当前位置:   article > 正文

helm安装_Kubernetes 包管理工具—Helm2.13安装和使用

helm安装 values.controller.extraargs "update-status") false

Helm概述

微服务和容器化给复杂应用部署与管理带来了极大的挑战。Helm是目前Kubernetes服务编排领域的唯一开源子项目,做为Kubernetes应用的一个包管理工具,可理解为Kubernetes的apt-get / yum,由 Deis 公司发起,该公司已经被微软收购。Helm通过软件打包的形式,支持发布的版本管理和控制,很大程度上简化了Kubernetes应用部署和管理的复杂性。

随着业务容器化与向微服务架构转变,通过分解巨大的单体应用为多个服务的方式,分解了单体应用的复杂性,使每个微服务都可以独立部署和扩展,实现了敏捷开发和快速迭代和部署。但任何事情都有两面性,虽然微服务给我们带来了很多便利,但由于应用被拆分成多个组件,导致服务数量大幅增加,对于Kubernetest编排来说,每个组件有自己的资源文件,并且可以独立的部署与伸缩,这给采用Kubernetes做应用编排带来了诸多挑战:

  1. 管理、编辑与更新大量的K8s配置文件

  2. 部署一个含有大量配置文件的复杂K8s应用

  3. 分享和复用K8s配置和应用

  4. 参数化配置模板支持多个环境

  5. 管理应用的发布:回滚、diff和查看发布历史

  6. 控制一个部署周期中的某一些环节

  7. 发布后的验证

Helm是kubernetes包管理工具,可以方便快捷的安装、管理、卸载kubernetes应用,类似于Linux操作系统中yum或apt-get软件的作用。

其主要的设计目的:

  • 创建新的chart包

  • 将charts包文件打包压缩tgz格式

  • 上传chart到chart仓库或从仓库中下载charts文件

  • 在kubernetes集群中安装及卸载charts

  • 管理通过helm安装的charts的应用发布周期

Helm 到底解决了什么问题?为什么 Kubernetes 需要 Helm?

Kubernetes 能够很好地组织和编排容器,但它缺少一个更高层次的应用打包工具,而 Helm 就是来干这件事的。

举个例子,我们需要部署一个MySQL服务,Kubernetes则需要部署以下对象:

① 为了能够让外界访问到MySQL,需要部署一个mysql的service;

②需要进行定义MySQL的密码,则需要部署一个Secret;

③Mysql的运行需要持久化的数据存储,此时还需要部署PVC;

④保证后端mysql的运行,还需要部署一个Deployment,以支持以上的对象。

针对以上对象,我们可以使用YAML文件进行定义并部署,但是仅仅对于单个的服务支持,如果应用需要由一个甚至几十个这样的服务组成,并且还需要考虑各种服务的依赖问题,可想而知,这样的组织管理应用的方式就显得繁琐。为此就诞生了一个工具Helm,就是为了解决Kubernetes这种应用部署繁重的现象。

Helm架构和组件

be17d1d4ea56c4cf4ad128bfe91a5ec6.png

Helm主要由Helm客户端、Tiller服务器和Charts仓库组成:

  • helm:Client,GO语言编写,实现管理本地的Chart仓库,可管理Chart,与Tiller服务进行交互,用于发送Chart,实例安装、查询、卸载等操作。

  • Tiller:Server,通常运行在K8S集群之上。用于接收helm发来的Charts和Conifg,合并生成release,完成部署。

简单的说:Helm 客户端负责管理 chart;Tiller 服务器负责管理 release。

Helm Client是用户命令行工具。

Helm Client主要负责如下:

  • 本地chart开发

  • charts仓库管理

  • 与Tiller sever交互

  • 发送预安装的chart

  • 查询release信息

  • 要求升级或卸载已存在的release

Tiller Server是一个部署在Kubernetes集群内部的server,Tiller Server与Helm client、Kubernetes API server进行交互。

Tiller server主要负责如下:

  • 监听来自Helm client的请求

  • 通过chart及其配置构建一次发布

  • 安装chart到Kubernetes集群,并跟踪随后的发布

  • 通过与Kubernetes交互升级或卸载chart

简单的说,client管理charts,而server管理发布release。

1268fc141c8a65631505b5e50df29ef9.png

Helm安装

Helm把Kubernetes资源(比如deployments、services或 ingress等) 打包到一个chart中,而chart被保存到chart仓库。通过chart仓库可用来存储和分享chart。Helm使发布可配置,支持发布应用配置的版本管理,简化了Kubernetes部署应用的版本控制、打包、发布、删除、更新等操作。

本节展示了Helm的Client、Server与本地Chart仓库的安装过程。

1.chart是什么?

helm程序包,比方说部署nginx,需要deployment 和 service的yaml,这两个yaml清单文件就是一个helm程序包,在k8s中把这些yaml清单文件叫做chart图表。

2.values.yaml文件

values.yaml文件为模板中的文件赋值,可以实现我们自定义安装,如果是chart开发者需要自定义模板,如果是chart使用者只需要修改values.yaml即可。

3.helm可理解如下

helm把kubernetes资源打包到一个chart中,制作并完成各个chart和chart本身依赖关系并利用chart仓库实现对外分发,而helm还可实现可配置的对外发布,通过values.yaml文件完成可配置的发布,如果chart版本更新了,helm自动支持滚更更新机制,还可以一键回滚,但是不是适合在生产环境使用,除非具有定义自制chart的能力,helm属于kubernetes一个项目。

4.repository、release、chart关系

repository:用于发布和存储 Chart 的仓库,Helm客户端通过HTTP协议来访问仓库中Chart的索引文件和压缩包。

chart: 一个 Helm 包,其中包含了运行一个应用所需要的镜像、依赖和资源定义等,还可能包含 Kubernetes 集群中的服务定义。

release:在 Kubernetes 集群上运行的 Chart 的一个实例。在同一个集群上,一个 Chart 可以安装很多次,每次安装都会创建一个新的 release。

chart--->通过values.yaml这个文件赋值-->生成release实例

helm 官方网站:

https://helm.sh/

helm2 官方文档地址:

https://v2.helm.sh/docs/

helm2 下载地址:

https://storage.googleapis.com/kubernetes-helm/

helm github地址:

https://github.com/helm/helm/releases

https://github.com/helm/helm/tags

helm 官方chart仓库:

https://hub.helm.sh/charts

https://hub.kubeapps.com/

5.安装helm客户端,在k8s的master01节点操作

从helm下载地址,下载适合自己的helm版本[root@k8s-master01 yaml]# mkdir helm && cd helm[root@k8s-master01 helm]# wget https://storage.googleapis.com/kubernetes-helm/helm-v2.13.1-linux-amd64.tar.gz[root@k8s-master01 helm]# tar -zxvf helm-v2.13.1-linux-amd64.tar.gz[root@k8s-master01 helm]# cp linux-amd64/helm /usr/local/bin[root@k8s-master01 helm]# helm versionClient: &version.Version{SemVer:"v2.13.1", GitCommit:"618447cbf203d147601b4b9bd7f8c37a5d39fbb4", GitTreeState:"clean"}Error: could not find tiller报错,显示没有tiller server端

安装Helm的服务端tiller (方案1 不推荐)

Helm Tiller是Helm的server,Tiller有多种安装方式,比如本地安装或以pod形式部署到Kubernetes集群中。方案1,使用yaml文件安装。

为 Tiller 创建具有集群管理员权限的Service Account

Tiller是Helm的服务端组件。Tiller将被安装在kubernetes集群中,Helm客户端会与其交互,从而使用Helm charts部署应用程序。

Helm将管理k8s集群资源。因此,我们需要向安装在集群kube-system命令空间中的tiller组件添加必要的权限。

所以需要做以下操作:

  1. 创建名称为tiller的Service Account

  2. 创建tiller对Service Account具有集群管理员权限的ClusterRoleBinding。

我们将在一个yaml文件中添加Service Account和clusterRoleBinding。

查看api-versions[root@k8s-master01 helm]# kubectl api-versions | grep rbac.authorization.k8s.iorbac.authorization.k8s.io/v1rbac.authorization.k8s.io/v1beta1创建一个名为tiller-rbac.yaml的文件。[root@k8s-master01 helm]# vim tiller-rbac.yamlapiVersion: v1kind: ServiceAccountmetadata:  name: tiller  namespace: kube-system---apiVersion: rbac.authorization.k8s.io/v1beta1kind: ClusterRoleBindingmetadata:  name: tillerroleRef:  apiGroup: rbac.authorization.k8s.io  kind: ClusterRole  name: cluster-adminsubjects:  - kind: ServiceAccount    name: tiller    namespace: kube-system
通过kubectl更新yaml文件[root@k8s-master01 helm]# kubectl apply -f tiller-rbac.yaml serviceaccount/tiller createdclusterrolebinding.rbac.authorization.k8s.io/tiller created查看创建的sa[root@k8s-master01 helm]# kubectl get sa -n kube-system | grep tillertiller                               1         7s
helm init命令进行初始化时,会用到gcr.io/kubernetes-helm/tiller:版本的镜像,需要提前下载,镜像标签和Helm同版本号tiller image镜像下载方式1:把tiller的镜像压缩包传到k8s的各个node节点,然后手动docker load -i 把镜像解压到本地,镜像所在百度网盘地址如下:百度网盘链接:https://pan.baidu.com/s/1Z_yuava-8W65mn5tla7cFw 提取码:bd2nnode节点导入测试的image镜像[root@k8s-node01 ~]# docker load -i tiler_2_13_1.tar.gz3fc64803ca2d: Loading layer [==================================================>]  4.463MB/4.463MB79395a173ae6: Loading layer [==================================================>]  6.006MB/6.006MBc33cd2d4c63e: Loading layer [==================================================>]  37.16MB/37.16MBd727bd750bf2: Loading layer [==================================================>]  36.89MB/36.89MBLoaded image: gcr.io/kubernetes-helm/tiller:v2.13.1tiller image镜像下载方式2:在 https://hub.docker.com/explore/ 搜索 tiller_v2.13.1node节点下载测试的image镜像,修改与yaml中image地址一致docker pull hekai/gcr.io_kubernetes-helm_tiller_v2.13.1docker tag hekai/gcr.io_kubernetes-helm_tiller_v2.13.1 gcr.io/kubernetes-helm/tiller:v2.13.1
[root@k8s-master01 helm]# vim tiller.yaml---apiVersion: apps/v1kind: Deploymentmetadata:  creationTimestamp: null  labels:    app: helm    name: tiller  name: tiller-deploy  namespace: kube-systemspec:  selector:    matchLabels:     app: helm     name: tiller  replicas: 1  strategy: {}  template:    metadata:      creationTimestamp: null      labels:        app: helm        name: tiller    spec:      automountServiceAccountToken: true      serviceAccount:  tiller      containers:      - env:        - name: TILLER_NAMESPACE          value: kube-system        - name: TILLER_HISTORY_MAX          value: "0"        image: gcr.io/kubernetes-helm/tiller:v2.13.1        imagePullPolicy: IfNotPresent        livenessProbe:          httpGet:            path: /liveness            port: 44135          initialDelaySeconds: 1          timeoutSeconds: 1        name: tiller        ports:        - containerPort: 44134          name: tiller        - containerPort: 44135          name: http        readinessProbe:          httpGet:            path: /readiness            port: 44135          initialDelaySeconds: 1          timeoutSeconds: 1        resources: {}status: {}---apiVersion: v1kind: Servicemetadata:  creationTimestamp: null  labels:    app: helm    name: tiller  name: tiller-deploy  namespace: kube-systemspec:  ports:  - name: tiller    port: 44134    targetPort: tiller  selector:    app: helm    name: tiller  type: ClusterIPstatus:  loadBalancer: {}...
通过kubectl更新yaml文件[root@k8s-master01 helm]# kubectl apply -f tiller.yamldeployment.apps/tiller-deploy createdservice/tiller-deploy created查看tiller pod是否部署完成[root@k8s-master01 helm]# kubectl get pods -n kube-system | grep tillertiller-deploy-7bd89687c8-qvhqq         1/1     Running   0          26s验证 helm 和 tiller版本[root@k8s-master01 helm]# helm versionClient: &version.Version{SemVer:"v2.13.1", GitCommit:"618447cbf203d147601b4b9bd7f8c37a5d39fbb4", GitTreeState:"clean"}Server: &version.Version{SemVer:"v2.13.1", GitCommit:"618447cbf203d147601b4b9bd7f8c37a5d39fbb4", GitTreeState:"clean"}

上面结果可看到helm客户端和服务端版本一致,都安装好了。

安装Helm的服务端tiller (方案2 推荐)

本文以pod安装为例,安装Tiller的最简单方式是helm init, 该命令会检查helm本地环境设置是否正确,helm init会连接kubectl默认连接的kubernetes集群(可以通过kubectl config view查看),一旦连接集群成功,tiller会被安装到kube-system namespace中。

1. helm的安装部署(省略)

2. tiller的安装部署

控制台执行 > helm init 命令,该命令会将从charts仓库中下载charts包,并根据其中的配置部署至kubernetes集群。

默认的charts仓库为 https://kubernetes-charts.storage.googleapis.com/index.yaml

默认使用的tiller镜像为 gcr.io/kubernetes-helm/tiller:v2.13.1

国内由于墙的原因无法直接访问,需要我们自行处理可替代的仓库和镜像版本,通过如下命令进行helm服务端的安装部署:

[root@k8s-master01 helm]# helm init --tiller-image registry.cn-hangzhou.aliyuncs.com/google_containers/tiller:v2.13.1 --stable-repo-url https://kubernetes.oss-cn-hangzhou.aliyuncs.com/charts     Creating /root/.helm/repository/repositories.yaml    Adding stable repo with URL: https://kubernetes.oss-cn-hangzhou.aliyuncs.com/charts    Adding local repo with URL: http://127.0.0.1:8879/charts    $HELM_HOME has been configured at /root/.helm.        Tiller (the Helm server-side component) has been installed into your Kubernetes Cluster.        Please note: by default, Tiller is deployed with an insecure 'allow unauthenticated users' policy.    To prevent this, run `helm init` with the --tiller-tls-verify flag.    For more information on securing your installation see: https://docs.helm.sh/using_helm/#securing-your-helm-installation    Happy Helming!     [root@k8s-master01 helm]# cat /root/.helm/repository/repositories.yamlapiVersion: v1generated: 2020-07-02T12:06:28.679177481+08:00repositories:- caFile: ""  cache: /root/.helm/repository/cache/stable-index.yaml  certFile: ""  keyFile: ""  name: stable  password: ""  url: https://kubernetes.oss-cn-hangzhou.aliyuncs.com/charts  username: ""- caFile: ""  cache: /root/.helm/repository/cache/local-index.yaml  certFile: ""  keyFile: ""  name: local  password: ""  url: http://127.0.0.1:8879/charts  username: ""

如果希望在安装时自定义一些参数,可以参考一下的一些参数:

  • --canary-image:安装canary分支,即项目的Master分支

  • --tiller-image:安装指定版本的镜像,默认和helm同版本

  • --kube-context:安装到指定的Kubernetes集群

  • --tiller-namespace:安装到指定的名称空间,默认为kube-system

稍等一会然后执行如下命令,看到如下输出说明安装成功:

验证 helm 和 tiller 版本[root@k8s-master01 helm]# helm versionClient: &version.Version{SemVer:"v2.13.1", GitCommit:"618447cbf203d147601b4b9bd7f8c37a5d39fbb4", GitTreeState:"clean"}Server: &version.Version{SemVer:"v2.13.1", GitCommit:"618447cbf203d147601b4b9bd7f8c37a5d39fbb4", GitTreeState:"clean"

访问授权

在上面的步骤中我们将tiller所需的资源部署到了kubernetes集群中,但是由于Deployment tiller-deploy没有定义授权的ServiceAccount导致访问apiserver拒绝,执行如下命令为tiller-deploy进行授权:

[root@k8s-master01 helm]# kubectl create serviceaccount --namespace kube-system tiller[root@k8s-master01 helm]# kubectl create clusterrolebinding tiller-cluster-rule --clusterrole=cluster-admin --serviceaccount=kube-system:tiller[root@k8s-master01 helm]# kubectl patch deploy --namespace kube-system tiller-deploy -p '{"spec":{"template":{"spec":{"serviceAccount":"tiller"}}}}'

从kubernetes集群中删除Tiller

Tiller将数据存储在ConfigMap资源当中,删除或重装不会导致数据丢失。

删除Tiller的方法有两种:

(1) kubectl delete deployment tiller-deploy --n kube-system(2) heml reset     由于某种原因,如果引发错误,请使用以下命令强制将其删除    helm reset --force

Helm命令及使用

[root@k8s-master01 helm]# helmThe Kubernetes package managerTo begin working with Helm, run the 'helm init' command:  $ helm initThis will install Tiller to your running Kubernetes cluster.It will also set up any necessary local configuration.Common actions from this point include:- helm search:    search for charts- helm fetch:     download a chart to your local directory to view- helm install:   upload the chart to Kubernetes- helm list:      list releases of chartsEnvironment:  $HELM_HOME           set an alternative location for Helm files. By default, these are stored in ~/.helm  $HELM_HOST           set an alternative Tiller host. The format is host:port  $HELM_NO_PLUGINS     disable plugins. Set HELM_NO_PLUGINS=1 to disable plugins.  $TILLER_NAMESPACE    set an alternative Tiller namespace (default "kube-system")  $KUBECONFIG          set an alternative Kubernetes configuration file (default "~/.kube/config")  $HELM_TLS_CA_CERT    path to TLS CA certificate used to verify the Helm client and Tiller server certificates (default "$HELM_HOME/ca.pem")  $HELM_TLS_CERT       path to TLS client certificate file for authenticating to Tiller (default "$HELM_HOME/cert.pem")  $HELM_TLS_KEY        path to TLS client key file for authenticating to Tiller (default "$HELM_HOME/key.pem")  $HELM_TLS_ENABLE     enable TLS connection between Helm and Tiller (default "false")  $HELM_TLS_VERIFY     enable TLS connection between Helm and Tiller and verify Tiller server certificate (default "false")  $HELM_TLS_HOSTNAME   the hostname or IP address used to verify the Tiller server certificate (default "127.0.0.1")  $HELM_KEY_PASSPHRASE set HELM_KEY_PASSPHRASE to the passphrase of your PGP private key. If set, you will not be prompted for                       the passphrase while signing helm chartsUsage:  helm [command]Available Commands:  completion  Generate autocompletions script for the specified shell (bash or zsh)  create      create a new chart with the given name  delete      given a release name, delete the release from Kubernetes  dependency  manage a chart's dependencies  fetch       download a chart from a repository and (optionally) unpack it in local directory  get         download a named release  help        Help about any command  history     fetch release history  home        displays the location of HELM_HOME  init        initialize Helm on both client and server  inspect     inspect a chart  install     install a chart archive  lint        examines a chart for possible issues  list        list releases  package     package a chart directory into a chart archive  plugin      add, list, or remove Helm plugins  repo        add, list, remove, update, and index chart repositories  reset       uninstalls Tiller from a cluster  rollback    roll back a release to a previous revision  search      search for a keyword in charts  serve       start a local http web server  status      displays the status of the named release  template    locally render templates  test        test a release  upgrade     upgrade a release  verify      verify that a chart at the given path has been signed and is valid  version     print the client/server version informationFlags:      --debug                           enable verbose output  -h, --help                            help for helm      --home string                     location of your Helm config. Overrides $HELM_HOME (default "/root/.helm")      --host string                     address of Tiller. Overrides $HELM_HOST      --kube-context string             name of the kubeconfig context to use      --kubeconfig string               absolute path to the kubeconfig file to use      --tiller-connection-timeout int   the duration (in seconds) Helm will wait to establish a connection to tiller (default 300)      --tiller-namespace string         namespace of Tiller (default "kube-system")Use "helm [command] --help" for more information about a command.=======================helm命令解释:- helm search:    搜索charts- helm fetch:     下载charts到本地目录- helm install:   安装charts- helm list:      列出charts的所有版本用法:  helm [command]命令可用选项:  completion  为指定的shell生成自动补全脚本(bash或zsh)  create      创建一个新的charts  delete      删除指定版本的release  dependency  管理charts的依赖  fetch       下载charts并解压到本地目录  (--untar 不解压)  get         下载一个release  history     release历史信息  home        显示helm的家目录  init        在客户端和服务端初始化helm  inspect     查看charts的详细信息  install     安装charts  lint        检测包的存在问题  list        列出release  package     将chart目录进行打包  plugin      add(增加), list(列出), 或 remove(移除) Helm 插件  repo        add(增加), list(列出), remove(移除), update(更新), and index(索引) chart仓库  reset       卸载tiller  rollback    release版本回滚  search      关键字搜索chart  serve       启动一个本地的http server  status      查看release状态信息  template    本地模板  test        release测试  upgrade     release更新  verify      验证chart的签名和有效期  version     打印客户端和服务端的版本信息=======================helm常用命令分类Charts:helm search   查找可用的Chartshelm inspect  查看指定Chart的基本信息helm install  根据指定的Chart 部署一个Release到K8shelm create   创建自己的Charthelm package  打包Chart,一般是一个压缩包文件release:helm list             列出已经部署的Releasehelm delete [RELEASE] 删除一个Release. 并没有物理删除, 出于审计需要,历史可查。helm status [RELEASE] 查看指定的Release信息,即使使用helm delete命令删除的Release.helm upgrade          升级某个Releasehelm rollback [RELEASE] [REVISION]  回滚Release到指定发布版本helm get values [RELEASE]  查看Release的配置文件值helm ls –deleted           列出已经删除的Releaserepo:helm repo listhelm repo add [RepoName] [RepoUrl]helm repo update=======================Helm 常用命令举例查看版本 helm version查看当前安装的chartshelm list查询nginx chartshelm search nginx下载nginx chartshelm fetch apphub/nginx            不解压helm fetch apphub/nginx --untar    解压查看package详细信息helm inspect chart安装chartshelm install --name nginx --namespaces prod bitnami/nginx查看charts状态helm status nginx删除chartshelm delete --purge nginx查看当前repo仓库资源helm repo list增加repo仓库资源helm repo add stable https://kubernetes.oss-cn-hangzhou.aliyuncs.com/chartshelm repo add --username admin --password password myps https://harbor.pt1.cn/chartrepo/charts更新repo仓库资源helm repo update创建chartshelm create helm_charts测试charts语法helm lint 打包chartscd helm_charts && helm package ./查看生成的yaml文件helm  template  helm_charts-0.1.1.tgz更新imagehelm upgrade --set image.tag=‘v201908‘ test update myharbor/study-api-en-oral回滚relasehelm rollback 2自定义 package 的选项:查询支持的选项helm inspect values stable/mysql自定义 password 持久化存储helm install --name db-mysql --set mysqlRootPassword=anoyi  stable/mysql发布到私有harbor仓库脚本request_url='https://harbor.qing.cn/api/chartrepo/charts/charts'user_name='admin'password='password'chart_file='helm_charts-0.1.3.tgz'curl -i -u "$user_name:$password" -k -X POST "${request_url}" \-H "accept: application/json" \-H "Content-Type: multipart/form-data" \-F "chart=@${chart_file};type=application/x-compressed"echo $result

设置helm命令自动补全 为了方便helm命令的使用

[root@k8s-master01 helm]# source [root@k8s-master01 helm]# echo "source > ~/.bashrc

Charts是Helm的程序包,它们都存在在Charts仓库当中。Kubernetes官方的仓库保存了一系列的Charts,仓库默认的名称为stable。安装Charts到集群时,Helm首先会到官方仓库获取相关的Charts,并创建release。

执行 helm search 查看当前可安装的 chart 

[root@k8s-master01 helm]# helm searchNAME                            CHART VERSION  APP VERSION    DESCRIPTION                                                 stable/acs-engine-autoscaler    2.1.3          2.1.1          Scales worker nodes within agent pools                      stable/aerospike                0.1.7          v3.14.1.2      A Helm chart for Aerospike in Kubernetes                    stable/anchore-engine           0.1.3          0.1.6          Anchore container analysis and policy evaluation engine s...stable/artifactory              7.0.3          5.8.4          Universal Repository Manager supporting all major packagi...stable/artifactory-ha           0.1.0          5.8.4          Universal Repository Manager supporting all major packagi...stable/aws-cluster-autoscaler   0.3.2                         Scales worker nodes within autoscaling groups.              stable/bitcoind                 0.1.0          0.15.1         Bitcoin is an innovative payment network and a new kind o...stable/buildkite                0.2.1          3              Agent for Buildkite                                         stable/centrifugo               2.0.0          1.7.3          Centrifugo is a real-time messaging server.                 .......

这些 chart 都是从哪里来的?

前面说过,Helm 可以像 yum 管理软件包一样管理 chart。 

yum 的软件包存放在仓库中,同样的,Helm 也有仓库。

Helm 安装时已经默认配置好了两个仓库:stable 和 local。

stable 是官方仓库,local 是用户存放自己开发的chart的本地仓库。

执行 helm repo list 进行查看。

[root@k8s-master01 helm]# helm repo listNAME    URL                                                   stable  https://kubernetes.oss-cn-hangzhou.aliyuncs.com/chartslocal   http://127.0.0.1:8879/charts

执行 helm repo update 更新chart仓库

仓库更新有时会提示无法连接[root@k8s-master01 helm]# helm repo update Hang tight while we grab the latest from your chart repositories......Skip local chart repository...Unable to get an update from the "stable" chart repository (https://kubernetes-charts.storage.googleapis.com):Error: Looks like "https://kubernetes-charts.storage.googleapis.com" is not a valid chart repository or cannot be reached: Get https://kubernetes-charts.storage.googleapis.com/index.yaml: dial tcp 172.217.160.112:443: connect: connection refusedUpdate Complete. ⎈ Happy Helming!⎈

执行 helm repo remove 删除默认的stable repo仓库

删除默认的stable repo[root@k8s-master01 helm]# helm repo remove stable"stable" has been removed from your repositories

因为helm默认使用官方的源,国内访问太慢。

执行 helm repod add 添加国内的源

增加阿里云的charts仓库增加微软azure的charts仓库[root@k8s-master01 helm]# helm repo add stable https://kubernetes.oss-cn-hangzhou.aliyuncs.com/charts[root@k8s-master01 helm]# helm repo add apphub https://apphub.aliyuncs.com[root@k8s-master01 helm]# helm repo add azure http://mirror.azure.cn/kubernetes/charts查看chart仓库列表[root@k8s-master01 helm]# helm repo listNAME    URL                                                   local   http://127.0.0.1:8879/charts                          apphub  https://apphub.aliyuncs.com                           azure   http://mirror.azure.cn/kubernetes/charts              stable  https://kubernetes.oss-cn-hangzhou.aliyuncs.com/charts再次更新repo[root@k8s-master01 helm]# helm repo updateHang tight while we grab the latest from your chart repositories......Skip local chart repository...Successfully got an update from the "stable" chart repository...Successfully got an update from the "apphub" chart repository...Successfully got an update from the "azure" chart repositoryUpdate Complete. ⎈ Happy Helming!⎈

与 yum 一样,helm 也支持关键字搜索。

执行 helm search 关键字

[root@k8s-master01 helm]# helm search mysqlNAME                              CHART VERSION  APP VERSION  DESCRIPTION                                                 apphub/mysql                      6.8.0          8.0.19       Chart to create a Highly available MySQL cluster            apphub/mysqldump                  2.6.0          2.4.1        A Helm chart to help backup MySQL databases using mysqldump apphub/mysqlha                    1.0.0          5.7.13       MySQL cluster with a single master and zero or more slave...apphub/prometheus-mysql-exporter  0.5.2          v0.11.0      A Helm chart for prometheus mysql exporter with cloudsqlp...azure/mysql                       1.6.4          5.7.30       Fast, reliable, scalable, and easy to use open-source rel...azure/mysqldump                   2.6.0          2.4.1        A Helm chart to help backup MySQL databases using mysqldump azure/prometheus-mysql-exporter   0.6.0          v0.11.0      A Helm chart for prometheus mysql exporter with cloudsqlp...stable/mysql                      0.3.5                       Fast, reliable, scalable, and easy to use open-source rel...apphub/percona                    1.2.0          5.7.17       free, fully compatible, enhanced, open source drop-in rep...apphub/percona-xtradb-cluster     1.0.3          5.7.19       free, fully compatible, enhanced, open source drop-in rep...apphub/phpmyadmin                 4.2.12         5.0.1        phpMyAdmin is an mysql administration frontend              azure/percona                     1.2.1          5.7.26       free, fully compatible, enhanced, open source drop-in rep...azure/percona-xtradb-cluster      1.0.4          5.7.19       free, fully compatible, enhanced, open source drop-in rep...azure/phpmyadmin                  4.3.5          5.0.1        DEPRECATED phpMyAdmin is an mysql administration frontend   stable/percona                    0.3.0                       free, fully compatible, enhanced, open source drop-in rep...stable/percona-xtradb-cluster     0.0.2          5.7.19       free, fully compatible, enhanced, open source drop-in rep...apphub/mariadb                    7.3.9          10.3.22      Fast, reliable, scalable, and easy to use open-source rel...apphub/mariadb-galera             0.8.1          10.4.12      MariaDB Galera is a multi-master database cluster solutio...azure/gcloud-sqlproxy             0.6.1          1.11         DEPRECATED Google Cloud SQL Proxy                           azure/mariadb                     7.3.14         10.3.22      DEPRECATED Fast, reliable, scalable, and easy to use open...stable/gcloud-sqlproxy            0.2.3                       Google Cloud SQL Proxy                                      stable/mariadb                    2.1.6          10.1.31      Fast, reliable, scalable, and easy to use open-source rel...

包括 DESCRIPTION 在内的所有信息,只要跟关键字匹配,都会显示在结果列表中。

安装 chart 也很简单,从上述列表中选择需要安装的MySQL版本。

执行 helm install 安装 MySQL

[root@k8s-master01 helm]# helm install apphub/mysql -n mysql-proNAME:   mysql-proLAST DEPLOYED: Thu Jul  2 14:57:03 2020NAMESPACE: kube-systemSTATUS: DEPLOYEDRESOURCES:==> v1/ConfigMapNAME              DATA  AGEmysql-pro-master  1     0smysql-pro-slave   1     0s==> v1/Pod(related)NAME                READY  STATUS   RESTARTS  AGEmysql-pro-master-0  0/1    Pending  0         0smysql-pro-slave-0   0/1    Pending  0         0s==> v1/SecretNAME       TYPE    DATA  AGEmysql-pro  Opaque  2     0s==> v1/ServiceNAME             TYPE       CLUSTER-IP     EXTERNAL-IP  PORT(S)   AGEmysql-pro        ClusterIP  10.97.248.71   3306/TCP  0smysql-pro-slave  ClusterIP  10.102.133.28  3306/TCP  0s==> v1/StatefulSetNAME              READY  AGEmysql-pro-master  0/1    0smysql-pro-slave   0/1    0sNOTES:Please be patient while the chart is being deployedTip:  Watch the deployment status using the command: kubectl get pods -w --namespace kube-systemServices:  echo Master: mysql-pro.kube-system.svc.cluster.local:3306  echo Slave:  mysql-pro-slave.kube-system.svc.cluster.local:3306Administrator credentials:  echo Username: root  echo Password : $(kubectl get secret --namespace kube-system mysql-pro -o jsonpath="{.data.mysql-root-password}" | base64 --decode)To connect to your database:  1. Run a pod that you can use as a client:      kubectl run mysql-pro-client --rm --tty -i --restart='Never' --image  docker.io/bitnami/mysql:8.0.19-debian-10-r0 --namespace kube-system --command -- bash  2. To connect to master service (read/write):      mysql -h mysql-pro.kube-system.svc.cluster.local -uroot -p my_database  3. To connect to slave service (read-only):      mysql -h mysql-pro-slave.kube-system.svc.cluster.local -uroot -p my_databaseTo upgrade this helm chart:  1. Obtain the password as described on the 'Administrator credentials' section and set the 'root.password' parameter as shown below:      ROOT_PASSWORD=$(kubectl get secret --namespace kube-system mysql-pro -o jsonpath="{.data.mysql-root-password}" | base64 --decode)      helm upgrade mysql-pro bitnami/mysql --set root.password=$ROOT_PASSWORD以上输出分为三部分:① chart 本次部署的描述信息:NAME      是 release 的名字,如果没用 -n 参数指定,Helm 会随机生成了一个 名字。NAMESPACE 是 release 部署的 namespace,默认是 default,也可以通过 --namespace 指定。STATUS    是 DEPLOYED,表示已经将 chart 部署到集群。② 当前 release 包含的资源:ConfigMap、Pod、Secret、Service 和 StatefulSet,其名字都是 mysql-pro,命名的格式为 ReleasName-ChartName。③ NOTES 部分显示的是 release 的使用方法。比如如何访问 Service,如何获取数据库密码,以及如何连接数据库等

执行 helm status 查看release状态:(输出和上面相同)

[root@k8s-master01 helm]# helm status mysql-proLAST DEPLOYED: Thu Jul  2 15:04:45 2020NAMESPACE: kube-systemSTATUS: DEPLOYED

执行 kubectl get 查看组成 该实例release 的各个对象:

[root@k8s-master01 helm]# kubectl get ConfigMap -n kube-system | grep mysql-promysql-pro-master                     1      50smysql-pro-slave                      1      50smysql-pro.v1                         1      50s[root@k8s-master01 helm]# kubectl get pod -n kube-system | grep mysql-promysql-pro-master-0                     0/1     Pending   0          76smysql-pro-slave-0                      0/1     Pending   0          76s[root@k8s-master01 helm]# kubectl describe pod lucky-sponge-mysql-master-0  -n kube-system......Events:  Type     Reason            Age        From               Message  ----     ------            ----       ----               -------  Warning  FailedScheduling  default-scheduler  running "VolumeBinding" filter plugin for pod "mysql-pro-master-0": pod has unbound immediate PersistentVolumeClaims  Warning  FailedScheduling  default-scheduler  running "VolumeBinding" filter plugin for pod "mysql-pro-master-0": pod has unbound immediate PersistentVolumeClaims[root@k8s-master01 helm]# kubectl get Secret -n kube-system | grep mysql-promysql-pro                                        Opaque                                2      2m45s[root@k8s-master01 helm]# kubectl get service -n kube-system | grep mysql-promysql-pro         ClusterIP   10.97.248.71    3306/TCP                 3m10smysql-pro-slave   ClusterIP   10.102.133.28   3306/TCP                 3m10s[root@k8s-master01 helm]#  kubectl get StatefulSet -n kube-systemNAME               READY   AGEmysql-pro-master   0/1     3m26smysql-pro-slave    0/1     3m26s

由于还没有准备 PersistentVolume 和 PersistentVolumeClaims,当前 实例release 还不可用。

执行 helm hist 查看指定release的历史部署版本信息

[root@k8s-master01 helm]# helm listNAME       REVISION  UPDATED                   STATUS    CHART        APP VERSION  NAMESPACE  mysql-pro  1         Thu Jul  2 14:57:03 2020  DEPLOYED  mysql-6.8.0  8.0.19       kube-system

执行 helm delete 删除 release

执行 helm delete --purge 彻底删除 release

删除chart[root@k8s-master01 helm]# helm delete mysql-prorelease "mysql-pro" deleted确认chart是否删除[root@k8s-master01 helm]# helm ls --all mysql-proNAME       REVISION  UPDATED                   STATUS   CHART        APP VERSION  NAMESPACE  mysql-pro  1         Thu Jul  2 14:57:03 2020  DELETED  mysql-6.8.0  8.0.19       kube-system以上删除,无法彻底删除,需要执行以下操作,彻底删除[root@k8s-master01 helm]# helm del --purge mysql-prorelease "mysql-pro" deleted再次确认chart是否删除[root@k8s-master01 helm]# helm ls --all mysql-pro[root@k8s-master01 helm]# helm hist mysql-proError: release: "mysql-pro" not found

Chart目录结构

chart 是 Helm 的应用打包格式。chart 由一系列文件组成,这些文件描述了 Kubernetes 部署应用时所需要的资源,比如 Service、Deployment、PersistentVolumeClaim、Secret、ConfigMap 等。

单个的 chart 可以非常简单,只用于部署一个服务,比如 Memcached;chart 也可以很复杂,部署整个应用,比如包含 HTTP Servers、 Database、消息中间件、cache 等。

chart 将这些文件放置在预定义的目录结构中,通常整个 chart 被打成 tar 包,而且标注上版本信息,便于 Helm 部署。

以前面 MySQL chart 为例。

一旦安装了某个 chart,可以在 /root/.helm/cache/archive 中找到 chart 的 tar 包。解压,可以看到mysql目录结构,包含各类yaml文件。

[root@k8s-master01 helm]# cd /root/.helm/cache/archive/[root@k8s-master01 archive]# ll-rw-r--r-- 1 root root 17860 Jul  2 12:39 mysql-6.8.0.tgz[root@k8s-master01 archive]# tar -zxvf mysql-6.8.0.tgz[root@k8s-master01 archive]# tree mysqlmysql├── Chart.yaml├── ci│   └── values-production.yaml├── files│   └── docker-entrypoint-initdb.d│       └── README.md├── README.md├── templates│   ├── _helpers.tpl│   ├── initialization-configmap.yaml│   ├── master-configmap.yaml│   ├── master-statefulset.yaml│   ├── master-svc.yaml│   ├── NOTES.txt│   ├── secrets.yaml│   ├── servicemonitor.yaml│   ├── slave-configmap.yaml│   ├── slave-statefulset.yaml│   └── slave-svc.yaml├── values-production.yaml└── values.yaml4 directories, 17 files

Chart仓库

chart仓库用来存储和分享打包的chart,官方chart仓库由Kubernetes Charts维护, Helm允许创建私有chart仓库。

创建chart仓库

chart仓库是一个可用来存储index.yml与打包的chart文件的HTTP server,当要分享chart时,需要上传chart文件到chart仓库。任何一个能能够提供YAML与tar文件的HTTP server都可以当做chart仓库,比如Google Cloud Storage (GCS) bucket、Amazon S3 bucket、Github Pages或创建你自己的web服务器。

Chart仓库结构

一个chart仓库由一个chart包与index.yaml文件组成,index.yaml记录了chart仓库中全部chart的索引,一个本地chart仓库的布局例子如下。

/root/.helm/repository/local/index.yaml文件中记录了chart的诸如名称、url、version等一些metadata信息。(默认是空的)
[root@k8s-master01 helm]# cat /root/.helm/repository/local/index.yaml apiVersion: v1entries: {}generated: 2020-07-02T12:06:29.172607551+08:00
创建本地chart仓库

创建chart仓库有多种方式,本文以创建一个本地仓库为例:

[root@k8s-master01 data]# helm serve &[1] 20630[root@k8s-master01 data]# Regenerating index. This may take a moment.Now serving you on 127.0.0.1:8879

默认情况下该服务只监听 127.0.0.1,如果你要绑定到其它网络接口,可使用以下命令:

[root@k8s-master01 data]# helm serve --address 192.168.56.10:8879 &[2] 25589[root@k8s-master01 data]# Regenerating index. This may take a moment.Now serving you on 192.168.56.10:8879

如果你想使用指定目录来做为 Helm Repository 的存储目录,可以加上 --repo-path 参数

#创建chart仓库目录[root@k8s-master01 helm]# mkdir -p /root/yaml/helm/data/[root@k8s-master01 data]# helm serve --address 192.168.56.10:8879 --repo-path /root/yaml/helm/data/ --url http://192.168.56.10:8879/charts/ &[1] 27156[root@k8s-master01 data]# Regenerating index. This may take a moment.Now serving you on 192.168.56.10:8879
将本地chart仓库加入仓库列表
[root@k8s-master01 data]# helm repo add local http://127.0.0.1:8879"local" has been added to your repositories[root@k8s-master01 helm]# helm repo add local-IP http://192.168.56.10:8879/charts/"local-IP" has been added to your repositories[root@k8s-master01 helm]# helm repo listNAME      URL                                                   local     http://127.0.0.1:8879                                 apphub    https://apphub.aliyuncs.com                           azure     http://mirror.azure.cn/kubernetes/charts              stable    https://kubernetes.oss-cn-hangzhou.aliyuncs.com/chartslocal-IP  http://192.168.56.10:8879/charts/

管理chart仓库

上面步骤中,已经创建了一个本地的chart仓库,接下来讲述如何在chart仓库中维护chart。chart须遵循 SemVer 2 规则填写正确的版本格式。

一旦chart目录已经存在,将chart打包,并移动到的一个新建目录,通过helm repo index 命令将chart的metadata记录在index.yaml文件中。

[root@k8s-master01 helm]# helm create mychartCreating mychart[root@k8s-master01 helm]# helm package mychartSuccessfully packaged chart and saved it to: /root/yaml/helm/mychart-0.1.0.tgz[root@k8s-master01 helm]# mkdir fantastic-charts[root@k8s-master01 helm]# mv mychart-0.1.0.tgz fantastic-charts/[root@k8s-master01 helm]# helm repo index fantastic-charts/ --url=http://192.168.56.10:8879/charts/

上传chart到chart仓库,通过helm repo add命令上传chart到chart仓库:

[root@k8s-master01 helm]# helm repo add fantastic-charts http://192.168.56.10:8879/charts/"fantastic-charts" has been added to your repositories

查看chart是否上传仓库成功:

[root@k8s-master01 helm]# helm repo listNAME              URL                                                   local             http://127.0.0.1:8879                                 apphub            https://apphub.aliyuncs.com                           azure             http://mirror.azure.cn/kubernetes/charts              stable            https://kubernetes.oss-cn-hangzhou.aliyuncs.com/chartslocal-IP          http://192.168.56.10:8879/charts/                     fantastic-charts  http://192.168.56.10:8879/charts/

查找上传的chart:

[root@k8s-master01 helm]# helm search mychart -lNAME           CHART VERSION  APP VERSION  DESCRIPTION                local/mychart  0.1.0          1.0          A Helm chart for Kubernetes

查看helm目录结构

[root@k8s-master01 helm]# tree /root/.helm//root/.helm/├── cache│   └── archive│       ├── mysql│       │   ├── Chart.yaml│       │   ├── ci│       │   │   └── values-production.yaml│       │   ├── files│       │   │   └── docker-entrypoint-initdb.d│       │   │       └── README.md│       │   ├── README.md│       │   ├── templates│       │   │   ├── _helpers.tpl│       │   │   ├── initialization-configmap.yaml│       │   │   ├── master-configmap.yaml│       │   │   ├── master-statefulset.yaml│       │   │   ├── master-svc.yaml│       │   │   ├── NOTES.txt│       │   │   ├── secrets.yaml│       │   │   ├── servicemonitor.yaml│       │   │   ├── slave-configmap.yaml│       │   │   ├── slave-statefulset.yaml│       │   │   └── slave-svc.yaml│       │   ├── values-production.yaml│       │   └── values.yaml│       └── mysql-6.8.0.tgz├── plugins├── repository│   ├── cache│   │   ├── apphub-index.yaml│   │   ├── azure-index.yaml│   │   ├── fantastic-charts-index.yaml│   │   ├── local-index.yaml -> /root/.helm/repository/local/index.yaml│   │   ├── local-IP-index.yaml│   │   └── stable-index.yaml│   ├── local│   │   ├── index.yaml│   │   └── mychart-0.1.0.tgz│   └── repositories.yaml└── starters12 directories, 27 files[root@k8s-master01 helm]# cat /root/.helm/repository/cache/apphub-index.yaml | less[root@k8s-master01 helm]# cat /root/.helm/repository/cache/azure-index.yaml[root@k8s-master01 helm]# cat /root/.helm/repository/cache/stable-index.yaml

Helm部署应用

案例1:Helm部署ingress Nginx应用

创建namespace命名空间[root@k8s-master01 ~]# kubectl create ns ingress-nginxnamespace/ingress-nginx created
查询需要部署的应用[root@k8s-master01 helm]# helm search nginx-ingressNAME                             CHART VERSION  APP VERSION  DESCRIPTION                                                 apphub/nginx-ingress             1.30.3         0.28.0       An nginx Ingress controller that uses ConfigMap to store ...apphub/nginx-ingress-controller  5.3.4          0.29.0       Chart for the nginx Ingress controller                      azure/nginx-ingress              1.39.1         0.32.0       An nginx Ingress controller that uses ConfigMap to store ...stable/nginx-ingress             1.39.1         0.32.0       An nginx Ingress controller that uses ConfigMap to store ...apphub/nginx-lego                0.3.1                       Chart for nginx-ingress-controller and kube-lego            azure/nginx-lego                 0.3.1                       Chart for nginx-ingress-controller and kube-lego            able/nginx-lego                  0.3.1                       Chart for nginx-ingress-controller and kube-lego 安装nginx-ingress chart包[root@k8s-master01 helm]# helm install apphub/nginx-ingress --set rbac.create=true -n nginx-ingress --namespace ingress-nginx
查看部署应用的pod[root@k8s-master01 helm]# kubectl get pod    (由于国外image无法下载,报错)NAME                                             READY   STATUS             RESTARTS   AGEnginx-ingress-controller-5895796f95-hfk87        0/1     ImagePullBackOff   0          5m27snginx-ingress-default-backend-7c868597f4-wm5rp   0/1     ImagePullBackOff   0          5m27s查看image镜像 (nginx-ingress-controller-5895796f95-hfk87)[root@k8s-master01 helm]# kubectl describe pod nginx-ingress-controller-5895796f95-hfk87 | grep image   在node节点下载image镜像 (由于下载dockerhub镜像较慢,先从阿里云容器镜像下载,再使用 docker tag 修改)[root@k8s-node01 k8s-images]# docker pull registry.cn-hangzhou.aliyuncs.com/google_containers/nginx-ingress-controller:0.32.0[root@k8s-node01 k8s-images]# docker tag registry.cn-hangzhou.aliyuncs.com/google_containers/nginx-ingress-controller:0.32.0 quay.io/kubernetes-ingress-controller/nginx-ingress-controller:0.32.0[root@k8s-node02 k8s-images]# docker pull registry.cn-hangzhou.aliyuncs.com/google_containers/nginx-ingress-controller:0.32.0[root@k8s-node02 k8s-images]# docker tag registry.cn-hangzhou.aliyuncs.com/google_containers/nginx-ingress-controller:0.32.0 quay.io/kubernetes-ingress-controller/nginx-ingress-controller:0.32.0查看image镜像 (nginx-ingress-default-backend-7c868597f4-wm5rp)[root@k8s-master01 helm]# kubectl describe pod nginx-ingress-default-backend-7c868597f4-wm5rp  | grep image在node节点下载image镜像 (由于下载dockerhub镜像较慢,先从阿里云容器镜像下载,再使用 docker tag 修改)[root@k8s-node01 k8s-images]# docker pull registry.cn-hangzhou.aliyuncs.com/lusifeng/defaultbackend-amd64:1.5[root@k8s-node01 k8s-images]# docker tag registry.cn-hangzhou.aliyuncs.com/lusifeng/defaultbackend-amd64:1.5 k8s.gcr.io/defaultbackend-amd64:1.5[root@k8s-node02 k8s-images]# docker pull registry.cn-hangzhou.aliyuncs.com/lusifeng/defaultbackend-amd64:1.5[root@k8s-node02 k8s-images]# docker tag registry.cn-hangzhou.aliyuncs.com/lusifeng/defaultbackend-amd64:1.5 k8s.gcr.io/defaultbackend-amd64:1.5在image镜像下载,docker tag修改完成后,应用pod正常运行[root@k8s-master01 helm]# kubectl get pod -o wideNAME                                             READY   STATUS    RESTARTS   AGE   IP            NODE         NOMINATED NODE   READINESS GATESnginx-ingress-controller-5895796f95-hfk87        1/1     Running   0          25m   10.244.2.78   k8s-node02   nginx-ingress-default-backend-7c868597f4-wm5rp   1/1     Running   0          25m   10.244.2.75   k8s-node02   [root@k8s-master01 helm]# kubectl get deployNAME                            READY   UP-TO-DATE   AVAILABLE   AGEnginx-ingress-controller        1/1     1            1           28mnginx-ingress-default-backend   1/1     1            1           28mSVC稍微有点问题[root@k8s-master01 helm]# kubectl get svcNAME                            TYPE           CLUSTER-IP       EXTERNAL-IP   PORT(S)                                     AGEnginx-ingress-controller        LoadBalancer   10.100.184.216        80:30792/TCP,443:31994/TCP                  28mnginx-ingress-default-backend   ClusterIP      10.101.4.146             80/TCP                                      28m

可以使用以下命令检查 helm ls 否安装。

[root@k8s-master01 helm]# helm lsNAME           REVISION  UPDATED                   STATUS    CHART                 APP VERSION  NAMESPACEnginx-ingress  1         Sun Jun  7 00:26:30 2020  DEPLOYED  nginx-ingress-1.39.1  0.32.0       mon

我们发现svc服务一直pending状态,无法对外服务,这是因为helm默认部署是以LoadBalancer方式,这种方式需要平台的支持,比如在AWS、GCE或者阿里云等平台。而我们自己搭建的集群没办法使用这种方式。不过我们可以通过设置externalIPs的方式使用内部ip,再通过统一的负载均衡接入外网。

可以使用 helm delete 命令删除刚才的部署。

删除 nginx-ingress release 实例[root@k8s-master01 helm]# helm delete nginx-ingressrelease "nginx-ingress" deleted[root@k8s-master01 helm]# helm install stable/nginx-ingress --name nginx-ingressError: a release named nginx-ingress already exists.Run: helm ls --all nginx-ingress; to check the status of the releaseOr run: helm del --purge nginx-ingress; to delete it[root@k8s-master01 helm]# helm ls --all nginx-ingressNAME           REVISION  UPDATED                   STATUS   CHART                 APP VERSION  NAMESPACEnginx-ingress  1         Sun Jun  7 00:26:30 2020  DELETED  nginx-ingress-1.39.1  0.32.0       mon  [root@k8s-master01 helm]# helm del --purge nginx-ingressrelease "nginx-ingress" deleted[root@k8s-master01 helm]# helm ls --all nginx-ingress
重新安装nginx-ingress chart包  [root@k8s-master01 helm]# helm install apphub/nginx-ingress --set controller.service.externalIPs[0]=192.168.56.88 -n nginx-ingress --namespace ingress-nginx[root@k8s-master01 helm]# kubectl get deployments.apps NAME                            READY   UP-TO-DATE   AVAILABLE   AGEnginx-ingress-controller        1/1     1            1           56snginx-ingress-default-backend   1/1     1            1           56s[root@k8s-master01 helm]# kubectl get svc    (外部实际IP地址 192.168.56.88)NAME                            TYPE           CLUSTER-IP      EXTERNAL-IP     PORT(S)                      AGEnginx-ingress-controller        LoadBalancer   10.98.177.150   192.168.56.88   80:30985/TCP,443:31203/TCP   62snginx-ingress-default-backend   ClusterIP      10.106.41.197   80/TCP                       61s[root@k8s-master01 helm]# curl 10.98.177.150default backend - 404[root@k8s-master01 helm]# curl 10.106.41.197default backend - 404[root@k8s-master01 helm]# curl 192.168.56.88  (实验环境不行,实际情况可以访问192.168.56.88,再转到内部 CLuster-IP)default backend - 404参考文章:使用helm部署ingress-nginxhttps://blog.csdn.net/u010039418/article/details/99894421

由于K8S集群内部还没有服务,因此请求转到nginx-ingress的default backend中。

创建应用并使用nginx-ingress对外暴露服务

我们创建一个nginx web服务器作为测试1

在node节点下载nginx:1.18镜像[root@k8s-node01 ~]# docker pull nginx:1.18  [root@k8s-node02 ~]# docker pull nginx:1.18
在master节点编辑 nginx-web.yaml[root@k8s-master01 helm]# vim nginx-web.yaml---apiVersion: v1kind: Podmetadata:name: nginxnamespace: ingress-nginxlabels:app: webspec:hostNetwork: falsecontainers:   - name: nginximage: nginx:1.18---kind: ServiceapiVersion: v1metadata:name: webservicenamespace: ingress-nginxspec:selector:app: webports:  - protocol: TCPport: 10080    targetPort: 8
[root@k8s-master01 helm]# kubectl apply -f nginx-web.yaml pod/nginx created[root@k8s-master01 helm]# kubectl get svcNAME                            TYPE           CLUSTER-IP      EXTERNAL-IP     PORT(S)                      AGEnginx-ingress-controller        LoadBalancer   10.98.177.150   192.168.56.88   80:30985/TCP,443:31203/TCP   15mnginx-ingress-default-backend   ClusterIP      10.106.41.197   <none>          80/TCP                       15mwebservice                      ClusterIP      10.97.67.174    <none>          10080/TCP                    86s[root@k8s-master01 helm]# kubectl get pod -o wideNAME                                             READY   STATUS        RESTARTS   AGE   IP            NODE         NOMINATED NODE   READINESS GATESnginx                                            1/1     Running       0          23m   10.244.2.96   k8s-node02   <none>           <none>nginx-ingress-controller-5f866545d4-jvxtb        1/1     Running       0          36m   10.244.2.95   k8s-node02   <none>           <none>nginx-ingress-default-backend-69b685fc96-cnppk   1/1     Running       0          36m   10.244.2.94   k8s-node02   <none>           <none

为了便于测试,我们进入这个nginx pod,将测试web服务的主页内容修改一下,

[root@k8s-master01 helm]# kubectl exec -ti nginx -n ingress-nginx -- /bin/bashroot@nginx:/#  echo "This is my webservice!" > /usr/share/nginx/html/index.htmlroot@nginx:/#  cat /usr/share/nginx/html/index.htmlThis is my webservice!root@nginx:/# exitexit

确保服务正常,我们在集群内通过 webservice 的 ClusterIP 测试一下,

[root@k8s-master01 helm]# kubectl get svcNAME                            TYPE           CLUSTER-IP      EXTERNAL-IP     PORT(S)                      AGEnginx-ingress-controller        LoadBalancer   10.98.177.150   192.168.56.88   80:30985/TCP,443:31203/TCP   19mnginx-ingress-default-backend   ClusterIP      10.106.41.197   80/TCP                       19mwebservice                      ClusterIP      10.97.67.174    10080/TCP                    6m48s[root@k8s-master01 helm]# curl 10.97.67.174:10080This is my webservice!

可见,服务创建正常,但是该服务目前只能在集群内使用,集群外无法访问该服务。

下面我们通过ingress-nginx来将该服务暴露到集群外部。

首先定义我们的服务的对外要暴露的访问域名,这里假设为 webservice.com

[root@k8s-master01 helm]# kubectl get ingress --all-namespacesNAMESPACE       NAME             CLASS    HOSTS              ADDRESS   PORTS   AGEingress-nginx   test-ingress        webservice.com               80      10s

创建好后,我们试着在K8S集群外部来访问。

[root@CentOS-7-2 ~]# curl -H "Host: webservice.com" http://192.168.56.8880This is my webservice!

可见,通过ingress-nginx,我们在K8S集群外部也能访问集群内部的服务。

此时可以查看下ingress-nginx内部的配置文件

[root@k8s-master01 helm]# kubectl exec -ti nginx-ingress-controller-5f866545d4-jvxtb -n ingress-nginx -- cat /etc/nginx/nginx.conf...    upstream ingress-nginx-webservice-10080 {# Load balance algorithm; empty for round robin, which is the default        least_conn;        keepalive 32;        server 10.244.2.96:80 max_fails=0 fail_timeout=0;    }  ...## start server webservice.com    server {        server_name webservice.com ;        listen 80;        listen [::]:80;        ...        location / {           port_in_redirect off;set $proxy_upstream_name "ingress-nginx-webservice-10080";           ...       }    ...    }## end server webservice.com

可见此时ingress-nginx里面已经自动生成了webservice服务的转发规则,无需我们手动再去添加。再看下后端服务ip 10.244.2.96是哪个pod。

[root@k8s-master01 helm]# kubectl get pods --all-namespaces -o wide | grep 10.244.2.96ingress-nginx          nginx                                            1/1     Running       0          25m     10.244.2.96     k8s-node02     <none>           <none>

正是我们启动的nginx服务。

这样就实现了将集群内部服务暴露到外部。

案例2:Helm部署WordPress应用

这里以一个典型的三层应用 Wordpress 为例,包括 MySQL、PHP 和 Apache。

查询需要部署的应用[root@k8s-master01 helm]# helm search wordpressNAME              CHART VERSION  APP VERSION  DESCRIPTION                                                 apphub/wordpress  8.1.3          5.3.2        Web publishing platform for building blogs and websites.    azure/wordpress   9.0.3          5.3.2        DEPRECATED Web publishing platform for building blogs and...stable/wordpress  0.8.8          4.9.4        Web publishing platform for building blogs and websites. 安装需要的应用 ()[root@k8s-master01 helm]# helm install --name wordpress-test --set "persistence.enabled=false,mariadb.persistence.enabled=false" apphub/wordpressNAME:   wordpress-testLAST DEPLOYED: Thu Jul  2 15:45:25 2020NAMESPACE: kube-systemSTATUS: DEPLOYEDRESOURCES:==> v1/ConfigMapNAME                          DATA  AGEwordpress-test-mariadb        1     1swordpress-test-mariadb-tests  1     1s==> v1/DeploymentNAME            READY  UP-TO-DATE  AVAILABLE  AGEwordpress-test  0/1    1           0          1s==> v1/Pod(related)NAME                             READY  STATUS             RESTARTS  AGEwordpress-test-5495b7fc87-2k4c6  0/1    ContainerCreating  0         0swordpress-test-mariadb-0         0/1    Pending            0         0s==> v1/SecretNAME                    TYPE    DATA  AGEwordpress-test          Opaque  1     1swordpress-test-mariadb  Opaque  2     1s==> v1/ServiceNAME                    TYPE          CLUSTER-IP      EXTERNAL-IP  PORT(S)                     AGEwordpress-test          LoadBalancer  10.100.184.216  80:31977/TCP,443:31916/TCP  1swordpress-test-mariadb  ClusterIP     10.100.156.107  3306/TCP                    1s==> v1/StatefulSetNAME                    READY  AGEwordpress-test-mariadb  0/1    1sNOTES:1. Get the WordPress URL:  NOTE: It may take a few minutes for the LoadBalancer IP to be available.        Watch the status with: 'kubectl get svc --namespace kube-system -w wordpress-test'  export SERVICE_IP=$(kubectl get svc --namespace kube-system wordpress-test --template "{{ range (index .status.loadBalancer.ingress 0) }}{{.}}{{ end }}")  echo "WordPress URL: http://$SERVICE_IP/"  echo "WordPress Admin URL: http://$SERVICE_IP/admin"2. Login with the following credentials to see your blog  echo Username: user  echo Password: $(kubectl get secret --namespace kube-system wordpress-test -o jsonpath="{.data.wordpress-password}" | base64 --decode)
查看部署应用的pod[root@k8s-master01 helm]# kubectl get pod -o wide | grep wordpress   (由于国外image无法下载,报错)wordpress-test-78db4789f8-xlhmr                  0/1     ContainerCreating   0          95s            k8s-node02   wordpress-test-mariadb-0                         0/1     Pending             0          94s    查看image镜像 (wordpress-test-78db4789f8-xlhmr)[root@k8s-master01 helm]# kubectl describe pod wordpress-test-5495b7fc87-2k4c6  | grep image  Normal  Pulling    11m        kubelet, localhost.localdomain  Pulling image "docker.io/bitnami/wordpress:5.3.2-debian-10-r0"在node节点下载image镜像 (由于下载dockerhub镜像较慢,在阿里云镜像仓库中未找到合适镜像 )[root@k8s-node01 k8s-images]# docker pull docker.io/bitnami/wordpress:5.3.2-debian-10-r0[root@k8s-node02 k8s-images]# docker pull docker.io/bitnami/wordpress:5.3.2-debian-10-r0查看image镜像 (wordpress-test-mariadb-0)[root@k8s-master01 helm]# kubectl describe pod wordpress-test-mariadb-0......Events:  Type     Reason            Age        From               Message  ----     ------            ----       ----               -------  Warning  FailedScheduling  default-scheduler  running "VolumeBinding" filter plugin for pod "wordpress-test-mariadb-0": pod has unbound immediate PersistentVolumeClaims  Warning  FailedScheduling  default-scheduler  running "VolumeBinding" filter plugin for pod "wordpress-test-mariadb-0": pod has unbound immediate PersistentVolumeClaims需要解决以上image问题 和 PersistentVolumeClaims问题[root@k8s-master01 helm]# kubectl get deploy -o wide | grep wordpresswordpress-test   0/1     1            0           13m     wordpress                             docker.io/bitnami/wordpress:5.3.2-debian-10-r0                          app=wordpress-test,release=wordpress-test[root@k8s-master01 helm]# kubectl get svc -o wide | grep wordpresswordpress-test                  LoadBalancer   10.96.19.254     80:30953/TCP,443:31387/TCP                  13m     app.kubernetes.io/instance=wordpress-test,app.kubernetes.io/name=wordpresswordpress-test-mariadb          ClusterIP      10.110.43.128    3306/TCP                                    13m     app=mariadb,component=master,release=wordpress-test[root@k8s-master01 helm]# kubectl get svc -o wide | grep wordpresswordpress-test           LoadBalancer   10.100.184.216   80:31977/TCP,443:31916/TCP   14m     app=wordpress-testwordpress-test-mariadb   ClusterIP      10.100.156.107   3306/TCP                     14m     app=mariadb,component=master,release=wordpress-test[root@k8s-master01 helm]# kubectl get secret --namespace default wordpress-test-wordpress -o jsonpath="{.data.wordpress-password}" | base64 --decode在浏览器中打开页面,并输入用户名和密码就可以看到搭建好的WordPress博客网站了
升级应用当有新的chart包发布时或者想改变已有release的配置时,可以通过 helm upgrade命令实现,比如:[root@k8s-master01 helm]# helm upgrade wordpress-test --set "persistence.enabled=true,mariadb.persistence.enabled=true" stable/wordpress

Helm对release的版本管理

在上面例子中,已经展示了Helm对release的非常强大的版本管理功能,比如通过”helm list -a”查看有哪些release,通过” helm hist“查看某一个具体的release发布过的历史版本,以及通过” helm get –revision”,查看某个release的一次历史版本对应的具体应用配置信息等。即使已经被删除的release仍然有记录,并且通过Helm能够快速回滚到已删除release的某个发布过的历史版本。Helm的这些版本管理功能,Kubernetes原生并不支持。

往期文章

Kubernetes 是什么?

Kubernetes 基础概念

Kubernetes 1.18.2集群部署 (单Master)+docker—kubeadm方式

Kubernetes 1.18.2集群部署 (多Master)+docker—kubeadm方式

Kubernetes 1.18.2集群部署 (多Master)+docker—二进制方式

Kubernetes Harbor v2.0.0私有镜像仓库部署-更新

Kubernetes kubectx/kubens切换context和namespace

Kubernetes 删除namespace时卡在Terminating状态

Kubernetes kubeadm初始化kubernetes集群延长证书过期时间

Kubernetes kubectl命令

Kubernetes 创建、更新应用

Kubernetes 资源清单

Kubernetes Pod状态和生命周期管理

Kubernetes Pod控制器

Kubernetes ReplicaSet控制器

Kubernetes Deployment控制器

Kubernetes DamonSet控制器

Kubernetes 服务发现Service

Kubernetes 内部服务发现—Coredns

Kubernetes 外部服务发现—Traefik ingress

Kubernetes 外部服务发现—Nginx Ingress Controller

Kubernetes 存储卷

Kubernetes 特殊存储卷—Secret和ConfigMap

Kubernetes StatefulSet控制器

Kubernetes 认证、授权和准入控制

Kubernetes dashboard认证访问

Kubernetes 网络模型和网络策略

Kubernetes 网络原理解析

Kubernetes 网络插件-flannel

Kubernetes 网络插件-calico

Kubernetes Pod资源调度

Kubernetes 资源指标和集群监控

Kubernetes 集群中部署Prometheus+Grafana+Alertmanager监控告警系统

Kubernetes Prometheus监控Nginx

Kubernetes Prometheus监控MySQL

Kubernetes Prometheus监控tomcat

Kubernetes 使用kubeadm升级集群

Kubernetes 使用StatefulSet部署MySQL高可用集群

Kubernetes 使用StatefulSet部署MongoDB高可用集群

Kubernetes 使用Elastic Stack构建Kubernetes全栈监控

Kubernetes 集群中部署Nginx+php-fpm+MySQL并运行Discuz

Kubernetes 集群中部署WordPress博客

声明:本文内容由网友自发贡献,不代表【wpsshop博客】立场,版权归原作者所有,本站不承担相应法律责任。如您发现有侵权的内容,请联系我们。转载请注明出处:https://www.wpsshop.cn/w/盐析白兔/article/detail/237244
推荐阅读
  

闽ICP备14008679号