Tekton — 通过tekton-operator部署tekton组件

Tektoncd Operator 作为一个 Kubernetes 的扩展，可以方便快捷地在 Kubernetes 集群上安装、升级和管理 Tekton Pipelines、Dashboard、Triggers 等组件。

那么本篇文章介绍在K8S集群中如何通过tekton-operator部署Tekton组件。

版本信息

K8S: v1.28.2
tekton-operator: v0.68.1

部署准备

官方已提供部署文件链接，可自行查找所需版本下载 => https://github.com/tektoncd/operator/releases

https://storage.googleapis.com/tekton-releases/operator/previous/v0.68.1/release.yaml

由于网络原因，我们是无法直接使用官方提供gcr.io镜像，因此需要提前准备好镜像并push到自己的私有仓库。

并且我们还要替换掉release.yaml里所有的gcr.io的镜像：
找到release.yaml里的tekton-operator-lifecycle容器，在环境变量里添加如下内容。
环境变量名规范是：{前缀}_{容器名/参数名}

			- name: IMAGE_PIPELINES_PROXY
              value: registry.cloud-devops.com/tekton-releases/github.com/tektoncd/operator/cmd/kubernetes/proxy-webhook:v0.68.1
            - name: IMAGE_PIPELINES_TEKTON_EVENTS_CONTROLLER
              value: registry.cloud-devops.com/tekton-releases/github.com/tektoncd/pipeline/cmd/events:v0.50.3
            - name: IMAGE_PIPELINES_WEBHOOK
              value: registry.cloud-devops.com/tekton-releases/github.com/tektoncd/pipeline/cmd/webhook:v0.50.3
            - name: IMAGE_PIPELINES_CONTROLLER
              value: registry.cloud-devops.com/tekton-releases/github.com/tektoncd/pipeline/cmd/resolvers:v0.50.3
            - name: IMAGE_PIPELINES_TEKTON_PIPELINES_CONTROLLER
              value: registry.cloud-devops.com/tekton-releases/github.com/tektoncd/pipeline/cmd/controller:v0.50.3
            - name: IMAGE_PIPELINES_ARGS__ENTRYPOINT_IMAGE
              value: registry.cloud-devops.com/tekton-releases/github.com/tektoncd/pipeline/cmd/entrypoint:v0.50.3
            - name: IMAGE_PIPELINES_ARGS__NOP_IMAGE
              value: registry.cloud-devops.com/tekton-releases/github.com/tektoncd/pipeline/cmd/nop:v0.50.3
            - name: IMAGE_PIPELINES_ARGS__SIDECARLOGRESULTS_IMAGE
              value: registry.cloud-devops.com/tekton-releases/github.com/tektoncd/pipeline/cmd/sidecarlogresults:v0.50.3
            - name: IMAGE_PIPELINES_ARGS__WORKINGDIRINIT_IMAGE
              value: registry.cloud-devops.com/tekton-releases/github.com/tektoncd/pipeline/cmd/workingdirinit:v0.50.3
            - name: IMAGE_TRIGGERS_TEKTON_TRIGGERS_CORE_INTERCEPTORS
              value: registry.cloud-devops.com/tekton-releases/github.com/tektoncd/triggers/cmd/interceptors:v0.25.2
            - name: IMAGE_TRIGGERS_TEKTON_TRIGGERS_CONTROLLER
              value: registry.cloud-devops.com/tekton-releases/github.com/tektoncd/triggers/cmd/controller:v0.25.2
            - name: IMAGE_TRIGGERS_WEBHOOK
              value: registry.cloud-devops.com/tekton-releases/github.com/tektoncd/triggers/cmd/webhook:v0.25.2
            - name: IMAGE_TRIGGERS_ARG__EL_IMAGE
              value: registry.cloud-devops.com/tekton-releases/github.com/tektoncd/triggers/cmd/eventlistenersink:v0.25.2
            - name: IMAGE_CHAINS_TEKTON_CHAINS_CONTROLLER
              value: registry.cloud-devops.com/tekton-releases/github.com/tektoncd/chains/cmd/controller:v0.17.1
            - name: IMAGE_JOB_PRUNER_TKN
              value: registry.cloud-devops.com/tekton-releases/dogfooding/tkn

在配置文件中，有如下配置。AUTOINSTALL_COMPONENTS控制tekton-operator在部署完成是否自动部署tekton组件。DEFAULT_TARGET_NAMESPACE指定组件所属命名空间。

apiVersion: v1
data:
  AUTOINSTALL_COMPONENTS: "true"
  DEFAULT_TARGET_NAMESPACE: tekton-pipelines
kind: ConfigMap
metadata:
  labels:
    operator.tekton.dev/release: devel
  name: tekton-config-defaults
  namespace: tekton-operator

如果AUTOINSTALL_COMPONENTS为false，在operator部署成功后，需要我们手动触发tekton组件的部署。
在tekton-operator中有个TektonConfig的自定义资源，创建其他组件的顶级 CRD。
当我们创建自己所需的 TektonConfig 对象后，operator会根据配置中的profile字段帮助我们安装相应的其他组件。

Tekton Operator 内置了 3 个 profile：lite、all、basic

Profile	Installed Component	Platform
lite	Pipeline	Kubernetes, Openshift
basic	Pipeline, Trigger, Chains	Kubernetes, Openshift
all	Pipeline, Trigger, Dashboard, Chains	Kubernetes
	Pipeline, Trigger, Addons, Pipelines as Code, Chains	Openshift

# To install pipelines, triggers, chains and dashboard (use profile 'all')
$ kubectl apply -f https://raw.githubusercontent.com/tektoncd/operator/main/config/crs/kubernetes/config/all/operator_v1alpha1_config_cr.yaml


apiVersion: operator.tekton.dev/v1alpha1
kind: TektonConfig
metadata:
  name: config
spec:
  profile: all     
  targetNamespace: tekton-pipelines   # 指定命名空间
  pruner:
    resources:  # 指定可以自动清理的资源
    - pipelinerun
    - taskrun
    keep: 100  # 清理时要保留的最大资源数
    schedule: "0 8 * * *"  # 清理资源的频率

安装

1. 部署tekton-operator

    $ kubectl create -f release.yaml
   

   查看部署结果：

   $ kubectl get pod -n tekton-operator
   

   

2. 部署tekton组件
   如果AUTOINSTALL_COMPONENTS为false执行这个命令。否则跳过，tekton会自动部署组件。

   $ kubectl create -f operator_v1alpha1_config_cr.yaml
   

   查看结果

   $ kubectl get pod -n tekton-pipelines
   

   

3. 替换dashboard镜像
   因为环境变量里还不支持dashboard的镜像替换，这里需要手动更改。

   $ kubectl get TektonInstallerSet
   

   
   编辑dashboard-main-deployment-xrlc2，替换镜像

   $ kubectl edit TektonInstallerSet dashboard-main-deployment-xrlc2
   

   查看结果
   

4. 访问
   Dashboard 的服务默认通过 ClusterIP 方式进行暴露，我们可以创建一个 Ingress 对象或者修改 Service 为 NodePort 方式进行暴露。

   

卸载tekton组件

卸载tekton其实也很简单，部署时我们知道operator会根据TektonConfig对象部署组件，那么卸载组件也只需要删除这个对象即可。

# 查询对象
$ kubectl get TektonConfig

删除TektonConfig对象

$ kubectl delete TektonConfig config

查看结果，各个组件在逐步删除