Is TCP reliable?Why?

Tektoncd Operator 是一个 Kubernetes 的扩展,用于在 Kubernetes 集群上安装、升级和管理 TektonCD Pipelines、Dashboard、Triggers 等.We just need to write the individual components yaml can be managed directly Tekton 的各种组件.

CRD	描述
​ ​TektonConfig​ ​	Configure what to install and manage Tekton 组件.
​ ​TektonPipeline​ ​	Configure installation management Tekton Pipeline 组件.
​ ​TektonTrigger​ ​	Configure installation management Tekton Trigger 组件.
​ ​TektonDashboard​ ​	Configure installation management Tekton Dashboard 组件.
​ ​TektonResult​ ​	Configure installation management Tekton Result 组件.
​ ​TektonAddon​ ​	Configure the installation management plugin,目前仅支持 Openshift.

安装

安装 Tektoncd Operator 有多种方式.

从 Operator Hub 安装

可以直接前往 Operator Hub 页面 https://operatorhub.io/operator/tektoncd-operator 进行安装,Its life cycle will be determined by  Operator Lifecycle Manager (OLM) 进行管理.

Install using a resource manifest file

可以直接从 Github Release 页面 https://github.com/tektoncd/operator/releases Get the resource manifest file,使用这种方式安装,需要自己管理 Operator 的生命周期.

Just use the following command to install it.

$ kubectl apply -f https://storage.googleapis.com/tekton-releases/operator/latest/release.yaml

由于官方使用的镜像是 gcr 的镜像,所以正常情况下我们是获取不到的,如果你的集群由于某些原因获取不到镜像,可以使用下面的资源清单文件：

$ kubectl apply -f https://my-oss-testing.oss-cn-beijing.aliyuncs.com/k8s/tekton/operator/release.v0.60.0.yml

默认情况下 Tektoncd Operator The created object will be used gcr 的镜像,比如 Tekton Pipelines Mirror of the controller,可以通过环境变量 IMAGE_PIPELINES_TEKTON_PIPELINES_CONTROLLER to specify the corresponding mirror,The environment variables shown below are to override the default gcr How the mirror is configured：

- name: IMAGE_PIPELINES_PROXY
  value: cnych/tekton-operator-proxy-webhook:v0.60.0
- name: IMAGE_JOB_PRUNER_TKN
  value: cnych/tekton-operator-pruner-tkn:v0.60.0
- name: IMAGE_PIPELINES_TEKTON_PIPELINES_CONTROLLER
  value: cnych/tekton-controller:v0.37.2
- name: IMAGE_PIPELINES_WEBHOOK
  value: cnych/tekton-webhook:v0.37.2
- name: IMAGE_PIPELINES_ARG__ENTRYPOINT_IMAGE
  value: cnych/tekton-entrypoint:v0.37.2
- name: IMAGE_PIPELINES_ARG__GIT_IMAGE
  value: cnych/tekton-git-init:v0.37.2
- name: IMAGE_PIPELINES_ARG__IMAGEDIGEST_EXPORTER_IMAGE
  value: cnych/tekton-imagedigestexporter:v0.37.2
- name: IMAGE_PIPELINES_ARG__KUBECONFIG_WRITER_IMAGE
  value: cnych/tekton-kubeconfigwriter:v0.37.2
- name: IMAGE_PIPELINES_ARG__NOP_IMAGE
  value: cnych/tekton-nop:v0.37.2
- name: IMAGE_TRIGGERS_TEKTON_TRIGGERS_CONTROLLER
  value: cnych/tekton-triggers-controller:v0.20.1
- name: IMAGE_TRIGGERS_WEBHOOK
  value: cnych/tekton-triggers-webhook:v0.20.1
- name: IMAGE_TRIGGERS_TEKTON_TRIGGERS_CORE_INTERCEPTORS
  value: cnych/tekton-triggers-interceptors:v0.20.1
- name: IMAGE_TRIGGERS_ARG__EL_IMAGE
  value: cnych/tekton-triggers-eventlistenersink:v0.20.1

The above method will create a file named  tekton-operator 的命名空间,其中包含一个 Operator 和一个 Webhook 的 Pod：

$ kubectl get pods -n tekton-operator
NAME                                       READY   STATUS    RESTARTS   AGE
tekton-operator-9d747548b-67t7m            2/2     Running   0          9m42s
tekton-operator-webhook-6cc769b85d-fssq9   1/1     Running   0          9m42s

安装 Operator 后,You can install what you need Tekton 组件,例如 Tekton Pipeline、Tekton Triggers.

每个 Tekton Components have a custom resource,Used to install and manage components.

$ kubectl get crd |grep tekton |grep operator
tektonchains.operator.tekton.dev                    2022-07-25T00:51:07Z
tektonconfigs.operator.tekton.dev                   2022-07-25T00:51:07Z
tektondashboards.operator.tekton.dev                2022-07-25T00:51:07Z
tektonhubs.operator.tekton.dev                      2022-07-25T00:51:07Z
tektoninstallersets.operator.tekton.dev             2022-07-25T00:51:07Z
tektonpipelines.operator.tekton.dev                 2022-07-25T00:51:07Z
tektonresults.operator.tekton.dev                   2022-07-25T00:51:07Z
tektontriggers.operator.tekton.dev                  2022-07-25T00:51:07Z

其中 TektonConfig is the top level for creating other components CRD.So we just create the one with the required configuration TektonConfig 对象即可,It will help us install the corresponding other components.

TektonConfig Will be created based on the config file passed to it TektonPipeline、TektonTriggers 和其他组件 CR 对象,其中有一个 profile 的字段,Can be used to determine all components to install.

Tektoncd Operator 内置了 3 个 profile：lite、all、basic.

• all：This configuration file will install all components.
• basic：This profile will only be installed TektonPipeline 和 TektonTrigger 组件.
• lite：This profile will only be installed TektonPipeline 组件.

比如我们要安装 pipelines、triggers 和 dashboard,可以使用 all 这个 profile 进行安装,如下资源清单所示：

# tekton-operator-profile-all.yaml
apiVersion: operator.tekton.dev/v1alpha1
kind: TektonConfig
metadata:
  name: config
spec:
  profile: all
  targetNamespace: tekton-pipelines
  pruner:
    resources:
      - pipelinerun
      - taskrun
    keep: 100
    schedule: "0 8 * * *"

其中 targetNamespace Used to specify installation Tekton 组件的命名空间,默认为 tekton-pipelines,pruner 为 Tekton Resources provide automatic cleanup.

• resources：Specifies resources that can be cleaned up automatically.
• keep：The maximum number of resources to keep when cleaning.
• schedule：How often to clean up resources.

You can directly install the above resource objects：

$ kubectl apply -f tekton-operator-profile-all.yaml
$ kubectl get tektonconfig
NAME     VERSION   READY   REASON
config   v0.60.0   True

上面的 TektonConfig The object we configured profile 为 all,会自动为我们创建 tektonpipelines、tektontriggers、tektondashboard 组件对象：

$ kubectl get tektonpipelines
NAME       VERSION   READY   REASON
pipeline   v0.37.0   True
$ kubectl get tektontriggers
NAME      VERSION   READY   REASON
trigger   v0.20.1   True
$ kubectl get tektondashboard
NAME        VERSION   READY   REASON
dashboard   v0.27.0   True

the above cr After the object is created, the corresponding component is automatically created,如下所示：

$ kubectl get pods -n tekton-pipelines
NAME                                                 READY   STATUS             RESTARTS   AGE
tekton-dashboard-84dc6f966b-g8flx                    0/1     ImagePullBackOff   0          3m48s
tekton-operator-proxy-webhook-7587596c79-ld8vm       1/1     Running            0          30m
tekton-pipelines-controller-78bc48896b-sd9fk         1/1     Running            0          30m
tekton-pipelines-webhook-5f48c855b4-js54q            1/1     Running            0          30m
tekton-triggers-controller-668b94cb5b-ggbk7          1/1     Running            0          27m
tekton-triggers-core-interceptors-66b7ddd78c-pq7gb   1/1     Running            0          27m
tekton-triggers-webhook-c8fd7755d-rknch              1/1     Running            0          27m

由于 dashboard Component images do not have corresponding overridden environment variables,So we need to modify it manually：

$ kubectl edit deploy tekton-dashboard -n tekton-pipelines
......
    image: cnych/tekton-dashboard:v0.28.0
......

Dashboard The service is passed by default ClusterIP 方式进行暴露,我们可以手动创建一个 Ingress object or modification Service 为 NodePort 方式进行暴露.

$ kubectl get svc -n tekton-pipelines
NAME                                TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)                              AGE
tekton-dashboard                    ClusterIP   10.102.221.101   <none>        9097/TCP                             28m
tekton-operator-proxy-webhook       ClusterIP   10.96.175.155    <none>        443/TCP                              33m
tekton-pipelines-controller         ClusterIP   10.99.0.85       <none>        9090/TCP,8008/TCP,8080/TCP           33m
tekton-pipelines-webhook            ClusterIP   10.106.195.14    <none>        9090/TCP,8008/TCP,443/TCP,8080/TCP   33m
tekton-triggers-controller          ClusterIP   10.99.84.154     <none>        9000/TCP                             30m
tekton-triggers-core-interceptors   ClusterIP   10.97.83.136     <none>        8443/TCP                             30m
tekton-triggers-webhook             ClusterIP   10.108.88.140    <none>        443/TCP                              30m

测试

Tekon After the components are installed,Next let's run a simple one Pipeline.

首先创建一个如下所示的 Task 任务,in the following tasks bash run in the container echo "Hello, world!" 的命令.

# hello-task.yaml
apiVersion: tekton.dev/v1beta1
kind: Task
metadata:
  name: hello
spec:
  steps:
    - name: hello
      image: bash:latest
      command:
        - echo
      args:
        - "Hello, world!"

Create another one in the same way goodbye 的任务,只需要将上面的 echo 内容修改为 goodbye 即可.

然后就可以定义一个 Pipeline 流水线了,如下所示：

# hello-goodbye-pipeline.yaml
apiVersion: tekton.dev/v1beta1
kind: Pipeline
metadata:
  name: hello-goodbye-pipeline
spec:
  tasks:
    - name: hello
      taskRef:
        name: hello
    - name: goodbye
      runAfter:
        - hello
      taskRef:
        name: goodbye

通过 taskRef 引用对应的 Task 对象.

直接创建上面的资源对象即可：

$ kubectl get pipeline
NAME                     AGE
hello-goodbye-pipeline   24s
$ kubectl get task
NAME      AGE
goodbye   101s
hello     107s

To execute the pipeline,我们还需要创建一个 PipelineRun object will actually execute.

# hello-goodbye-pipeline-run.yaml
apiVersion: tekton.dev/v1beta1
kind: PipelineRun
metadata:
  generateName: hello-goodbye-pipeline-
spec:
  pipelineRef:
    name: hello-goodbye-pipeline

Create the above resource,Note what we use here ​ ​generateName​ ​ 属性,需要使用 ​ ​kubectl create​ ​ 命令进行创建.It will follow the above shortly after normal creation Pipeline The description goes to perform two tasks.

要想卸载 Tekton We also only need to define  ​ ​TektonConfig​ ​ 对象删除即可.

如果不想使用 TektonCD Operator 内置的几个 ​ ​profile​ ​,We can also manually configure different components by ourselves CR 实例,另外 TektonCD Operator There are not many configurable ways available at this stage,Especially for domestic users, the problem of mirroring is a long-standing problem,It can only be overridden globally through environment variables Operator,Some mirrors are not covered at all,If it is used domestically and cannot be used gcr If mirroring, the experience may not be so smooth.