基于metrics-server弹性伸缩
目录
一、环境准备
1.1、主机初始化配置
1.2、部署docker环境
二、部署kubernetes集群
2.1、组件介绍
2.2、配置阿里云yum源
2.3、安装kubelet kubeadm kubectl
2.4、配置init-config.yaml
2.5、安装master节点
2.6、安装node节点
2.7、安装flannel、cni
2.8 节点管理命令
三、安装Dashboard UI
3.1、部署Dashboard
3.2、开放端口设置
3.3、权限配置
3.4、访问Token配置
四、metrics-server服务部署
4.1、在Node节点上下载镜像
4.2、修改 Kubernetes apiserver 启动参数
4.3、Master上进行部署
五、弹性伸缩
5.1、弹性伸缩介绍
5.2、弹性伸缩工作原理
5.3、弹性伸缩实战
一、环境准备
| 操作系统 |
IP地址 |
主机名 |
组件 |
| CentOS7.5 |
192.168.147.137 |
k8s-master |
kubeadm、kubelet、kubectl、docker-ce |
| CentOS7.5 |
192.168.147.139 |
k8s-node01 |
kubeadm、kubelet、kubectl、docker-ce |
| CentOS7.5 |
192.168.147.140 |
k8s-node02 |
kubeadm、kubelet、kubectl、docker-ce |
注意：所有主机配置推荐CPU：2C+ Memory：2G+
1.1、主机初始化配置
所有主机配置禁用防火墙和selinux
[root@localhost ~]# setenforce 0
[root@localhost ~]# iptables -F
[root@localhost ~]# systemctl stop firewalld
[root@localhost ~]# systemctl disable firewalld
[root@localhost ~]# systemctl stop NetworkManager
[root@localhost ~]# systemctl disable NetworkManager
[root@localhost ~]# sed -i '/^SELINUX=/s/enforcing/disabled/' /etc/selinux/config
配置主机名并绑定hosts，不同主机名称不同
[root@client2 ~]# hostname k8s-master
[root@client2 ~]# bash
[root@k8s-master ~]# cat << EOF >> /etc/hosts
> 192.168.147.137 k8s-master
> 192.168.147.139 k8s-node01
> 192.168.147.140 k8s-node02
> EOF
[root@k8s-master ~]# scp /etc/hosts 192.168.147.139:/etc/hosts
[root@k8s-master ~]# scp /etc/hosts 192.168.147.140:/etc/hosts
主机配置初始化
[root@k8s-master ~]# yum -y install vim wget net-tools lrzsz
[root@k8s-master ~]# swapoff -a
[root@k8s-master ~]# sed -i '/swap/s/^/#/' /etc/fstab
[root@k8s-node01 ~]# cat << EOF >> /etc/sysctl.conf
net.bridge.bridge-nf-call-ip6tables = 1
net.bridge.bridge-nf-call-iptables = 1
EOF
[root@k8s-node01 ~]# modprobe br_netfilter
[root@k8s-node01 ~]# sysctl -p
1.2、部署docker环境
三台主机上分别部署 Docker 环境，因为 Kubernetes 对容器的编排需要 Docker 的支持
[root@k8s-master ~]# wget -O /etc/yum.repos.d/CentOS-Base.repo http://mirrors.aliyun.com/repo/Centos-7.repo
[root@k8s-master ~]# yum install -y yum-utils device-mapper-persistent-data lvm2
使用 YUM 方式安装 Docker 时，推荐使用阿里的 YUM 源。阿里的官方开源站点地址是：https://developer.aliyun.com/mirror/，可以在站点内找到 Docker 的源地址。
[root@k8s-master ~]# yum-config-manager --add-repo https://mirrors.aliyun.com/docker-ce/linux/centos/docker-ce.repo
[root@k8s-master ~]# yum clean all && yum makecache fast
[root@k8s-master ~]# yum -y install docker-ce
[root@k8s-master ~]# systemctl start docker
[root@k8s-master ~]# systemctl enable docker
镜像加速器（所有主机配置）
很多镜像都是在国外的服务器上，由于网络上存在的问题，经常导致无法拉取镜像的错误，所以最好将镜像拉取地址设置成国内的。目前国内很多公有云服务商都提供了镜像加速服务。镜像加速配置如下所示。
https://dockerhub.azk8s.cn //Azure 中国镜像
https://hub-mirror.c.163.com //网易云加速器
[root@k8s-master ~]# cat << END > /etc/docker/daemon.json
{
"registry-mirrors":[ "https://nyakyfun.mirror.aliyuncs.com" ]
}
END
[root@k8s-master ~]# systemctl daemon-reload
[root@k8s-master ~]# systemctl restart docker
将镜像加速地址直接写入/etc/docker/daemon.json 文件内，如果文件不存在，可直接新建文件并保存。通过该文件扩展名可以看出，daemon.json 的内容必须符合 json 格式，书写时要注意。同时，由于单一镜像服务存在不可用的情况，在配置加速时推荐配置两个或多个加速地址，从而达到冗余、高可用的目的。
二、部署kubernetes集群
2.1、组件介绍
三个节点都需要安装下面三个组件
- kubeadm：安装工具，使所有的组件都会以容器的方式运行
- kubectl：客户端连接K8S API工具
- kubelet：运行在node节点，用来启动容器的工具
2.2、配置阿里云yum源
推荐使用阿里云的yum源安装：
[root@k8s-master ~]# cat < /etc/yum.repos.d/kubernetes.repo
[kubernetes]
name=Kubernetes
baseurl=https://mirrors.aliyun.com/kubernetes/yum/repos/kubernetes-el7-x86_64/
enabled=1
gpgcheck=1
repo_gpgcheck=1
gpgkey=https://mirrors.aliyun.com/kubernetes/yum/doc/yum-key.gpg
https://mirrors.aliyun.com/kubernetes/yum/doc/rpm-package-key.gpg
EOF
[root@k8s-master ~]# ls /etc/yum.repos.d/
backup Centos-7.repo CentOS-Media.repo CentOS-x86_64-kernel.repo docker-ce.repo kubernetes.repo
2.3、安装kubelet kubeadm kubectl
所有主机配置
17、19版本都可以
[root@k8s-master ~]# yum install -y kubelet-1.17.0 kubeadm-1.17.0 kubectl-1.17.0
[root@k8s-master ~]# systemctl enable kubelet
kubelet 刚安装完成后，通过 systemctl start kubelet 方式是无法启动的，需要加入节点或初始化为 master 后才可启动成功。
如果在命令执行过程中出现索引 gpg 检查失败的情况, 请使用 yum install -y --nogpgcheck kubelet kubeadm kubectl 来安装。
2.4、配置init-config.yaml
Kubeadm 提供了很多配置项，Kubeadm 配置在 Kubernetes 集群中是存储在ConfigMap 中的，也可将这些配置写入配置文件，方便管理复杂的配置项。Kubeadm 配内容是通过 kubeadm config 命令写入配置文件的。
在master节点安装，master 定于为192.168.147.137，通过如下指令创建默认的init-config.yaml文件：
[root@k8s-master ~]# kubeadm config print init-defaults > init-config.yaml其中，kubeadm config 除了用于输出配置项到文件中，还提供了其他一些常用功能，如下所示。
- kubeadm config view：查看当前集群中的配置值。
- kubeadm config print join-defaults：输出 kubeadm join 默认参数文件的内容。
- kubeadm config images list：列出所需的镜像列表。
- kubeadm config images pull：拉取镜像到本地。
- kubeadm config upload from-flags：由配置参数生成 ConfigMap。
init-config.yaml配置
[root@k8s-master ~]# cat init-config.yaml
apiVersion: kubeadm.k8s.io/v1beta2
bootstrapTokens:
- groups:
- system:bootstrappers:kubeadm:default-node-token
token: abcdef.0123456789abcdef
ttl: 24h0m0s
usages:
- signing
- authentication
kind: InitConfiguration
localAPIEndpoint:
advertiseAddress: 192.168.147.137 //master节点IP地址
bindPort: 6443
nodeRegistration:
criSocket: /var/run/dockershim.sock
name: k8s-master //如果使用域名保证可以解析，或直接使用 IP 地址
taints:
- effect: NoSchedule
key: node-role.kubernetes.io/master
---
apiServer:
timeoutForControlPlane: 4m0s
apiVersion: kubeadm.k8s.io/v1beta2
certificatesDir: /etc/kubernetes/pki
clusterName: kubernetes
controllerManager: {}
dns:
type: CoreDNS
etcd:
local:
dataDir: /var/lib/etcd //etcd 容器挂载到本地的目录
imageRepository: registry.aliyuncs.com/google_containers //修改为国内地址
kind: ClusterConfiguration
kubernetesVersion: v1.17.0
networking:
dnsDomain: cluster.local
serviceSubnet: 10.96.0.0/12
podSubnet: 10.244.0.0/16 //新增加 Pod 网段
scheduler: {}
2.5、安装master节点
拉取所需镜像
[root@k8s-master ~]# kubeadm config images list --config init-config.yaml
W0817 09:27:02.869422 20103 validation.go:28] Cannot validate kube-proxy config - no validator is available
W0817 09:27:02.869465 20103 validation.go:28] Cannot validate kubelet config - no validator is available
registry.aliyuncs.com/google_containers/kube-apiserver:v1.17.0
registry.aliyuncs.com/google_containers/kube-controller-manager:v1.17.0
registry.aliyuncs.com/google_containers/kube-scheduler:v1.17.0
registry.aliyuncs.com/google_containers/kube-proxy:v1.17.0
registry.aliyuncs.com/google_containers/pause:3.1
registry.aliyuncs.com/google_containers/etcd:3.4.3-0
registry.aliyuncs.com/google_containers/coredns:1.6.5[root@k8s-master ~]# kubeadm config images pull --config init-config.yaml
W0817 09:27:46.519233 20108 validation.go:28] Cannot validate kube-proxy config - no validator is available
W0817 09:27:46.519271 20108 validation.go:28] Cannot validate kubelet config - no validator is available
[config/images] Pulled registry.aliyuncs.com/google_containers/kube-apiserver:v1.17.0
[config/images] Pulled registry.aliyuncs.com/google_containers/kube-controller-manager:v1.17.0
[config/images] Pulled registry.aliyuncs.com/google_containers/kube-scheduler:v1.17.0
[config/images] Pulled registry.aliyuncs.com/google_containers/kube-proxy:v1.17.0
[config/images] Pulled registry.aliyuncs.com/google_containers/pause:3.1
[config/images] Pulled registry.aliyuncs.com/google_containers/etcd:3.4.3-0
[config/images] Pulled registry.aliyuncs.com/google_containers/coredns:1.6.5
安装matser节点
[root@k8s-master ~]# kubeadm init --config=init-config.yaml //初始化安装K8S根据提示操作
kubectl 默认会在执行的用户家目录下面的.kube 目录下寻找config 文件。这里是将在初始化时[kubeconfig]步骤生成的admin.conf 拷贝到.kube/config
[root@k8s-master ~]# mkdir -p $HOME/.kube
[root@k8s-master ~]# cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
[root@k8s-master ~]# chown $(id -u):$(id -g) $HOME/.kube/config
kubeadm init 主要执行了以下操作：
[init]：指定版本进行初始化操作
[preflight] ：初始化前的检查和下载所需要的Docker镜像文件
[kubelet-start] ：生成kubelet 的配置文件”/var/lib/kubelet/config.yaml”，没有这个文件kubelet无法启动，所以初始化之前的kubelet 实际上启动失败。
[certificates]：生成Kubernetes 使用的证书，存放在/etc/kubernetes/pki 目录中。
[kubeconfig] ：生成 Kubeconfig 文件，存放在/etc/kubernetes 目录中，组件之间通信需要使用对应文件。
[control-plane]：使用/etc/kubernetes/manifest 目录下的YAML 文件，安装 Master 组件。
[etcd]：使用/etc/kubernetes/manifest/etcd.yaml 安装Etcd 服务。
[wait-control-plane]：等待control-plan 部署的Master 组件启动。
[apiclient]：检查Master组件服务状态。
[uploadconfig]：更新配置
[kubelet]：使用configMap 配置kubelet。
[patchnode]：更新CNI信息到Node 上，通过注释的方式记录。
[mark-control-plane]：为当前节点打标签，打了角色Master，和不可调度标签，这样默认就不会使用Master 节点来运行Pod。
[bootstrap-token]：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
[addons]：安装附加组件CoreDNS 和kube-proxy
Kubeadm 通过初始化安装是不包括网络插件的，也就是说初始化之后是不具备相关网络功能的，比如 k8s-master 节点上查看节点信息都是“Not Ready”状态、Pod 的 CoreDNS无法提供服务等。
2.6、安装node节点
根据master安装时的提示信息
[root@k8s-node01 ~]# kubeadm join 192.168.147.137:6443 --token abcdef.0123456789abcdef --discovery-token-ca-cert-hash sha256:d3d51c71ca07191581fff8a91c70fbfb5ffe9c620eafa71e6e256b6fb0ea3488
[root@k8s-master ~]# kubectl get nodes
NAME STATUS ROLES AGE VERSION
k8s-master NotReady master 100s v1.17.0
k8s-node01 NotReady 30s v1.17.0
k8s-node02 NotReady 25s v1.17.0
前面已经提到，在初始化 k8s-master 时并没有网络相关配置，所以无法跟 node 节点通信，因此状态都是“NotReady”。但是通过 kubeadm join 加入的 node 节点已经在k8s-master 上可以看到。
2.7、安装flannel、cni
Master 节点NotReady 的原因就是因为没有使用任何的网络插件，此时Node 和Master的连接还不正常。目前最流行的Kubernetes 网络插件有Flannel、Calico、Canal、Weave 这里选择使用flannel。
所有主机：
master上传kube-flannel.yml，所有主机上传flannel_v0.12.0-amd64.tar、cni-plugins-linux-amd64-v0.8.6.tgz
[root@k8s-master ~]# docker load < flannel_v0.12.0-amd64.tar
256a7af3acb1: Loading layer 5.844MB/5.844MB
d572e5d9d39b: Loading layer 10.37MB/10.37MB
57c10be5852f: Loading layer 2.249MB/2.249MB
7412f8eefb77: Loading layer 35.26MB/35.26MB
05116c9ff7bf: Loading layer 5.12kB/5.12kB
Loaded image: quay.io/coreos/flannel:v0.12.0-amd64
[root@k8s-master ~]# tar xf cni-plugins-linux-amd64-v0.8.6.tgz
[root@k8s-master ~]# cp flannel /opt/cni/bin/
master主机：上传并创建kube-flannel.yaml
[root@k8s-master ~]# kubectl apply -f kube-flannel.yml
podsecuritypolicy.policy/psp.flannel.unprivileged created
clusterrole.rbac.authorization.k8s.io/flannel created
clusterrolebinding.rbac.authorization.k8s.io/flannel created
serviceaccount/flannel created
configmap/kube-flannel-cfg created
daemonset.apps/kube-flannel-ds-amd64 created
daemonset.apps/kube-flannel-ds-arm64 created
daemonset.apps/kube-flannel-ds-arm created
daemonset.apps/kube-flannel-ds-ppc64le created
daemonset.apps/kube-flannel-ds-s390x created
[root@k8s-master ~]# kubectl get nodes
NAME STATUS ROLES AGE VERSION
k8s-master Ready master 5m6s v1.17.0
k8s-node01 Ready 3m56s v1.17.0
k8s-node02 Ready 3m51s v1.17.0
[root@k8s-master ~]# kubectl get pods -n kube-system
NAME READY STATUS RESTARTS AGE
coredns-9d85f5447-dmf55 1/1 Running 0 16m
coredns-9d85f5447-mx5f8 1/1 Running 0 16m
etcd-k8s-master 1/1 Running 0 16m
kube-apiserver-k8s-master 1/1 Running 0 16m
kube-controller-manager-k8s-master 1/1 Running 0 16m
kube-flannel-ds-amd64-hdfz4 1/1 Running 0 11m
kube-flannel-ds-amd64-td6jh 1/1 Running 0 11m
kube-flannel-ds-amd64-wggsc 1/1 Running 0 11m
kube-proxy-8tscv 1/1 Running 0 15m
kube-proxy-sf5qz 1/1 Running 0 16m
kube-proxy-z54n6 1/1 Running 0 15m
kube-scheduler-k8s-master 1/1 Running 0 16m
2.8 节点管理命令
以下命令无需执行，仅作为了解
重置master和node配置
[root@k8s-master ~]# kubeadm reset删除node配置
[root@k8s-master ~]# kubectl delete node 192.168.147.137
[root@k8s-node01 ~]# docker rm -f $(docker ps -aq)
[root@k8s-node01 ~]# systemctl stop kubelet
[root@k8s-node01 ~]# rm -rf /etc/kubernetes/*
[root@k8s-node01 ~]# rm -rf /var/lib/kubelet/*
三、安装Dashboard UI
3.1、部署Dashboard
dashboard的github仓库地址：https://github.com/kubernetes/dashboard
代码仓库当中，有给出安装示例的相关部署文件，我们可以直接获取之后，直接部署即可
[root@k8s-master ~]# wget https://raw.githubusercontent.com/kubernetes/dashboard/master/aio/deploy/recommended.yaml默认这个部署文件当中，会单独创建一个名为kubernetes-dashboard的命名空间，并将kubernetes-dashboard部署在该命名空间下。dashboard的镜像来自docker hub官方，所以可不用修改镜像地址，直接从官方获取即可。
3.2、开放端口设置
在默认情况下，dashboard并不对外开放访问端口，这里简化操作，直接使用nodePort的方式将其端口暴露出来，修改serivce部分的定义：
所有主机下载镜像
[root@k8s-master ~]# docker pull kubernetesui/dashboard:v2.0.0
v2.0.0: Pulling from kubernetesui/dashboard
2a43ce254c7f: Pull complete
Digest: sha256:06868692fb9a7f2ede1a06de1b7b32afabc40ec739c1181d83b5ed3eb147ec6e
Status: Downloaded newer image for kubernetesui/dashboard:v2.0.0
docker.io/kubernetesui/dashboard:v2.0.0
[root@k8s-master ~]# docker pull kubernetesui/metrics-scraper:v1.0.4
v1.0.4: Pulling from kubernetesui/metrics-scraper
07008dc53a3e: Pull complete
1f8ea7f93b39: Pull complete
04d0e0aeff30: Pull complete
Digest: sha256:555981a24f184420f3be0c79d4efb6c948a85cfce84034f85a563f4151a81cbf
Status: Downloaded newer image for kubernetesui/metrics-scraper:v1.0.4
docker.io/kubernetesui/metrics-scraper:v1.0.4
[root@k8s-master ~]# vim recommended.yaml
39 spec:
40 type: NodePort
41 ports:
42 - port: 443
43 targetPort: 8443
44 nodePort: 32443
45 selector:
3.3、权限配置
配置一个超级管理员权限
[root@k8s-master ~]# vim recommended.yaml
164 name: cluster-admin
[root@k8s-master ~]# kubectl apply -f recommended.yaml
namespace/kubernetes-dashboard created
serviceaccount/kubernetes-dashboard created
service/kubernetes-dashboard created
secret/kubernetes-dashboard-certs created
secret/kubernetes-dashboard-csrf created
secret/kubernetes-dashboard-key-holder created
configmap/kubernetes-dashboard-settings created
role.rbac.authorization.k8s.io/kubernetes-dashboard created
clusterrole.rbac.authorization.k8s.io/kubernetes-dashboard created
rolebinding.rbac.authorization.k8s.io/kubernetes-dashboard created
clusterrolebinding.rbac.authorization.k8s.io/kubernetes-dashboard created
deployment.apps/kubernetes-dashboard created
service/dashboard-metrics-scraper created
deployment.apps/dashboard-metrics-scraper created
[root@k8s-master ~]# kubectl get pods -n kubernetes-dashboard
NAME READY STATUS RESTARTS AGE
dashboard-metrics-scraper-c79c65bb7-k9xv9 1/1 Running 0 5m40s
kubernetes-dashboard-56484d4c5-dkfgm 1/1 Running 0 5m40s
[root@k8s-master ~]# kubectl get pods -A -o wide
NAMESPACE NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
kube-system coredns-9d85f5447-dmf55 1/1 Running 0 29m 10.244.1.2 k8s-node01
kube-system coredns-9d85f5447-mx5f8 1/1 Running 0 29m 10.244.2.2 k8s-node02
kube-system etcd-k8s-master 1/1 Running 0 30m 192.168.147.137 k8s-master
kube-system kube-apiserver-k8s-master 1/1 Running 0 30m 192.168.147.137 k8s-master
kube-system kube-controller-manager-k8s-master 1/1 Running 0 30m 192.168.147.137 k8s-master
kube-system kube-flannel-ds-amd64-hdfz4 1/1 Running 0 25m 192.168.147.139 k8s-node01
kube-system kube-flannel-ds-amd64-td6jh 1/1 Running 0 25m 192.168.147.140 k8s-node02
kube-system kube-flannel-ds-amd64-wggsc 1/1 Running 0 25m 192.168.147.137 k8s-master
kube-system kube-proxy-8tscv 1/1 Running 0 28m 192.168.147.140 k8s-node02
kube-system kube-proxy-sf5qz 1/1 Running 0 29m 192.168.147.137 k8s-master
kube-system kube-proxy-z54n6 1/1 Running 0 29m 192.168.147.139 k8s-node01
kube-system kube-scheduler-k8s-master 1/1 Running 0 30m 192.168.147.137 k8s-master
kubernetes-dashboard dashboard-metrics-scraper-c79c65bb7-k9xv9 1/1 Running 0 6m6s 10.244.2.3 k8s-node02
kubernetes-dashboard kubernetes-dashboard-56484d4c5-dkfgm 1/1 Running 0 6m6s 10.244.1.3 k8s-node01
3.4、访问Token配置
使用谷歌浏览器测试访问 https://192.168.147.137:32443
接受风险并继续
可以看到出现如上图画面，需要我们输入一个kubeconfig文件或者一个token。事实上在安装dashboard时，也为我们默认创建好了一个serviceaccount，为kubernetes-dashboard，并为其生成好了token，我们可以通过如下指令获取该sa的token：
[root@k8s-master ~]# kubectl describe secret -n kubernetes-dashboard $(kubectl get secret -n kubernetes-dashboard |grep kubernetes-dashboard-token | awk '{print $1}') |grep token | awk '{print $2}'
kubernetes-dashboard-token-kxcj9
kubernetes.io/service-account-token
eyJhbGciOiJSUzI1NiIsImtpZCI6IkxfVmx1aDJVblJzODJiZFFlcG8zX0ZhbFpqTXlZOEhLMlhPcDJFdDVkeGsifQ.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.fqUzKZ0n4gWM2pZc2Eyl7AfVQ5NLEQm5GqYvbmsWl2BO5OYuNZr_9LOWZmjzstz4MwqdFktf4C8FYhqZ52DsqAxdXSu1Ob8dVFFgX4cUvI7okSUW1rZQXgheKDC_NZQKWtYRfBbtb86Nk1sfdy2xtKs4KPQA7h-XH1HO3MOvo7yd_yRpv8QkVnwGuJCjgtdgqXSpH20Sqg2FfL4wFi5fhwkskQaNPiwYRnQH24DS6BEZBWEE22BxuLyZm76dge8xEizYmzXwvvhY7EFv6P4U3LV7JENq-6oNFYzlT0o5XXBtHZyO_H78VfCYYnLUg-6nHzNR0oUwfG-i4VXNWLIPrQ
四、metrics-server服务部署
4.1、在Node节点上下载镜像
heapster已经被metrics-server取代，metrics-server是K8S中的资源指标获取工具
所有node节点上
[root@k8s-node02 ~]# docker pull bluersw/metrics-server-amd64:v0.3.6
v0.3.6: Pulling from bluersw/metrics-server-amd64
e8d8785a314f: Pull complete
b2f4b24bed0d: Pull complete
Digest: sha256:c9c4e95068b51d6b33a9dccc61875df07dc650abbf4ac1a19d58b4628f89288b
Status: Downloaded newer image for bluersw/metrics-server-amd64:v0.3.6
docker.io/bluersw/metrics-server-amd64:v0.3.6
[root@k8s-node02 ~]# docker tag bluersw/metrics-server-amd64:v0.3.6 k8s.gcr.io/metrics-server-amd64:v0.3.6
4.2、修改 Kubernetes apiserver 启动参数
在kube-apiserver项中添加如下配置选项 修改后apiserver会自动重启
添加一行
[root@k8s-master ~]# vim /etc/kubernetes/manifests/kube-apiserver.yaml
40 - --enable-aggregator-routing=true
4.3、Master上进行部署
[root@k8s-master ~]# wget https://github.com/kubernetes-sigs/metrics-server/releases/download/v0.3.6/components.yaml修改安装脚本：
[root@k8s-master ~]# vim components.yaml
添加2行
91 - --kubelet-preferred-address-types=InternalIP
92 - --kubelet-insecure-tls
[root@k8s-master ~]# kubectl create -f components.yaml
clusterrole.rbac.authorization.k8s.io/system:aggregated-metrics-reader created
clusterrolebinding.rbac.authorization.k8s.io/metrics-server:system:auth-delegator created
rolebinding.rbac.authorization.k8s.io/metrics-server-auth-reader created
apiservice.apiregistration.k8s.io/v1beta1.metrics.k8s.io created
serviceaccount/metrics-server created
deployment.apps/metrics-server created
service/metrics-server created
clusterrole.rbac.authorization.k8s.io/system:metrics-server created
clusterrolebinding.rbac.authorization.k8s.io/system:metrics-server created
等待1-2分钟查看结果
[root@k8s-master ~]# kubectl top nodes
NAME CPU(cores) CPU% MEMORY(bytes) MEMORY%
k8s-master 67m 3% 1066Mi 62%
k8s-node01 26m 1% 846Mi 49%
k8s-node02 29m 1% 808Mi 47% 再回到dashboard界面可以看到CPU和内存使用情况了
五、弹性伸缩
5.1、弹性伸缩介绍
HPA（Horizontal Pod Autoscaler，Pod水平自动伸缩）的操作对象是replication controller, deployment, replica set, stateful set 中的pod数量。注意，Horizontal Pod Autoscaling不适用于无法伸缩的对象，例如DaemonSets。
HPA根据观察到的CPU使用量与用户的阈值进行比对，做出是否需要增减实例（Pods）数量的决策。控制器会定期调整副本控制器或部署中副本的数量，以使观察到的平均CPU利用率与用户指定的目标相匹配。
5.2、弹性伸缩工作原理
Horizontal Pod Autoscaler 会实现为一个控制循环，其周期由--horizontal-pod-autoscaler-sync-period选项指定（默认15秒）。
在每个周期内，controller manager都会根据每个HorizontalPodAutoscaler定义的指定的指标去查询资源利用率。 controller manager从资源指标API（针对每个pod资源指标）或自定义指标API（针对所有其他指标）获取指标。
5.3、弹性伸缩实战
[root@k8s-master ~]# mkdir hpa
[root@k8s-master ~]# cd hpa
创建hpa测试应用的deployment
[root@k8s-master hpa]# vim nginx.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: hpa-test
labels:
app: hpa
spec:
replicas: 1
selector:
matchLabels:
app: nginx
template:
metadata:
labels:
app: nginx
spec:
containers:
- name: nginx
image: nginx:1.19.6
ports:
- containerPort: 80
resources:
requests:
cpu: 0.010
memory: 100Mi
limits:
cpu: 0.010
memory: 100Mi
使用的资源是： CPU 0.010个核，内存100M
[root@k8s-master hpa]# kubectl apply -f nginx.yaml
deployment.apps/hpa-test created
[root@k8s-master hpa]# kubectl get pod
NAME READY STATUS RESTARTS AGE
hpa-test-5bfcdd875f-8ntvv 1/1 Running 0 2m16s
创建hpa策略
[root@k8s-master hpa]# kubectl autoscale --max=10 --min=1 --cpu-percent=5 deployment hpa-test
horizontalpodautoscaler.autoscaling/hpa-test autoscaled
[root@k8s-master hpa]# kubectl get hpa
NAME REFERENCE TARGETS MINPODS MAXPODS REPLICAS AGE
hpa-test Deployment/hpa-test /5% 1 10 0 7s
模拟业务压力测试
[root@k8s-master hpa]# kubectl get pod -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
hpa-test-5bfcdd875f-bh849 1/1 Running 0 7m57s 10.244.2.5 k8s-node02
hpa-test-5bfcdd875f-hcnv8 1/1 Running 0 7m57s 10.244.1.4 k8s-node01
hpa-test-5bfcdd875f-l4fj7 1/1 Running 0 7m57s 10.244.2.4 k8s-node02
[root@k8s-master ~]# while true;do curl -I 10.244.2.5 ;done 观察资源使用情况及弹性伸缩情况
[root@k8s-master ~]# kubectl get hpa
NAME REFERENCE TARGETS MINPODS MAXPODS REPLICAS AGE
hpa-test Deployment/hpa-test 0%/5% 1 10 3 5m15s
[root@k8s-master ~]# kubectl get pod -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
hpa-test-5bfcdd875f-8ntvv 1/1 Running 0 9m31s 10.244.1.4 k8s-node01
[root@k8s-master ~]# while true;do curl -l 10.244.1.4 ;done
观察资源使用情况及弹性伸缩情况
[root@k8s-master ~]# kubectl get hpa
NAME REFERENCE TARGETS MINPODS MAXPODS REPLICAS AGE
hpa-test Deployment/hpa-test 110%/5% 1 10 10 8m18s
[root@k8s-master ~]# kubectl get pod
NAME READY STATUS RESTARTS AGE
hpa-test-5bfcdd875f-277zn 0/1 ImagePullBackOff 0 4m23s
hpa-test-5bfcdd875f-7lqdh 0/1 ContainerCreating 0 4m7s
hpa-test-5bfcdd875f-878gj 0/1 ImagePullBackOff 0 4m38s
hpa-test-5bfcdd875f-8ntvv 1/1 Running 0 15m
hpa-test-5bfcdd875f-bljwc 1/1 Running 0 4m23s
hpa-test-5bfcdd875f-cgmdx 1/1 Running 0 4m38s
hpa-test-5bfcdd875f-jvpv9 0/1 ImagePullBackOff 0 4m38s
hpa-test-5bfcdd875f-m6gz4 1/1 Running 0 4m23s
hpa-test-5bfcdd875f-tc2hw 0/1 ContainerCreating 0 4m23s
hpa-test-5bfcdd875f-xkblv 1/1 Running 0 4m7s
CPU、内存立马开始上升了
根据电脑性能，弹缩时间不等，多等一会
[root@k8s-master ~]# kubectl get pod
NAME READY STATUS RESTARTS AGE
hpa-test-5bfcdd875f-277zn 1/1 Running 0 22m
hpa-test-5bfcdd875f-7lqdh 1/1 Running 0 21m
hpa-test-5bfcdd875f-878gj 1/1 Running 0 22m
hpa-test-5bfcdd875f-8ntvv 1/1 Running 0 33m
hpa-test-5bfcdd875f-bljwc 1/1 Running 0 22m
hpa-test-5bfcdd875f-cgmdx 1/1 Running 0 22m
hpa-test-5bfcdd875f-jvpv9 1/1 Running 0 22m
hpa-test-5bfcdd875f-m6gz4 1/1 Running 0 22m
hpa-test-5bfcdd875f-tc2hw 1/1 Running 0 22m
hpa-test-5bfcdd875f-xkblv 1/1 Running 0 21m
最终会全部running的
将压力测试终止后，稍等一小会儿pod数量会自动缩减到1
[root@k8s-master ~]# kubectl get pod
NAME READY STATUS RESTARTS AGE
hpa-test-5bfcdd875f-277zn 1/1 Running 0 22m删除hpa策略
[root@k8s-master ~]# kubectl delete hpa hpa-test
horizontalpodautoscaler.autoscaling "hpa-test" deleted
[root@k8s-master ~]# kubectl get hpa
No resources found in default namespace.