附022.Kubernetes_v1.18.3高可用部署架构一

2020 年 6 月 12 日
笔记
dashboard, helm, Ingress, keepalived, kubeadm, Kubernetes, Longhorn, metrics, Nginx高可用, 集群相关, 高可用

kubeadm介绍

kubeadm概述

参考附003.Kubeadm部署Kubernetes。

kubeadm功能

参考附003.Kubeadm部署Kubernetes。

本方案描述

本方案采用kubeadm部署Kubernetes 1.18.3版本；
etcd采用混部方式；
Keepalived：实现VIP高可用；
Nginx：以Pod形式运行与Kubernetes之上，即in Kubernetes模式，提供反向代理至3个master 6443端口；
其他主要部署组件包括：
- Metrics：度量；
- Dashboard：Kubernetes 图形UI界面；
- Helm：Kubernetes Helm包管理工具；
- Ingress：Kubernetes 服务暴露；
- Longhorn：Kubernetes 动态存储组件。

部署规划

节点规划

节点主机名	IP	类型	运行服务
master01	172.24.8.71	Kubernetes master节点	docker、etcd、kube-apiserver、kube-scheduler、kube-controller-manager、kubectl、kubelet、metrics、calico
master02	172.24.8.72	Kubernetes master节点	docker、etcd、kube-apiserver、kube-scheduler、kube-controller-manager、kubectl、kubelet、metrics、calico
master03	172.24.8.73	Kubernetes master节点	docker、etcd、kube-apiserver、kube-scheduler、kube-controller-manager、kubectl、kubelet、metrics、calico
worker01	172.24.8.74	Kubernetes worker节点	docker、kubelet、proxy、calico
worker02	172.24.8.75	Kubernetes worker节点	docker、kubelet、proxy、calico
worker03	172.24.8.76	Kubernetes worker节点	docker、kubelet、proxy、calico

Kubernetes的高可用主要指的是控制平面的高可用，即指多套Master节点组件和Etcd组件，工作节点通过负载均衡连接到各Master。

Kubernetes高可用架构中etcd与Master节点组件混布方式特点：

Etcd混布方式
所需机器资源少
部署简单，利于管理
容易进行横向扩展
风险大，一台宿主机挂了，master和etcd就都少了一套，集群冗余度受到的影响比较大。

初始准备

[root@master01 ~]# hostnamectl set-hostname master01 #其他节点依次修改

[root@master01 ~]# cat >> /etc/hosts << EOF
172.24.8.71 master01··
172.24.8.72 master02
172.24.8.73 master03
172.24.8.74 worker01
172.24.8.75 worker02
172.24.8.76 worker03
EOF

[root@master01 ~]# vi k8sinit.sh

#!/bin/sh
#****************************************************************#
# ScriptName: k8sinit.sh
# Author: xhy
# Create Date: 2020-05-30 16:30
# Modify Author: xhy
# Modify Date: 2020-05-30 16:30
# Version: 
#***************************************************************#
# Initialize the machine. This needs to be executed on every machine.

# Add docker user
useradd -m docker

# Disable the SELinux.
sed -i 's/^SELINUX=.*/SELINUX=disabled/' /etc/selinux/config

# Turn off and disable the firewalld.
systemctl stop firewalld
systemctl disable firewalld

# Modify related kernel parameters & Disable the swap.
cat > /etc/sysctl.d/k8s.conf << EOF
net.ipv4.ip_forward = 1
net.bridge.bridge-nf-call-ip6tables = 1
net.bridge.bridge-nf-call-iptables = 1
net.ipv4.tcp_tw_recycle = 0
vm.swappiness = 0
vm.overcommit_memory = 1
vm.panic_on_oom = 0
net.ipv6.conf.all.disable_ipv6 = 1
EOF
sysctl -p /etc/sysctl.d/k8s.conf >&/dev/null
swapoff -a
sed -i '/ swap / s/^\(.*\)$/#\1/g' /etc/fstab
modprobe br_netfilter

# Add ipvs modules
cat > /etc/sysconfig/modules/ipvs.modules <<EOF
#!/bin/bash
modprobe -- ip_vs
modprobe -- ip_vs_rr
modprobe -- ip_vs_wrr
modprobe -- ip_vs_sh
modprobe -- nf_conntrack_ipv4
modprobe -- nf_conntrack
EOF

chmod 755 /etc/sysconfig/modules/ipvs.modules
bash /etc/sysconfig/modules/ipvs.modules

# Install rpm
yum install -y conntrack ntpdate ntp ipvsadm ipset jq iptables curl sysstat libseccomp wget

# Install Docker Compose
sudo curl -L "//down.linuxsb.com:8888/docker/compose/releases/download/1.25.4/docker-compose-$(uname -s)-$(uname -m)" -o /usr/local/bin/docker-compose
sudo chmod +x /usr/local/bin/docker-compose

# Update kernel
rpm --import //down.linuxsb.com:8888/RPM-GPG-KEY-elrepo.org
rpm -Uvh //down.linuxsb.com:8888/elrepo-release-7.0-4.el7.elrepo.noarch.rpm
yum --disablerepo="*" --enablerepo="elrepo-kernel" install -y kernel-ml
sed -i 's/^GRUB_DEFAULT=.*/GRUB_DEFAULT=0/' /etc/default/grub
grub2-mkconfig -o /boot/grub2/grub.cfg
yum update -y

# Reboot the machine.
# reboot

提示：对于某些特性，可能需要升级内核，内核升级操作见《018.Linux升级内核》。4.19版及以上内核nf_conntrack_ipv4已经改为nf_conntrack。

互信配置

为了更方便远程分发文件和执行命令，本实验配置master节点到其它节点的 ssh 信任关系。

[root@master01 ~]# ssh-keygen -f ~/.ssh/id_rsa -N ''
[root@master01 ~]# ssh-copy-id -i ~/.ssh/id_rsa.pub root@master01
[root@master01 ~]# ssh-copy-id -i ~/.ssh/id_rsa.pub root@master02
[root@master01 ~]# ssh-copy-id -i ~/.ssh/id_rsa.pub root@master03
[root@master01 ~]# ssh-copy-id -i ~/.ssh/id_rsa.pub root@worker01
[root@master01 ~]# ssh-copy-id -i ~/.ssh/id_rsa.pub root@worker02
[root@master01 ~]# ssh-copy-id -i ~/.ssh/id_rsa.pub root@worker03

提示：此操作仅需要在master节点操作。

其他准备

[root@master01 ~]# vi environment.sh

#!/bin/sh
#****************************************************************#
# ScriptName: environment.sh
# Author: xhy
# Create Date: 2020-05-30 16:30
# Modify Author: xhy
# Modify Date: 2020-05-30 16:30
# Version: 
#***************************************************************#
# 集群 MASTER 机器 IP 数组
export MASTER_IPS=(172.24.8.71 172.24.8.72 172.24.8.73)

# 集群 MASTER IP 对应的主机名数组
export MASTER_NAMES=(master01 master02 master03)

# 集群 NODE 机器 IP 数组
export NODE_IPS=(172.24.8.74 172.24.8.75 172.24.8.76)

# 集群 NODE IP 对应的主机名数组
export NODE_NAMES=(worker01 worker02 worker03)

# 集群所有机器 IP 数组
export ALL_IPS=(172.24.8.71 172.24.8.72 172.24.8.73 172.24.8.74 172.24.8.75 172.24.8.76)

# 集群所有IP 对应的主机名数组
export ALL_NAMES=(master01 master02 master03 worker01 worker02 worker03)

[root@master01 ~]# source environment.sh
[root@master01 ~]# chmod +x *.sh
[root@master01 ~]# for all_ip in ${ALL_IPS[@]}
  do
    echo ">>> ${all_ip}"
    scp -rp /etc/hosts root@${all_ip}:/etc/hosts
    scp -rp k8sinit.sh root@${all_ip}:/root/
    ssh root@${all_ip} "bash /root/k8sinit.sh"
  done

集群部署

Docker安装

[root@master01 ~]# for all_ip in ${ALL_IPS[@]}
  do
    echo ">>> ${all_ip}"
    ssh root@${all_ip} "yum -y install yum-utils device-mapper-persistent-data lvm2"
    ssh root@${all_ip} "yum-config-manager --add-repo //mirrors.aliyun.com/docker-ce/linux/centos/docker-ce.repo"
    ssh root@${all_ip} "yum -y install docker-ce"
    ssh root@${all_ip} "mkdir /etc/docker"
    ssh root@${all_ip} "cat > /etc/docker/daemon.json <<EOF
{
  \"registry-mirrors\": [\"//dbzucv6w.mirror.aliyuncs.com\"],
  \"exec-opts\": [\"native.cgroupdriver=systemd\"],
  \"log-driver\": \"json-file\",
  \"log-opts\": {
    \"max-size\": \"100m\"
  },
  \"storage-driver\": \"overlay2\",
  \"storage-opts\": [
    \"overlay2.override_kernel_check=true\"
  ]
}
EOF"
    ssh root@${all_ip} "systemctl restart docker"
    ssh root@${all_ip} "systemctl enable docker"
    ssh root@${all_ip} "systemctl status docker"
    ssh root@${all_ip} "iptables -nvL"
  done

提示：如上仅需Master01节点操作，从而实现所有节点自动化安装。

正式安装

[root@master01 ~]# for all_ip in ${ALL_IPS[@]}
  do
    echo ">>> ${all_ip}"
    ssh root@${all_ip} "cat <<EOF > /etc/yum.repos.d/kubernetes.repo
[kubernetes]
name=Kubernetes
baseurl=//mirrors.aliyun.com/kubernetes/yum/repos/kubernetes-el7-x86_64/
enabled=1
gpgcheck=1
repo_gpgcheck=1
gpgkey=//mirrors.aliyun.com/kubernetes/yum/doc/yum-key.gpg //mirrors.aliyun.com/kubernetes/yum/doc/rpm-package-key.gpg
EOF"
    ssh root@${all_ip} "yum install -y kubeadm-1.18.3-0.x86_64 kubelet-1.18.3-0.x86_64 kubectl-1.18.3-0.x86_64 --disableexcludes=kubernetes"
    ssh root@${all_ip} "systemctl enable kubelet"
done

[root@master01 ~]# yum search -y kubelet --showduplicates #查看相应版本

提示：如上仅需Master01节点操作，从而实现所有节点自动化安装，同时此时不需要启动kubelet，初始化的过程中会自动启动的，如果此时启动了会出现报错，忽略即可。

说明：同时安装了cri-tools, kubernetes-cni, socat三个依赖：
socat：kubelet的依赖；
cri-tools：即CRI(Container Runtime Interface)容器运行时接口的命令行工具。

部署高可用组件I

Keepalived安装

[root@master01 ~]# for master_ip in ${MASTER_IPS[@]}
  do
    echo ">>> ${master_ip}"
    ssh root@${master_ip} "yum -y install gcc gcc-c++ make libnl libnl-devel libnfnetlink-devel openssl-devel"
    ssh root@${master_ip} "wget //down.linuxsb.com:8888/software/keepalived-2.0.20.tar.gz"
    ssh root@${master_ip} "tar -zxvf keepalived-2.0.20.tar.gz"
    ssh root@${master_ip} "cd keepalived-2.0.20/ && ./configure --sysconf=/etc --prefix=/usr/local/keepalived && make && make install"
    ssh root@${master_ip} "systemctl enable keepalived && systemctl start keepalived"
  done

提示：如上仅需Master01节点操作，从而实现所有节点自动化安装。

创建配置文件

[root@master01 ~]# wget //down.linuxsb.com:8888/ngkek8s.sh		#拉取自动部署脚本
[root@master01 ~]# chmod u+x ngkek8s.sh

[root@master01 ~]# vi ngkek8s.sh
#!/bin/sh
#****************************************************************#
# ScriptName: k8s_ha.sh
# Author: xhy
# Create Date: 2020-05-13 16:32
# Modify Author: xhy
# Modify Date: 2020-06-12 12:53
# Version: v2
#***************************************************************#

#######################################
# set variables below to create the config files, all files will create at ./config directory
#######################################

# master keepalived virtual ip address
export K8SHA_VIP=172.24.8.100

# master01 ip address
export K8SHA_IP1=172.24.8.71

# master02 ip address
export K8SHA_IP2=172.24.8.72

# master03 ip address
export K8SHA_IP3=172.24.8.73

# master01 hostname
export K8SHA_HOST1=master01

# master02 hostname
export K8SHA_HOST2=master02

# master03 hostname
export K8SHA_HOST3=master03

# master01 network interface name
export K8SHA_NETINF1=eth0

# master02 network interface name
export K8SHA_NETINF2=eth0

# master03 network interface name
export K8SHA_NETINF3=eth0

# keepalived auth_pass config
export K8SHA_KEEPALIVED_AUTH=412f7dc3bfed32194d1600c483e10ad1d

# kubernetes CIDR pod subnet
export K8SHA_PODCIDR=10.10.0.0

# kubernetes CIDR svc subnet
export K8SHA_SVCCIDR=10.20.0.0

[root@master01 ~]# ./ngkek8s.sh

解释：如上仅需Master01节点操作。执行ngkek8s.sh脚本后，会自动生成以下配置文件：

kubeadm-config.yaml：kubeadm初始化配置文件，位于当前目录
keepalived：keepalived配置文件，位于各个master节点的/etc/keepalived目录
nginx-lb：nginx-lb负载均衡配置文件，位于各个master节点的/etc/kubernetes/nginx-lb/目录
calico.yaml：calico网络组件部署文件，位于config/calico/目录

[root@master01 ~]# cat kubeadm-config.yaml		#检查集群初始化配置
apiVersion: kubeadm.k8s.io/v1beta2
kind: ClusterConfiguration
networking:
  serviceSubnet: "10.20.0.0/16"			     	#设置svc网段
  podSubnet: "10.10.0.0/16"			        #设置Pod网段
  dnsDomain: "cluster.local"
kubernetesVersion: "v1.18.3"			    	#设置安装版本
controlPlaneEndpoint: "172.24.8.100:16443"		#设置相关API VIP地址
apiServer:
  certSANs:
  - master01
  - master02
  - master03
  - 127.0.0.1
  - 172.24.8.71
  - 172.24.8.72
  - 172.24.8.73
  - 172.24.8.100
  timeoutForControlPlane: 4m0s
certificatesDir: "/etc/kubernetes/pki"
imageRepository: "k8s.gcr.io"
---
apiVersion: kubeproxy.config.k8s.io/v1alpha1
kind: KubeProxyConfiguration
featureGates:
  SupportIPVSProxyMode: true
mode: ipvs

提示：如上仅需Master01节点操作，更多config文件参考：//godoc.org/k8s.io/kubernetes/cmd/kubeadm/app/apis/kubeadm/v1beta2。
此kubeadm部署初始化配置更多参考：//pkg.go.dev/k8s.io/kubernetes/cmd/kubeadm/app/apis/kubeadm/v1beta2?tab=doc。

启动Keepalived

[root@master01 ~]# cat /etc/keepalived/check_apiserver.sh	#确认Keepalived配置

[root@master01 ~]# for master_ip in ${MASTER_IPS[@]}
  do
    echo ">>> ${master_ip}"
    ssh root@${master_ip} "systemctl start keepalived.service && systemctl enable keepalived.service"
    ssh root@${master_ip} "systemctl status keepalived.service | grep Active"
  done

[root@master01 ~]# for all_ip in ${ALL_IPS[@]}
  do
    echo ">>> ${all_ip}"
    ssh root@${all_ip} "ping -c1 172.24.8.100"
  done								#等待10s左右执行检查

提示：如上仅需Master01节点操作，从而实现所有节点自动启动服务。

启动Nginx

执行ngkek8s.sh脚本后，nginx-lb的配置文件会自动复制到各个master的节点的/etc/kubernetes/nginx-lb目录。

[root@master01 ~]# for master_ip in ${MASTER_IPS[@]}
  do
    echo ">>> ${master_ip}"
    ssh root@${master_ip} "cd /etc/kubernetes/nginx-lb/ && docker-compose up -d"
    ssh root@${master_ip} "docker-compose ps"
  done

提示：如上仅需Master01节点操作，从而实现所有节点自动启动服务。

初始化集群-Master

拉取镜像

[root@master01 ~]# kubeadm --kubernetes-version=v1.18.3 config images list #列出所需镜像

[root@master01 ~]# cat config/downimage.sh			#确认版本，提前下载镜像
#!/bin/sh
#****************************************************************#
# ScriptName: downimage.sh
# Author: xhy
# Create Date: 2020-05-29 19:55
# Modify Author: xhy
# Modify Date: 2020-05-30 16:07
# Version: v2
#***************************************************************#

KUBE_VERSION=v1.18.3
CALICO_VERSION=v3.14.1
CALICO_URL=calico
KUBE_PAUSE_VERSION=3.2
ETCD_VERSION=3.4.3-0
CORE_DNS_VERSION=1.6.7
GCR_URL=k8s.gcr.io
METRICS_SERVER_VERSION=v0.3.6
INGRESS_VERSION=0.32.0
CSI_PROVISIONER_VERSION=v1.4.0
CSI_NODE_DRIVER_VERSION=v1.2.0
CSI_ATTACHER_VERSION=v2.0.0
CSI_RESIZER_VERSION=v0.3.0 
ALIYUN_URL=registry.cn-hangzhou.aliyuncs.com/google_containers
UCLOUD_URL=uhub.service.ucloud.cn/uxhy
QUAY_URL=quay.io

kubeimages=(kube-proxy:${KUBE_VERSION}
kube-scheduler:${KUBE_VERSION}
kube-controller-manager:${KUBE_VERSION}
kube-apiserver:${KUBE_VERSION}
pause:${KUBE_PAUSE_VERSION}
etcd:${ETCD_VERSION}
coredns:${CORE_DNS_VERSION}
metrics-server-amd64:${METRICS_SERVER_VERSION}
)

for kubeimageName in ${kubeimages[@]} ; do
docker pull $UCLOUD_URL/$kubeimageName
docker tag $UCLOUD_URL/$kubeimageName $GCR_URL/$kubeimageName
docker rmi $UCLOUD_URL/$kubeimageName
done

calimages=(cni:${CALICO_VERSION}
pod2daemon-flexvol:${CALICO_VERSION}
node:${CALICO_VERSION}
kube-controllers:${CALICO_VERSION})

for calimageName in ${calimages[@]} ; do
docker pull $UCLOUD_URL/$calimageName
docker tag $UCLOUD_URL/$calimageName $CALICO_URL/$calimageName
docker rmi $UCLOUD_URL/$calimageName
done

ingressimages=(nginx-ingress-controller:${INGRESS_VERSION})

for ingressimageName in ${ingressimages[@]} ; do
docker pull $UCLOUD_URL/$ingressimageName
docker tag $UCLOUD_URL/$ingressimageName $QUAY_URL/kubernetes-ingress-controller/$ingressimageName
docker rmi $UCLOUD_URL/$ingressimageName
done

csiimages=(csi-provisioner:${CSI_PROVISIONER_VERSION}
csi-node-driver-registrar:${CSI_NODE_DRIVER_VERSION}
csi-attacher:${CSI_ATTACHER_VERSION}
csi-resizer:${CSI_RESIZER_VERSION}
)

for csiimageName in ${csiimages[@]} ; do
docker pull $UCLOUD_URL/$csiimageName
docker tag $UCLOUD_URL/$csiimageName $QUAY_URL/k8scsi/$csiimageName
docker rmi $UCLOUD_URL/$csiimageName
done

[root@master01 ~]# for all_ip in ${ALL_IPS[@]}
  do
    echo ">>> ${all_ip}"
    scp -rp config/downimage.sh root@${all_ip}:/root/
    ssh root@${all_ip} "bash downimage.sh &"
  done

提示：如上仅需Master01节点操作，从而实现所有节点自动拉取镜像。
[root@master01 ~]# docker images #确认验证

Master上初始化

[root@master01 ~]# kubeadm init --config=kubeadm-config.yaml --upload-certs  

保留如下命令用于后续节点添加：
You can now join any number of the control-plane node running the following command on each as root:

  kubeadm join 172.24.8.100:16443 --token xb9wda.v0yf7tlsgo8mdrhk \
    --discovery-token-ca-cert-hash sha256:249884d81a23bd821e38d3345866a99e6d55e443b545825c3c448f30f8e52c3b \
    --control-plane --certificate-key e30428776a47ed2c7e18c9e2951d9e40e068c9ecec5a4858457f1475f1a2a39a

Please note that the certificate-key gives access to cluster sensitive data, keep it secret!
As a safeguard, uploaded-certs will be deleted in two hours; If necessary, you can use
"kubeadm init phase upload-certs --upload-certs" to reload certs afterward.

Then you can join any number of worker nodes by running the following on each as root:

kubeadm join 172.24.8.100:16443 --token xb9wda.v0yf7tlsgo8mdrhk \
    --discovery-token-ca-cert-hash sha256:249884d81a23bd821e38d3345866a99e6d55e443b545825c3c448f30f8e52c3b

注意：如上token具有默认24小时的有效期，token和hash值可通过如下方式获取：
kubeadm token list
如果 Token 过期以后，可以输入以下命令，生成新的 Token:

kubeadm token create
openssl x509 -pubkey -in /etc/kubernetes/pki/ca.crt | openssl rsa -pubin -outform der 2>/dev/null | openssl dgst -sha256 -hex | sed 's/^.* //'***

[root@master01 ~]# mkdir -p $HOME/.kube
[root@master01 ~]# sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
[root@master01 ~]# sudo chown $(id -u):$(id -g) $HOME/.kube/config

[root@master01 ~]# cat << EOF >> ~/.bashrc
export KUBECONFIG=$HOME/.kube/config
EOF							#设置KUBECONFIG环境变量

[root@master01 ~]# echo "source <(kubectl completion bash)" >> ~/.bashrc
[root@master01 ~]# source ~/.bashrc

附加：初始化过程大致步骤如下：

[kubelet-start] 生成kubelet的配置文件”/var/lib/kubelet/config.yaml”
[certificates]生成相关的各种证书
[kubeconfig]生成相关的kubeconfig文件
[bootstraptoken]生成token记录下来，后边使用kubeadm join往集群中添加节点时会用到

提示：初始化仅需要在master01上执行，若初始化异常可通过kubeadm reset && rm -rf $HOME/.kube重置。

添加其他master节点

[root@master02 ~]# kubeadm join 172.24.8.100:16443 --token xb9wda.v0yf7tlsgo8mdrhk \
    --discovery-token-ca-cert-hash sha256:249884d81a23bd821e38d3345866a99e6d55e443b545825c3c448f30f8e52c3b \
    --control-plane --certificate-key e30428776a47ed2c7e18c9e2951d9e40e068c9ecec5a4858457f1475f1a2a39a

[root@master02 ~]# mkdir -p $HOME/.kube
[root@master02 ~]# sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
[root@master02 ~]# sudo chown $(id -u):$(id -g) $HOME/.kube/config
 [root@master02 ~]# cat << EOF >> ~/.bashrc`
export KUBECONFIG=$HOME/.kube/config
EOF						               	#设置KUBECONFIG环境变量
[root@master02 ~]# echo "source <(kubectl completion bash)" >> ~/.bashrc
[root@master02 ~]# source ~/.bashrc

提示：master03也如上执行添加至集群的controlplane。
提示：若添加异常可通过kubeadm reset && rm -rf $HOME/.kube重置。

Tags: dashboard helm Ingress keepalived kubeadm Kubernetes Longhorn metrics Nginx高可用集群相关高可用