfg]# cat token.csv 
a081e7ba91d597006cbdacfa8ee114ac,kubelet-bootstrap,10001,"system:kubelet-bootstrap"

apiserver配置文件

[root@k8s-master cfg]# cat kube-apiserver 
KUBE_APISERVER_OPTS="--logtostderr=true \
--v=4 \
--etcd-servers=https://192.168.248.66:2379,https://192.168.248.67:2379,https://192.168.248.68:2379 \
--bind-address=192.168.248.65 \
--secure-port=6443 \
--advertise-address=192.168.248.65 \
--allow-privileged=true \
--service-cluster-ip-range=10.0.0.0/24 \
--admission-control=NamespaceLifecycle,LimitRanger,SecurityContextDeny,ServiceAccount,ResourceQuota,NodeRestriction \
--authorization-mode=RBAC,Node \
--enable-bootstrap-token-auth \
--token-auth-file=/cloud/k8s/kubernetes/cfg/token.csv \
--service-node-port-range=30000-50000 \
--tls-cert-file=/cloud/k8s/kubernetes/ssl/server.pem  \
--tls-private-key-file=/cloud/k8s/kubernetes/ssl/server-key.pem \
--client-ca-file=/cloud/k8s/kubernetes/ssl/ca.pem \
--service-account-key-file=/cloud/k8s/kubernetes/ssl/ca-key.pem \
--etcd-cafile=/cloud/k8s/etcd/ssl/ca.pem \
--etcd-certfile=/cloud/k8s/etcd/ssl/server.pem \
--etcd-keyfile=/cloud/k8s/etcd/ssl/server-key.pem"

kube-apiserver启动文件

[root@k8s-master cfg]# cat /usr/lib/systemd/system/kube-apiserver.service 
[Unit]
Description=Kubernetes API Server
Documentation=https://github.com/kubernetes/kubernetes

[Service]
EnvironmentFile=-/cloud/k8s/kubernetes/cfg/kube-apiserver
ExecStart=/cloud/k8s/kubernetes/bin/kube-apiserver $KUBE_APISERVER_OPTS
Restart=on-failure

[Install]
WantedBy=multi-user.target

启动kube-apiserver服务

[root@k8s-master cfg]# systemctl daemon-reload
[root@k8s-master cfg]# systemctl enable kube-apiserver
[root@k8s-master cfg]# systemctl start kube-apiserver
[root@k8s-master cfg]# ps -ef |grep kube-apiserver
root       1050      1  4 09:02 ?        00:25:21 /cloud/k8s/kubernetes/bin/kube-apiserver --logtostderr=true --v=4 --etcd-servers=https://192.168.248.66:2379,https://192.168.248.67:2379,https://192.168.248.68:2379 --bind-address=192.168.248.65 --secure-port=6443 --advertise-address=192.168.248.65 --allow-privileged=true --service-cluster-ip-range=10.0.0.0/24 --admission-control=NamespaceLifecycle,LimitRanger,SecurityContextDeny,ServiceAccount,ResourceQuota,NodeRestriction --authorization-mode=RBAC,Node --enable-bootstrap-token-auth --token-auth-file=/cloud/k8s/kubernetes/cfg/token.csv --service-node-port-range=30000-50000 --tls-cert-file=/cloud/k8s/kubernetes/ssl/server.pem --tls-private-key-file=/cloud/k8s/kubernetes/ssl/server-key.pem --client-ca-file=/cloud/k8s/kubernetes/ssl/ca.pem --service-account-key-file=/cloud/k8s/kubernetes/ssl/ca-key.pem --etcd-cafile=/cloud/k8s/etcd/ssl/ca.pem --etcd-certfile=/cloud/k8s/etcd/ssl/server.pem --etcd-keyfile=/cloud/k8s/etcd/ssl/server-key.pem
root       1888   1083  0 18:15 pts/0    00:00:00 grep --color=auto kube-apiserve

部署kube-scheduler组件

创建kube-scheduler配置文件

[root@k8s-master cfg]# cat /cloud/k8s/kubernetes/cfg/kube-scheduler 
KUBE_SCHEDULER_OPTS="--logtostderr=true --v=4 --master=127.0.0.1:8080 --leader-elect"

创建kube-scheduler启动文件

[root@k8s-master cfg]# cat /usr/lib/systemd/system/kube-scheduler.service 
[Unit]
Description=Kubernetes Scheduler
Documentation=https://github.com/kubernetes/kubernetes

[Service]
EnvironmentFile=-/cloud/k8s/kubernetes/cfg/kube-scheduler
ExecStart=/cloud/k8s/kubernetes/bin/kube-scheduler $KUBE_SCHEDULER_OPTS
Restart=on-failure

[Install]
WantedBy=multi-user.target

启动kube-scheduler服务

[root@k8s-master cfg]# systemctl daemon-reload 
[root@k8s-master cfg]# systemctl enable kube-scheduler.service  
[root@k8s-master cfg]# systemctl start kube-scheduler.service
[root@k8s-master cfg]# ps -ef |grep kube-scheduler
root       1716      1  0 16:12 ?        00:00:19 /cloud/k8s/kubernetes/bin/kube-scheduler --logtostderr=true --v=4 --master=127.0.0.1:8080 --leader-elect
root       1897   1083  0 18:21 pts/0    00:00:00 grep --color=auto kube-scheduler

部署kube-controller-manager组件

创建kube-controller-manager配置文件

[root@k8s-master cfg]# cat /cloud/k8s/kubernetes/cfg/kube-controller-manager 
KUBE_CONTROLLER_MANAGER_OPTS="--logtostderr=true \
--v=4 \
--master=127.0.0.1:8080 \
--leader-elect=true \
--address=127.0.0.1 \
--service-cluster-ip-range=10.0.0.0/24 \
--cluster-name=kubernetes \
--cluster-signing-cert-file=/cloud/k8s/kubernetes/ssl/ca.pem \
--cluster-signing-key-file=/cloud/k8s/kubernetes/ssl/ca-key.pem  \
--root-ca-file=/cloud/k8s/kubernetes/ssl/ca.pem \
--service-account-private-key-file=/cloud/k8s/kubernetes/ssl/ca-key.pem"

创建kube-controller-manager启动文件

[root@k8s-master cfg]# cat /usr/lib/systemd/system/kube-controller-manager.service 
[Unit]
Description=Kubernetes Controller Manager
Documentation=https://github.com/kubernetes/kubernetes

[Service]
EnvironmentFile=/cloud/k8s/kubernetes/cfg/kube-controller-manager
ExecStart=/cloud/k8s/kubernetes/bin/kube-controller-manager $KUBE_CONTROLLER_MANAGER_OPTS
Restart=on-failure

[Install]
WantedBy=multi-user.target

启动kube-controller-manager服务

[root@k8s-master cfg]# systemctl daemon-reload
[root@k8s-master cfg]# systemctl enable kube-controller-manager
[root@k8s-master cfg]# systemctl start kube-controller-manager
[root@k8s-master cfg]# ps -ef |grep kube-controller-manager
root       1709      1  2 16:12 ?        00:03:11 /cloud/k8s/kubernetes/bin/kube-controller-manager --logtostderr=true --v=4 --master=127.0.0.1:8080 --leader-elect=true --address=127.0.0.1 --service-cluster-ip-range=10.0.0.0/24 --cluster-name=kubernetes --cluster-signing-cert-file=/cloud/k8s/kubernetes/ssl/ca.pem --cluster-signing-key-file=/cloud/k8s/kubernetes/ssl/ca-key.pem --root-ca-file=/cloud/k8s/kubernetes/ssl/ca.pem --service-account-private-key-file=/cloud/k8s/kubernetes/ssl/ca-key.pem
root       1907   1083  0 18:29 pts/0    00:00:00 grep --color=auto kube-controller-manager

查看集群状态

[root@k8s-master cfg]# kubectl get cs
NAME                 STATUS    MESSAGE             ERROR
controller-manager   Healthy   ok                  
scheduler            Healthy   ok                  
etcd-0               Healthy   {"health":"true"}   
etcd-2               Healthy   {"health":"true"}   
etcd-1               Healthy   {"health":"true"}

8、部署node节点组件(所有node节点操作)

解压node节点安装包

[root@k8s-node1 install]# tar xf kubernetes-node-linux-amd64.tar.gz
[root@k8s-node1 install]# cp kubernetes/node/bin/{kubelet,kube-proxy} /cloud/k8s/kubernetes/bin/

创建kubelet bootstrap.kubeconfig 文件

[root@k8s-master kubernetes]# cat environment.sh 
# 创建kubelet bootstrapping kubeconfig
BOOTSTRAP_TOKEN=a081e7ba91d597006cbdacfa8ee114ac
KUBE_APISERVER="https://192.168.248.65:6443"
# 设置集群参数
kubectl config set-cluster kubernetes \
  --certificate-authority=./ca.pem \
  --embed-certs=true \
  --server=${KUBE_APISERVER} \
  --kubeconfig=bootstrap.kubeconfig

# 设置客户端认证参数
kubectl config set-credentials kubelet-bootstrap \
  --token=${BOOTSTRAP_TOKEN} \
  --kubeconfig=bootstrap.kubeconfig

# 设置上下文参数
kubectl config set-context default \
  --cluster=kubernetes \
  --user=kubelet-bootstrap \
  --kubeconfig=bootstrap.kubeconfig

# 设置默认上下文
kubectl config use-context default --kubeconfig=bootstrap.kubeconfig

# 执行environment.sh生成bootstrap.kubeconfig[object Object]

创建 kubelet.kubeconfig 文件

[root@k8s-master kubernetes]# cat envkubelet.kubeconfig.sh
# 创建kubelet bootstrapping kubeconfig
BOOTSTRAP_TOKEN=a081e7ba91d597006cbdacfa8ee114ac
KUBE_APISERVER="https://192.168.248.65:6443"

# 设置集群参数
kubectl config set-cluster kubernetes \
  --certificate-authority=./ca.pem \
  --embed-certs=true \
  --server=${KUBE_APISERVER} \
  --kubeconfig=kubelet.kubeconfig

# 设置客户端认证参数
kubectl config set-credentials kubelet \
  --token=${BOOTSTRAP_TOKEN} \
  --kubeconfig=kubelet.kubeconfig

# 设置上下文参数
kubectl config set-context default \
  --cluster=kubernetes \
  --user=kubelet \
  --kubeconfig=kubelet.kubeconfig

# 设置默认上下文
kubectl config use-context default --kubeconfig=kubelet.kubeconfig

#执行envkubelet.kubeconfig.sh脚本，生成kubelet.kubeconfig[object Object]

创建kube-proxy.kubeconfig文件

[root@k8s-master kubernetes]# cat env_proxy.sh
# 创建kube-proxy kubeconfig文件

BOOTSTRAP_TOKEN=a081e7ba91d597006cbdacfa8ee114ac
KUBE_APISERVER="https://192.168.248.65:6443"

kubectl config set-cluster kubernetes \
  --certificate-authority=./ca.pem \
  --embed-certs=true \
  --server=${KUBE_APISERVER} \
  --kubeconfig=kube-proxy.kubeconfig

kubectl config set-credentials kube-proxy \
  --client-certificate=./kube-proxy.pem \
  --client-key=./kube-proxy-key.pem \
  --embed-certs=true \
  --kubeconfig=kube-proxy.kubeconfig

kubectl config set-context default \
  --cluster=kubernetes \
  --user=kube-proxy \
  --kubeconfig=kube-proxy.kubeconfig

kubectl config use-context default --kubeconfig=kube-proxy.kubeconfig

#执行env_proxy.sh脚本生成kube-proxy.kubeconfig文件

将以上生成的kubeconfig复制到所有node节点

[root@k8s-master kubernetes]# scp bootstrap.kubeconfig kubelet.kubeconfig kube-proxy.kubeconfig k8s-node1:/cloud/k8s/kubernetes/cfg/
[root@k8s-master kubernetes]# scp bootstrap.kubeconfig kubelet.kubeconfig kube-proxy.kubeconfig k8s-node2:/cloud/k8s/kubernetes/cfg/
[root@k8s-master kubernetes]# scp bootstrap.kubeconfig kubelet.kubeconfig kube-proxy.kubeconfig k8s-node3:/cloud/k8s/kubernetes/cfg/

所有node节点创建kubelet 参数配置模板文件

[root@k8s-node1 cfg]# cat kubelet.config 
kind: KubeletConfiguration
apiVersion: kubelet.config.k8s.io/v1beta1
address: 192.168.248.66
port: 10250
readOnlyPort: 10255
cgroupDriver: cgroupfs
clusterDNS: ["10.0.0.2"]
clusterDomain: cluster.local.
failSwapOn: false
authentication:
  anonymous:
    enabled: true

[root@k8s-node2 cfg]# cat kubelet.config 
kind: KubeletConfiguration
apiVersion: kubelet.config.k8s.io/v1beta1
address: 192.168.248.67
port: 10250
readOnlyPort: 10255
cgroupDriver: cgroupfs
clusterDNS: ["10.0.0.2"]
clusterDomain: cluster.local.
failSwapOn: false
authentication:
  anonymous:
    enabled: true
    
[root@k8s-node3 cfg]# cat kubelet.config 
kind: KubeletConfiguration
apiVersion: kubelet.config.k8s.io/v1beta1
address: 192.168.248.68
port: 10250
readOnlyPort: 10255
cgroupDriver: cgroupfs
clusterDNS: ["10.0.0.2"]
clusterDomain: cluster.local.
failSwapOn: false
authentication:
  anonymous:
    enabled: true

创建kubelet配置文件

[root@k8s-node1 cfg]# cat /cloud/k8s/kubernetes/cfg/kubelet
KUBELET_OPTS="--logtostderr=true \
--v=4 \
--hostname-override=k8s-node1 \
--kubeconfig=/cloud/k8s/kubernetes/cfg/kubelet.kubeconfig \
--bootstrap-kubeconfig=/cloud/k8s/kubernetes/cfg/bootstrap.kubeconfig \
--config=/cloud/k8s/kubernetes/cfg/kubelet.config \
--cert-dir=/cloud/k8s/kubernetes/ssl \
--pod-infra-container-image=registry.cn-hangzhou.aliyuncs.com/google-containers/pause-amd64:3.0"

[root@k8s-node2 cfg]# cat /cloud/k8s/kubernetes/cfg/kubelet
KUBELET_OPTS="--logtostderr=true \
--v=4 \
--hostname-override=k8s-node2 \
--kubeconfig=/cloud/k8s/kubernetes/cfg/kubelet.kubeconfig \
--bootstrap-kubeconfig=/cloud/k8s/kubernetes/cfg/bootstrap.kubeconfig \
--config=/cloud/k8s/kubernetes/cfg/kubelet.config \
--cert-dir=/cloud/k8s/kubernetes/ssl \
--pod-infra-container-image=registry.cn-hangzhou.aliyuncs.com/google-containers/pause-amd64:3.0"

[root@k8s-node3 cfg]# cat /cloud/k8s/kubernetes/cfg/kubelet
KUBELET_OPTS="--logtostderr=true \
--v=4 \
--hostname-override=k8s-node3 \
--kubeconfig=/cloud/k8s/kubernetes/cfg/kubelet.kubeconfig \
--bootstrap-kubeconfig=/cloud/k8s/kubernetes/cfg/bootstrap.kubeconfig \
--config=/cloud/k8s/kubernetes/cfg/kubelet.config \
--cert-dir=/cloud/k8s/kubernetes/ssl \
--pod-infra-container-image=registry.cn-hangzhou.aliyuncs.com/google-containers/pause-amd64:3.0"

创建kubelet启动文件

[root@k8s-node1 cfg]# cat /usr/lib/systemd/system/kubelet.service 
[Unit]
Description=Kubernetes Kubelet
After=docker.service
Requires=docker.service

[Service]
EnvironmentFile=/cloud/k8s/kubernetes/cfg/kubelet
ExecStart=/cloud/k8s/kubernetes/bin/kubelet $KUBELET_OPTS
Restart=on-failure
KillMode=process

[Install]
WantedBy=multi-user.target

将kubelet-bootstrap用户绑定到系统集群角色(不绑定角色，kubelet将无法启动成功)

kubectl create clusterrolebinding kubelet-bootstrap \
  --clusterrole=system:node-bootstrapper \
  --user=kubelet-bootstrap

启动kubelet服务(所有node节点)

[root@k8s-node1 cfg]# systemctl daemon-reload
[root@k8s-node1 cfg]# systemctl enable kubelet
[root@k8s-node1 cfg]# systemctl start kubelet
[root@k8s-node1 cfg]# ps -ef |grep kubelet
root       3306      1  2 09:02 ?        00:14:47 /cloud/k8s/kubernetes/bin/kubelet --logtostderr=true --v=4 --hostname-override=k8s-node1 --kubeconfig=/cloud/k8s/kubernetes/cfg/kubelet.kubeconfig --bootstrap-kubeconfig=/cloud/k8s/kubernetes/cfg/bootstrap.kubeconfig --config=/cloud/k8s/kubernetes/cfg/kubelet.config --cert-dir=/cloud/k8s/kubernetes/ssl --pod-infra-container-image=registry.cn-hangzhou.aliyuncs.com/google-containers/pause-amd64:3.0
root      87181  12020  0 19:22 pts/0    00:00:00 grep --color=auto kubelet

在master节点上approve kubelet CSR 请求

kubectl get csr
kubectl certificate approve $NAME
csr 状态变为 Approved,Issued 即可

查看集群状态及node节点

[root@k8s-master kubernetes]# kubectl  get cs,node
NAME                                 STATUS    MESSAGE             ERROR
componentstatus/controller-manager   Healthy   ok                  
componentstatus/scheduler            Healthy   ok                  
componentstatus/etcd-2               Healthy   {"health":"true"}   
componentstatus/etcd-0               Healthy   {"health":"true"}   
componentstatus/etcd-1               Healthy   {"health":"true"}   

NAME             STATUS   ROLES    AGE    VERSION
node/k8s-node1   Ready       4d2h   v1.15.0
node/k8s-node2   Ready       4d2h   v1.15.0
node/k8s-node3   Ready       4d2h   v1.15.0

部署 node kube-proxy 组件

[root@k8s-node1 cfg]# cat /cloud/k8s/kubernetes/cfg/kube-proxy
KUBE_PROXY_OPTS="--logtostderr=true \
--v=4 \
--hostname-override=k8s-node1 \
--cluster-cidr=10.0.0.0/24 \
--kubeconfig=/cloud/k8s/kubernetes/cfg/kube-proxy.kubeconfig"

创建kube-proxy启动文件

[root@k8s-node1 cfg]# cat /usr/lib/systemd/system/kube-proxy.service 
[Unit]
Description=Kubernetes Proxy
After=network.target

[Service]
EnvironmentFile=/cloud/k8s/kubernetes/cfg/kube-proxy
ExecStart=/cloud/k8s/kubernetes/bin/kube-proxy $KUBE_PROXY_OPTS
Restart=on-failure

[Install]
WantedBy=multi-user.target

启动kube-proxy服务

[root@k8s-node1 cfg]# systemctl daemon-reload
[root@k8s-node1 cfg]# systemctl enable kube-proxy
[root@k8s-node1 cfg]# systemctl start kube-proxy
[root@k8s-node1 cfg]# ps -ef |grep kube-proxy
root        966      1  0 09:02 ?        00:01:20 /cloud/k8s/kubernetes/bin/kube-proxy --logtostderr=true --v=4 --hostname-override=k8s-node1 --cluster-cidr=10.0.0.0/24 --kubeconfig=/cloud/k8s/kubernetes/cfg/kube-proxy.kubeconfig
root      87093  12020  0 19:22 pts/0    00:00:00 grep --color=auto kube-proxy

部署Coredns组件

[root@k8s-master ~]# cat coredns.yaml 
# Warning: This is a file generated from the base underscore template file: coredns.yaml.base

apiVersion: v1
kind: ServiceAccount
metadata:
  name: coredns
  namespace: kube-system
  labels:
      kubernetes.io/cluster-service: "true"
      addonmanager.kubernetes.io/mode: Reconcile
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
  labels:
    kubernetes.io/bootstrapping: rbac-defaults
    addonmanager.kubernetes.io/mode: Reconcile
  name: system:coredns
rules:
- apiGroups:
  - ""
  resources:
  - endpoints
  - services
  - pods
  - namespaces
  verbs:
  - list
  - watch
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
  annotations:
    rbac.authorization.kubernetes.io/autoupdate: "true"
  labels:
    kubernetes.io/bootstrapping: rbac-defaults
    addonmanager.kubernetes.io/mode: EnsureExists
  name: system:coredns
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: ClusterRole
  name: system:coredns
subjects:
- kind: ServiceAccount
  name: coredns
  namespace: kube-system
---
apiVersion: v1
kind: ConfigMap
metadata:
  name: coredns
  namespace: kube-system
  labels:
      addonmanager.kubernetes.io/mode: EnsureExists
data:
  Corefile: |
    .:53 {
        errors
        health
        kubernetes cluster.local in-addr.arpa ip6.arpa {
            pods insecure
            upstream
            fallthrough in-addr.arpa ip6.arpa
        }
        prometheus :9153
        proxy . /etc/resolv.conf
        cache 30
        loop
        reload
        loadbalance
    }
---
apiVersion: extensions/v1beta1
kind: Deployment
metadata:
  name: coredns
  namespace: kube-system
  labels:
    k8s-app: kube-dns
    kubernetes.io/cluster-service: "true"
    addonmanager.kubernetes.io/mode: Reconcile
    kubernetes.io/name: "CoreDNS"
spec:
  # replicas: not specified here:
  # 1. In order to make Addon Manager do not reconcile this replicas parameter.
  # 2. Default is 1.
  # 3. Will be tuned in real time if DNS horizontal auto-scaling is turned on.
  replicas: 3
  strategy:
    type: RollingUpdate
    rollingUpdate:
      maxUnavailable: 1
  selector:
    matchLabels:
      k8s-app: kube-dns
  template:
    metadata:
      labels:
        k8s-app: kube-dns
      annotations:
        seccomp.security.alpha.kubernetes.io/pod: 'docker/default'
    spec:
      serviceAccountName: coredns
      tolerations:
        - key: node-role.kubernetes.io/master
          effect: NoSchedule
        - key: "CriticalAddonsOnly"
          operator: "Exists"
      containers:
      - name: coredns
        image: coredns/coredns:1.3.1
        imagePullPolicy: IfNotPresent
        resources:
          limits:
            memory: 170Mi
          requests:
            cpu: 100m
            memory: 70Mi
        args: [ "-conf", "/etc/coredns/Corefile" ]
        volumeMounts:
        - name: config-volume
          mountPath: /etc/coredns
          readOnly: true
        ports:
        - containerPort: 53
          name: dns
          protocol: UDP
        - containerPort: 53
          name: dns-tcp
          protocol: TCP
        - containerPort: 9153
          name: metrics
          protocol: TCP
        livenessProbe:
          httpGet:
            path: /health
            port: 8080
            scheme: HTTP
          initialDelaySeconds: 60
          timeoutSeconds: 5
          successThreshold: 1
          failureThreshold: 5
        securityContext:
          allowPrivilegeEscalation: false
          capabilities:
            add:
            - NET_BIND_SERVICE
            drop:
            - all
          readOnlyRootFilesystem: true
      dnsPolicy: Default
      volumes:
        - name: config-volume
          configMap:
            name: coredns
            items:
            - key: Corefile
              path: Corefile
---
apiVersion: v1
kind: Service
metadata:
  name: kube-dns
  namespace: kube-system
  annotations:
    prometheus.io/port: "9153"
    prometheus.io/scrape: "true"
  labels:
    k8s-app: kube-dns
    kubernetes.io/cluster-service: "true"
    addonmanager.kubernetes.io/mode: Reconcile
    kubernetes.io/name: "CoreDNS"
spec:
  selector:
    k8s-app: kube-dns
  clusterIP: 10.0.0.2 
  ports:
  - name: dns
    port: 53
    protocol: UDP
  - name: dns-tcp
    port: 53
    protocol: TCP

[root@k8s-master ~]# kubectl apply -f coredns.yaml 
serviceaccount/coredns unchanged
clusterrole.rbac.authorization.k8s.io/system:coredns unchanged
clusterrolebinding.rbac.authorization.k8s.io/system:coredns unchanged
configmap/coredns unchanged
deployment.extensions/coredns unchanged
service/kube-dns unchanged

[root@k8s-master ~]# kubectl get deployment -n kube-system   
NAME      READY   UP-TO-DATE   AVAILABLE   AGE
coredns   3/3     3            3           33h
[root@k8s-master ~]# kubectl get deployment -n kube-system -o wide
NAME      READY   UP-TO-DATE   AVAILABLE   AGE   CONTAINERS   IMAGES                  SELECTOR
coredns   3/3     3            3           33h   coredns      coredns/coredns:1.3.1   k8s-app=kube-dns
[root@k8s-master ~]# kubectl get pod -n kube-system -o wide          
NAME                      READY   STATUS    RESTARTS   AGE   IP            NODE        NOMINATED NODE   READINESS GATES
coredns-b49c586cf-nwzv6   1/1     Running   1          33h   172.18.54.3   k8s-node3              
coredns-b49c586cf-qv5b9   1/1     Running   1          33h   172.18.22.3   k8s-node1              
coredns-b49c586cf-rcqhc   1/1     Running   1          33h   172.18.95.2   k8s-node2              
[root@k8s-master ~]# kubectl get svc -n kube-system -o wide   
NAME       TYPE        CLUSTER-IP   EXTERNAL-IP   PORT(S)         AGE   SELECTOR
kube-dns   ClusterIP   10.0.0.2             53/UDP,53/TCP   33h   k8s-app=kube-dns

到此kubernetes V1.15.0乞丐版部署完成。

分享标题：二进制部署Kubernetes集群参考文档（V1.15.0）
当前地址：http://cdxtjz.cn/article/jsdddi.html