使用Deployment声明式升级应用

╰半夏微凉° 2023-08-17 16:42 35阅读 0赞

> 镜像来源：https://github.com/luksa/kubernetes-in-action/blob/master/Chapter09

# 一、应用滚动升级 #

1.发布应用、服务

[root@m-xxx ~]# kubectl create -f kubia-deployment-v1.yaml --record # 该选项是记录历史版本号
    ---
    apiVersion: apps/v1beta1
    kind: Deployment
    metadata:
      name: kubia
    spec:
      replicas: 3
      template:
        metadata:
          name: kubia
          labels:
            nginx: kubia
        spec:
          containers:
          - image: kubia:v1
            name: nodejs
            imagePullPolicy: IfNotPresent       
    ---
    apiVersion: v1
    kind: Service
    metadata:
      name: kubia
    spec:
      type: NodePort
      selector:
        app: kubia
      ports:
      - port: 80
        targetPort: 8080

2.展示Deployment滚动过程中的状态

[root@m-xxx ~]# kubectl rollout status deployment kubia 
    deployment "kubia" successfully rolled out
    [root@m-xxx~]# kubectl get deployment 
    NAME           READY   UP-TO-DATE   AVAILABLE   AGE
    kubia          3/3     3            3           4m9s
    [root@m-xxx ~]# kubectl get rs
    NAME                      DESIRED   CURRENT   READY   AGE
    kubia-5957c597b9          3         3         3       4m41s
    [root@m-xxx ~]# kubectl get pods
    NAME                            READY   STATUS    RESTARTS   AGE
    kubia-5957c597b9-5zf7r          1/1     Running   0          5m3s
    kubia-5957c597b9-dvh76          1/1     Running   0          5m3s
    kubia-5957c597b9-h5j5p          1/1     Running   0          5m3s

3.触发滚动升级

[root@m-xxx ~]# kubectl set image deployment kubia nodejs=kubia:v2
    deployment.extensions/kubia image updated

4.列出新旧rs，其过程是deployment控制新旧rs扩缩容，直至到0：期望值

[root@m-xxx ~]# kubectl get rs
    NAME                      DESIRED   CURRENT   READY   AGE
    kubia-5957c597b9          0         0         0       55m
    kubia-88f66b894           3         3         3       6m7s

# 二、应用回滚 #

v3版本存在一个bug，程序只能正确请求前面四个请求，镜像源看文章头部引用。  
1.部署v3版本应用

[root@m-xxx ~]# kubectl set image deployment kubia nodejs=kubia:v3
    [root@m-xxx ~]# kubectl rollout status deployment kubia
    Waiting for deployment "kubia" rollout to finish: 1 old replicas are pending termination...
    Waiting for deployment "kubia" rollout to finish: 1 old replicas are pending termination...
    deployment "kubia" successfully rolled out

2.访问观察

[root@xxx ~]# while true;do curl 任意节点ip:30275;sleep 0.5;done
    This is v3 running in pod kubia-bf846cc6c-8xb4m
    ... ...
    This is v3 running in pod kubia-bf846cc6c-9rkhr
    Some internal error has occurred! This is pod kubia-bf846cc6c-8xb4m
    ... ...
    Some internal error has occurred! This is pod kubia-bf846cc6c-9rkhr

3.回滚到上一版本，也可在滚动升级过程中执行，会直接停止滚动升级

[root@m-xxx ~]# kubectl rollout undo deployment kubia # 回滚到上一版本
    deployment.extensions/kubia rolled back
    [root@m-xxx ~]# kubectl rollout status deployment kubia
    Waiting for deployment "kubia" rollout to finish: 1 out of 3 new replicas have been updated...
    Waiting for deployment "kubia" rollout to finish: 1 out of 3 new replicas have been updated...
    Waiting for deployment "kubia" rollout to finish: 1 out of 3 new replicas have been updated...
    Waiting for deployment "kubia" rollout to finish: 2 out of 3 new replicas have been updated...
    Waiting for deployment "kubia" rollout to finish: 2 out of 3 new replicas have been updated...
    Waiting for deployment "kubia" rollout to finish: 2 out of 3 new replicas have been updated...
    Waiting for deployment "kubia" rollout to finish: 2 out of 3 new replicas have been updated...
    Waiting for deployment "kubia" rollout to finish: 1 old replicas are pending termination...
    Waiting for deployment "kubia" rollout to finish: 1 old replicas are pending termination...
    Waiting for deployment "kubia" rollout to finish: 1 old replicas are pending termination...
    deployment "kubia" successfully rolled out

4.显示滚动升级历史

[root@m-xxx ~]# kubectl rollout history deployment kubia
    deployment.extensions/kubia 
    REVISION  CHANGE-CAUSE
    1         kubectl create --filename=kubia-deployment-v1.yaml --record=true
    3         kubectl create --filename=kubia-deployment-v1.yaml --record=true
    4         kubectl create --filename=kubia-deployment-v1.yaml --record=true

5.回滚特定版本

[root@m-xxx ~]# kubectl rollout undo deployment kubia --to-revision=1

6.暂停滚动升级(金丝雀发布)

[root@m-xxx ~]# kubectl set image deployment kubia nodejs=kubia:v4
    deployment.extensions/kubia image updated
    [root@m-xxx ~]# kubectl rollout pause deployment kubia
    deployment.extensions/kubia paused

7.恢复滚动升级

[root@m-xxx~]# kubectl rollout resume deployment kubia

# 三、阻止出错版本的滚动升级 #

使用minReadySeconds和就绪探针自动避免错误版本升级  
1.使用apply更新Deployment

[root@m-xxx ~]# kubectl apply -f kubia-deployment-v3-with-readinesscheck.yaml 
    ---
    kind: Deployment
    apiVersion: apps/v1beta1
    metadata:
      name: kubia
    spec:
      replicas: 3
      template:
        metadata:
          name: kubia
          labels:
            nginx: kubia
        spec:
          containers:
          - name: nodejs
            image: kubia:v3
            imagePullPolicy: IfNotPresent
            readinessProbe:
              periodSeconds: 1  # 每一秒钟执行一次
              httpGet:
                path: /
                port: 8080
      strategy:
        type: RollingUpdate  # 滚动升级策略，另外一个是Recreate(全部杀死再升级)
        rollingUpdate:  # 控制一次替换多少个pod
          maxUnavailable: 0  # 升级过程中，至少replicas-0可用，不可理解为最多0个不可用
          maxSurge: 1  # 升级过程中，副本实例可超1
      minReadySeconds: 30 # 新建pod正确运行30秒后才视为可用
      revisionHistoryLimit: 2 #保存2个历史版本号
      progressDeadlineSeconds: 600 # 默认10分钟滚动升级未完成，则自动取消滚动升级 
    [root@m-xxx ~]# kubectl rollout status deployment kubia 
    Waiting for deployment "kubia" rollout to finish: 0 of 3 updated replicas are available...

2.取消出错版本的滚动升级

[root@m-xxx ~]# kubectl rollout undo deployment kubia