初试 kubevela-技术圈

KubeVela 是一个简单易用且高度可扩展的应用管理平台与核心引擎，KubeVela 是基于 Kubernetes 与 Open Application Model（OAM）技术构建的。

OAM 全称是 Open Application Model，从名称上来看它所定义的就是一种模型，同时也实现了基于 OAM 的我认为这种模型旨在定义了云原生应用的标准。

开放（Open）：支持异构的平台、容器运行时、调度系统、云供应商、硬件配置等，总之与底层无关
应用（Application）：云原生应用
模型（Model）：定义标准，以使其与底层平台无关

在 OAM 中，一个应用程序包含三个核心理念。

第一个核心理念是组成应用程序的组件（Component），它可能包含微服务集合、数据库和云负载均衡器；
第二个核心理念是描述应用程序运维特征（Trait）的集合，例如，弹性伸缩和 Ingress 等功能。它们对应用程序的运行至关重要，但在不同环境中其实现方式各不相同；
最后，为了将这些描述转化为具体的应用程序，运维人员使用应用配置（Application Configuration）来组合组件和相应的特征，以构建应部署的应用程序的具体实例

对于开发人员来说，KubeVela本身是一个易于使用的工具，能够以最小的工作量描述应用程序并将其发布到Kubernetes。只需管理一个以应用程序为中心的工作流程即可轻松地与任何CI / CD管道集成，无需管理少量的Kubernetes YAML文件，只需一个简单的docker -compose样式的Appfile。

对于平台开发人员来说，KubeVela是一个框架，使他们能够轻松创建面向开发人员的高度可扩展的平台。详细地说，KubeVela通过执行以下操作减轻了构建此类平台的麻烦：

以应用为中心。在Appfile的后面，KubeVela实施了一个应用程序概念，因为它的主要API和ALL KubeVela的功能仅满足应用程序的需求。这是通过采用开放应用程序模型作为KubeVela的核心API来实现的。
本地扩展。KubeVela中的应用程序由各种可插拔工作负载类型和操作功能（即特征）组成。Kubernetes生态系统的功能可以随时通过Kubernetes CRD注册机制作为新的工作负载类型或特征添加到KubeVela中。
简单但可扩展的抽象机制。KubeVela的主要用户界面（即Appfile和CLI）使用基于CUELang的抽象引擎构建，该引擎将面向用户的模式转换为下划线的Kubernetes资源。KubeVela提供了一组内置的抽象对象，平台构建者可以随时随意对其进行修改。抽象更改将在运行时生效，无需重新编译或重新部署KubeVela。

架构

KubeVela 整体架构如下图所示：

在架构上，KubeVela 只有一个 controller 并且以插件的方式运行在 Kubernetes 之上，为 Kubernetes 带来了面向应用层的抽象，以及以此为基础的面向用户的使用界面，即Appfile。Appfile 乃至 KubeVela 运行机制背后的核心，则是其能力管理模型 Open Application Model (OAM) 。基于这个模型，KubeVela 为系统管理员提供了一套基于注册与自发现的能力装配流程，来接入 Kubernetes 生态中的任意能力到 KubeVela 中，从而以“一套核心框架搭配不同能力”的方式，适配各种使用场景

概念和术语

从上面的架构图上我们可以看到kubevela中存在一些专业术语application，service，workload type，trait，他们之间的关系如下图：

workload type：声明运行时基础设施应该考虑到应用管理的一些特性。workload type类型可以是“长期运行的服务”或“一次性任务”。

trait: 定义一个组件所需的运维策略与配置，例如环境变量、Ingress、AutoScaler、Volume 等。

Service: service 不同于k8s中的对象service,在这里定义了一个服务在Kubernetes中运行应用程序所需的运行时配置（即workload type，trait）。service是KubeVela中基本可部署单元的描述符

Application: application在k8s中时应用的集合，描述了开发人员需要定义的内容，application由KubeVela中的Appfile（vela.yaml默认命名）定义。

安装

要求：

Kubernetes集群> = v1.15.0
安装并配置了kubectl

下载kubevela

通过脚本下载

1	curl -fsSl https://kubevela.io/install.sh \| bash

从github下载

vela从发行页面下载最新的二进制文件。
解压缩vela二进制文件并将其添加$PATH到入门中。

1	$ sudo mv ./vela /usr/local/bin/vela

初始化kubevela

执行vela install安装KubeVela服务器组件及其依赖组件。

将安装以下依赖项组件以及Vela服务器组件：

Prometheus Stack
Cert-manager
Flagger
KEDA

注意：如果k8s集群的monitoring这个namespace下已经安装了Prometheus-operator，则kubevela安装的 prometheus stack 会造成冲突。

配置保存在“ vela-system / vela-config”中的ConfigMap中

[root@master-01 kubevela]# vela install
- Installing Vela Core Chart:
install chart vela-core, version 0.1.0, desc : A Helm chart for Kube Vela core, contains 35 file
Successfully installed the chart, status: deployed, last deployed time = 2020-12-03 11:06:34.3800069 +0800 CST m=+6.951945903                                 Automatically discover capabilities successfully ✅ Add(8) Update(0) Delete(0)

TYPE            CATEGORY        DESCRIPTION                                                                   
+task           workload        One-off task to run a piece of code or script to completion                   
+webservice     workload        Long-running scalable service with stable endpoint to receive external traffic
+worker         workload        Long-running scalable backend worker without network endpoint                 
+autoscale      trait           Automatically scale the app following certain triggers or metrics             
+metrics        trait           Configure metrics targets to be monitored for the app                         
+rollout        trait           Configure canary deployment strategy to release the app                       
+route          trait           Configure route policy to the app                                             
+scaler         trait           Manually scale the app                                                        

- Finished successfully.
[root@master-01 kubevela]# kubectl get pod -n vela-system 
NAME                                 READY   STATUS    RESTARTS   AGE
flagger-7846864bbf-m6wxt             1/1     Running   0          50s
kubevela-vela-core-f8b987775-mdjqm   0/1     Running   0          65s
[root@master-01 kubevela]# kubectl get pod -n cert-manager 
NAME                                       READY   STATUS    RESTARTS   AGE
cert-manager-79c5f9946-gx9s9               1/1     Running   0          70s
cert-manager-cainjector-76c9d55b6f-f478g   1/1     Running   0          70s
cert-manager-webhook-6d4c5c44bb-bw9vh      1/1     Running   0          70s
[root@master-01 kubevela]# kubectl get pod -n keda
NAME                                               READY   STATUS    RESTARTS   AGE
keda-operator-566d494bf-mqpn8                      0/1     Running   0          68s
keda-operator-metrics-apiserver-698865dc8b-fg4gn   1/1     Running   0          68s

卸载

执行：

1 2	$ helm uninstall -n vela-system kubevela $ rm -r ~/.vela

这将卸载KubeVela服务器组件及其依赖组件。这还将清理本地CLI缓存。

然后清理CRD（默认情况下，不会通过chart删除CRD）：

$ kubectl delete crd \
  applicationconfigurations.core.oam.dev \
  applicationdeployments.core.oam.dev \
  autoscalers.standard.oam.dev \
  certificaterequests.cert-manager.io \
  certificates.cert-manager.io \
  challenges.acme.cert-manager.io \
  clusterissuers.cert-manager.io \
  components.core.oam.dev \
  containerizedworkloads.core.oam.dev \
  healthscopes.core.oam.dev \
  issuers.cert-manager.io \
  manualscalertraits.core.oam.dev \
  metricstraits.standard.oam.dev \
  orders.acme.cert-manager.io \
  podspecworkloads.standard.oam.dev \
  routes.standard.oam.dev \
  scopedefinitions.core.oam.dev \
  servicemonitors.monitoring.coreos.com \
  traitdefinitions.core.oam.dev \
  workloaddefinitions.core.oam.dev

使用vela 部署一个服务

下载官方提供的示例文件

1 2	$ git clone https://github.com/oam-dev/kubevela.git $ cd kubevela/docs/examples/testapp

该示例包含NodeJS应用程序代码，用于构建应用程序的Dockerfile。

注意：要将image: misterli/testapp:v1用户替换为自己用户，以便进行推送。

[root@master-01 testapp]# ls 
Dockerfile  package.json  server.js  vela.yaml
[root@master-01 testapp]# cat vela.yaml 
name: testapp

services:
  express-server:
    # this image will be used in both build and deploy steps
    image: misterli/testapp:v1

    build:
      # Here more runtime specific build templates will be supported, like NodeJS, Go, Python, Ruby.
      docker:
        file: Dockerfile
        context: .

      # Uncomment the following to push to local kind cluster
      # push:
      #   local: kind

    # type: webservice (default) | worker | task

    cmd: ["node", "server.js"]
    port: 8080

    # scaler:
      # replicas: 1

    # route:
    #   domain: example.com
    #   rules:
    #     - path: /testapp
    #       rewriteTarget: /

    # metrics:
    #   format: "prometheus"
    #   port: 8080
    #   path: "/metrics"
    #   scheme:  "http"
    #   enabled: true

    # autoscale:
    #   min: 1
    #   max: 4
    #   cron:
    #     startAt:  "14:00"
    #     duration: "2h"
    #     days:     "Monday, Thursday"
    #     replicas: 2
    #     timezone: "America/Los_Angeles"

  # pi:
  #   image: perl 
  #   cmd: ["perl",  "-Mbignum=bpi", "-wle", "print bpi(2000)"]

部署

[root@master-01 testapp]# vela  up 
Parsing vela.yaml ...
Loading templates ...

Building service (express-server)...
Sending build context to Docker daemon   7.68kB
Step 1/10 : FROM mhart/alpine-node:12
12: Pulling from mhart/alpine-node
31603596830f: Pulling fs layer
a1768851dab2: Pulling fs layer
31603596830f: Verifying Checksum
31603596830f: Download complete
31603596830f: Pull complete
a1768851dab2: Verifying Checksum
a1768851dab2: Download complete
a1768851dab2: Pull complete
Digest: sha256:31eebb77c7e3878c45419a69e5e7dddd376d685e064279e024e488076d97c7e4
Status: Downloaded newer image for mhart/alpine-node:12
 ---> b13e0277346d
Step 2/10 : WORKDIR /app
 ---> Running in ab10b920fb85
Removing intermediate container ab10b920fb85
 ---> 9f6c8afc0ac4
Step 3/10 : COPY package.json ./
 ---> a4432016a818
Step 4/10 : RUN npm install
 ---> Running in c13d25b9a074
npm notice created a lockfile as package-lock.json. You should commit this file.
npm WARN docker_web_app@1.0.0 No repository field.
npm WARN docker_web_app@1.0.0 No license field.

added 50 packages from 37 contributors and audited 50 packages in 6.037s
found 0 vulnerabilities

Removing intermediate container c13d25b9a074
 ---> ba5e090aa522
Step 5/10 : RUN npm ci --prod
 ---> Running in f0fa46706fdc
npm WARN prepare removing existing node_modules/ before installation
added 50 packages in 0.3s
Removing intermediate container f0fa46706fdc
 ---> a9345a48a79a
Step 6/10 : FROM mhart/alpine-node:slim-12
slim-12: Pulling from mhart/alpine-node
31603596830f: Already exists
de802a068b6a: Pulling fs layer
de802a068b6a: Verifying Checksum
de802a068b6a: Download complete
de802a068b6a: Pull complete
Digest: sha256:12e59927fda21237348acf1a229ad09cf37fb232d251c3e54e1dac3ddac6feeb
Status: Downloaded newer image for mhart/alpine-node:slim-12
 ---> 6d25d4327eff
Step 7/10 : WORKDIR /app
 ---> Running in d541b38c1823
Removing intermediate container d541b38c1823
 ---> 1e0777fd03d8
Step 8/10 : COPY --from=0 /app .
 ---> abe26ca579ed
Step 9/10 : COPY . .
 ---> 6e9f13fd2777
Step 10/10 : CMD ["node", "server.js"]
 ---> Running in e2a66724e4f1
Removing intermediate container e2a66724e4f1
 ---> 533e1502cb2c
Successfully built 533e1502cb2c
Successfully tagged misterli/testapp:v1
pushing image (misterli/testapp:v1)...
The push refers to repository [docker.io/misterli/testapp]
c84892b4351c: Preparing
fac1f8a2295d: Preparing
5d57bb81c0cc: Preparing
2864da400028: Preparing
89ae5c4ee501: Preparing
89ae5c4ee501: Mounted from mhart/alpine-node
2864da400028: Mounted from mhart/alpine-node
5d57bb81c0cc: Pushed
c84892b4351c: Pushed
fac1f8a2295d: Pushed
v1: digest: sha256:6ac7865710892ddd57c0604d02560f1dd9bbf007b23fbacfa45fdbf718a41669 size: 1365

Rendering configs for service (express-server)...
Writing deploy config to (.vela/deploy.yaml)

Applying deploy configs ...
Checking if app has been deployed...
App has not been deployed, creating a new deployment...
✅ App has been deployed ???
    Port forward: vela port-forward testapp
             SSH: vela exec testapp
         Logging: vela logs testapp
      App status: vela status testapp
  Service status: vela status testapp --svc express-server

[root@master-01 testapp]# vela status testapp 
About:

  Name:         testapp                                
  Namespace:    default                                
  Created at:   2020-12-03 11:17:10.202380171 +0800 CST
  Updated at:   2020-12-03 11:17:10.202380322 +0800 CST

Services:
                                                                                                                                                                - Name: express-server
    Type: webservice
    HEALTHY Ready:1/1 
    Traits:

    Last Deployment:
      Created at: 2020-12-03 11:17:10 +0800 CST
      Updated at: 2020-12-03T11:17:10+08:00
[root@master-01 rabbitmq]# kubectl get pod 
NAME                                  READY   STATUS    RESTARTS   AGE
busybox-deployment-7bfd6d554c-nqrln   1/1     Running   831        6d1h
busybox-deployment-7bfd6d554c-s6lrw   1/1     Running   831        6d1h
check-ecs-price-7cdc97b997-j9w9q      1/1     Running   0          7d
express-server-7b5d47c867-hcq99       1/1     Running   0          88s

我们可以看到执行vela up之后,先进行了docker镜像的build push 操作，然后根据vela.yaml文件内容生成.vela/deploy.yaml文件然后进行apply操作，我们查看一下.vela/deploy.yaml文件内容：

[root@master-01 testapp]# cat .vela/deploy.yaml 
apiVersion: core.oam.dev/v1alpha2
kind: ApplicationConfiguration
metadata:
  creationTimestamp: null
  name: testapp
  namespace: default
spec:
  components:
  - componentName: express-server
    scopes:
    - scopeRef:
        apiVersion: core.oam.dev/v1alpha2
        kind: HealthScope
        name: testapp-default-health
    traits:
    - trait:
        apiVersion: core.oam.dev/v1alpha2
        kind: ManualScalerTrait
        metadata:
          labels:
            trait.oam.dev/type: scaler
        spec:
          replicaCount: 2
status:
  dependency: {}
  observedGeneration: 0

---
apiVersion: core.oam.dev/v1alpha2
kind: Component
metadata:
  creationTimestamp: null
  name: express-server
  namespace: default
spec:
  workload:
    apiVersion: apps/v1
    kind: Deployment
    metadata:
      labels:
        workload.oam.dev/type: webservice
    spec:
      selector:
        matchLabels:
          app.oam.dev/component: express-server
      template:
        metadata:
          labels:
            app.oam.dev/component: express-server
        spec:
          containers:
          - command:
            - node
            - server.js
            image: misterli/testapp:v1
            name: express-server
            ports:
            - containerPort: 8080
status:
  observedGeneration: 0

---
apiVersion: core.oam.dev/v1alpha2
kind: HealthScope
metadata:
  creationTimestamp: null
  name: testapp-default-health
  namespace: default
spec:
  workloadRefs: []
status:
  scopeHealthCondition:
    healthStatus: ""

我们修改vela.yaml文件中，取消如下注释并修改为2

1 2	scaler: replicas: 2

我们在执行vela up ，pod的副本数将变为2个

[root@master-01 rabbitmq]# kubectl get pod 
NAME                                  READY   STATUS    RESTARTS   AGE
busybox-deployment-7bfd6d554c-nqrln   1/1     Running   832        6d1h
busybox-deployment-7bfd6d554c-s6lrw   1/1     Running   832        6d1h
check-ecs-price-7cdc97b997-j9w9q      1/1     Running   0          7d
express-server-7b5d47c867-g4jbh       1/1     Running   0          70s
express-server-7b5d47c867-hcq99       1/1     Running   0          5m33s

注意：我们这里删除服务如果使用kubectl 删除deployment 是不起作用的，vela会自动拉取新的deployment,删除服务我们需要使用vela delete APP_NAME

#错误的删除方式
[root@master-01 testapp]# kubectl get deployments.apps 
NAME              READY   UP-TO-DATE   AVAILABLE   AGE
check-ecs-price   1/1     1            1           8d
express-server    1/1     1            1           3m45s
[root@master-01 testapp]# kubectl delete  deployments.apps  express-server 
deployment.apps "express-server" deleted
[root@master-01 testapp]# kubectl get deployments.apps 
NAME              READY   UP-TO-DATE   AVAILABLE   AGE
check-ecs-price   1/1     1            1           8d
express-server    0/1     1            0           1s

#正确的删除方式
[root@master-01 testapp]# vela ls 
SERVICE         APP     TYPE            TRAITS          STATUS          CREATED-TIME                 
express-server  testapp webservice      metrics,scaler  Deployed        2020-12-03 11:17:10 +0800 CST
[root@master-01 testapp]# vela delete testapp 
Deleting Application "testapp"
delete apps succeed testapp from default
[root@master-01 testapp]# vela ls 
SERVICE APP     TYPE    TRAITS  STATUS  CREATED-TIME
[root@master-01 testapp]# kubectl get deployments.apps 
NAME              READY   UP-TO-DATE   AVAILABLE   AGE
check-ecs-price   1/1     1            1           8d

点击屏末 | 阅读原文 | 即刻学习