第一步：首先搭建etcd集群

yum install -y etcd

配置文件 /etc/etcd/etcd.conf
ETCD_NAME=infra1

ETCD_DATA_DIR="/var/lib/etcd"

ETCD_LISTEN_PEER_URLS="https://172.20.0.113:2380"

ETCD_LISTEN_CLIENT_URLS="https://172.20.0.113:2379"

#[cluster]

ETCD_INITIAL_ADVERTISE_PEER_URLS="https://172.20.0.113:2380"

ETCD_INITIAL_CLUSTER_TOKEN="etcd-cluster"

ETCD_ADVERTISE_CLIENT_URLS="https://172.20.0.113:2379,http://127.0.0.1:2379"
#配置集群IP

ETCD_INITIAL_CLUSTER="infra1=http://172.20.0.113:2380,infra2=http://172.20.0.114:2380,infra3=http://172.20.0.115:2380"
第二步：安装keepalived虚拟IP代理
yum install -y keepalived

# 添加以下内容

net.ipv4.ip_forward = 1

net.ipv4.ip_nonlocal_bind = 1

# 验证并生效

$ sysctl -p

# 验证是否生效

$ cat /proc/sys/net/ipv4/ip_forward


配置文件 /etc/keepalived/keepalived.conf


! Configuration File for keepalived

global_defs {

   notification_email {

   }

   router_id <cluster-difference-name>

}

vrrp_script check_haproxy {

    # 自身状态检测

    script "killall -0 haproxy"

    interval 3

    weight 5

}

vrrp_instance haproxy-vip {

    # 使用单播通信，默认是组播通信

    unicast_src_ip 192.168.1.137

    unicast_peer {

        192.168.1.138

    }

    # 初始化状态

    state MASTER(BACKUP)

    # 虚拟ip 绑定的网卡 （这里根据你自己的实际情况选择网卡）

    interface eth0

    # 此ID 要与Backup 配置一致

    virtual_router_id 51<cluster id same>

    # 默认启动优先级，要比Backup 大点，但要控制量，保证自身状态检测生效

    priority 100

    advert_int 1

    authentication {

        auth_type PASS

        auth_pass 1111

    }

    virtual_ipaddress {

        # 虚拟ip 地址

        192.168.1.139

    }

    track_script {

        check_k8s

    }

}

virtual_server 192.168.1.139 80 {

  delay_loop 5

  lvs_sched wlc

  lvs_method NAT

  persistence_timeout 1800

  protocol TCP

  real_server 192.168.1.137 80 {

    weight 1

    TCP_CHECK {

      connect_port 80

      connect_timeout 3

    }

  }

}

virtual_server 192.168.1.139 443 {

  delay_loop 5

  lvs_sched wlc

  lvs_method NAT

  persistence_timeout 1800

  protocol TCP

  real_server 192.168.1.137 443 {

    weight 1

    TCP_CHECK {

      connect_port 443

      connect_timeout 3

    }

  }

}
第三步：安装docker、准备必要镜像
关闭防火墙
systemctl stop firewalld
关闭SELinux
setenforce 0
sed -i -e 's/SELINUX=enforcing/SELINUX=disable/g' /etc/sysconfig/selinux
安装docker
yum install -y docker
#镜像包
etcd-amd64_v3.1.11
flannel:v0.9.1-amd64_1.14.7
k8s-dns-dnsmasq-nanny-amd64_1.14.7
k8s-dns-sidecar-amd64_1.14.7
kube-apiserver-amd-v1.9.2
kube-controller-manager-amd64-v1.9.2
kube-proxy-amd64-v1.9.2
kube-scheduler-amd64-v1.9.2
pause-amd64_3.0

第四步：配置kubeadm-config.yaml


apiVersion: kubeadm.k8s.io/v1alpha1

kind: MasterConfiguration

api:

  advertiseAddress: 192.168.4.24

etcd:

  endpoints:

  - http://192.168.4.24:2379

  - http://192.168.4.25:2379

  - http://192.168.4.26:2379

imageRepository: k8s.gcr.io #配置私有仓库

networking: podSubnet: 10.1.0.0/16 #和flanneld的网段一致

   endpoint-reconciler-type: lease EOL ##初始化kubernetes集群
kubeadm init --config kubeadm-config.yaml

注意:如果flanneld使用服务安装则需要添加

etcdctl --endpoints=https://172.20.0.113:2379,https://172.20.0.114:2379,https://172.20.0.115:2379 \

  --ca-file=/etc/kubernetes/ssl/ca.pem \

  --cert-file=/etc/kubernetes/ssl/kubernetes.pem \

  --key-file=/etc/kubernetes/ssl/kubernetes-key.pem \

  mkdir /kube-centos/network

etcdctl --endpoints=https://172.20.0.113:2379,https://172.20.0.114:2379,https://172.20.0.115:2379 \

  --ca-file=/etc/kubernetes/ssl/ca.pem \

  --cert-file=/etc/kubernetes/ssl/kubernetes.pem \

  --key-file=/etc/kubernetes/ssl/kubernetes-key.pem \

  mk /kube-centos/network/config '{"Network":"172.30.0.0/16","SubnetLen":24,"Backend":{"Type":"vxlan"}}'
 错误NetworkPlugin cni failed to set up pod
  停掉集群删除flannel，避免网络污染

  rm -rf /var/lib/cni/flannel/* && rm -rf /var/lib/cni/networks/cbr0/* && ip link delete cni0