k8s控制器之Deployment第四弹之回滚Deployment

在某些情况下，如:Deployment 不稳定（可能是不断地崩溃）,我们可能需要回滚（rollback）Deployment。默认情况下，kubernetes 将保存 Deployment 的所有更新（rollout）历史。您可以设定 revision history limit 来确定保存的历史版本数量。

当且仅当 Deployment 的 .spec.template 字段被修改时（例如，您修改了容器的镜像），kubernetes 将为其创建一个 Deployment revision（版本）。Deployment 的其他更新（例如：修改 .spec.replicas 字段）将不会创建新的 Deployment reviesion（版本）。

为了让大家可以更好的理解Deployment的回滚机制，我们这里设计了一个小实验来模拟更新错误，从而方便大家更好的理解

模拟更新错误

• 假设你在更新Deployment忘了打一个点，将nginx:1.9.1 写成了 nginx:19.1

kubectl set image deployment.v1.apps/nginx-deployment nginx=nginx:19.1 --record=true

输出结果:

deployment.apps/nginx-deployment image updated

• 该更新将卡住，您可以执行命令 kubectl rollout status deployment.v1.apps/nginx-deployment 检查其状态，

kubectl rollout status deployment.v1.apps/nginx-deployment

输出结果：

Waiting for deployment "nginx-deployment" rollout to finish: 1 out of 3 new replicas have been updated...

• 执行命令 kubectl get rs 您将看到两个旧的 ReplicaSet（nginx-deployment-84df99548d and nginx-deployment-8bf4959b6 ）和一个新的 ReplicaSet (nginx-deployment-6549cf8856)

kubectl get rs

输出结果：

NAME                          DESIRED   CURRENT   READY   AGE
nginx-deployment-6549cf8856   1         1         0       35s
nginx-deployment-84df99548d   0         0         0       21h
nginx-deployment-8bf4959b6    3         3         3       6h11m

• 执行命令 kubectl get pods，您将看到 1 个由新 ReplicaSet 创建的 Pod 卡在抓取 image 的死循环里：

kubectl get pods

输出结果：

NAME                                READY   STATUS             RESTARTS   AGE
nginx-deployment-6549cf8856-j88tr   0/1     ImagePullBackOff   0          79s
nginx-deployment-8bf4959b6-pxf77    1/1     Running            0          6h11m
nginx-deployment-8bf4959b6-w4r85    1/1     Running            0          6h11m
nginx-deployment-8bf4959b6-w8lgr    1/1     Running            0          6h11m

Deployment Controller 会自动停止有问题的更新（rollout），不会继续 scale up 新的 ReplicaSet。maxUnavailable 参数指定了最多会有几个 Pod 副本卡住，该参数的默认值是 25%。

• 执行命令 kubectl describe deployment 查看 Deployment 的详情

  kubectl describe deployment
  

输出结果：

Name:                   nginx-deployment
Namespace:              default
CreationTimestamp:      Mon, 22 May 2023 19:51:32 +0800
Labels:                 app=nginx
Annotations:            deployment.kubernetes.io/revision: 3
                        kubernetes.io/change-cause: kubectl set image deployment.v1.apps/nginx-deployment nginx=nginx:19.1 --record=true
Selector:               app=nginx
Replicas:               3 desired | 1 updated | 4 total | 3 available | 1 unavailable
StrategyType:           RollingUpdate
MinReadySeconds:        0
RollingUpdateStrategy:  25% max unavailable, 25% max surge
Pod Template:
  Labels:  app=nginx
  Containers:
   nginx:
    Image:        nginx:19.1
    Port:         80/TCP
    Host Port:    0/TCP
    Environment:  <none>
    Mounts:       <none>
  Volumes:        <none>
Conditions:
  Type           Status  Reason
  ----           ------  ------
  Available      True    MinimumReplicasAvailable
  Progressing    True    ReplicaSetUpdated
OldReplicaSets:  nginx-deployment-8bf4959b6 (3/3 replicas created)
NewReplicaSet:   nginx-deployment-6549cf8856 (1/1 replicas created)
Events:
  Type    Reason             Age   From                   Message
  ----    ------             ----  ----                   -------
  Normal  ScalingReplicaSet  2m4s  deployment-controller  Scaled up replica set nginx-deployment-6549cf8856 to 1

这时一个更新错误便被我们模拟出来了。

下面我便带大家来逐步解决这个问题

检查Deployment 的更新历史

1. 检查 Deployment 的历史版本

kubectl rollout history deployment.v1.apps/nginx-deployment

输出结果：

deployment.apps/nginx-deployment 
REVISION  CHANGE-CAUSE
1         kubectl apply --filename=nginx-deployment.yaml --record=true
2         kubectl deployment.apps/nginx-deployment set image deployment.v1.apps/nginx-deployment nginx=nginx:1.9.1 --record=true
3         kubectl set image deployment.v1.apps/nginx-deployment nginx=nginx:19.1 --record=true

CHANGE-CAUSE 是该 revision（版本）创建时从 Deployment 的 annotation kubernetes.io/change-cause 拷贝而来。

可以通过如下方式制定 CHANGE-CAUSE 信息：

• 为 Deployment 增加注解，kubectl annotate deployment.v1.apps/nginx-deployment kubernetes.io/change-cause="image updated to 1.9.1"
• 执行 kubectl apply 命令时，增加 --record 选项
• 手动编辑 Deployment 的 .metadata.annotation 信息

2. 查看 revision（版本）的详细信息

   kubectl rollout history deployment.v1.apps/nginx-deployment --revision=2
   

   输出内容：

   deployment.apps/nginx-deployment with revision #2
   Pod Template:
     Labels:	app=nginx
   	pod-template-hash=8bf4959b6
     Annotations:	kubernetes.io/change-cause:
   	  kubectl deployment.apps/nginx-deployment set image deployment.v1.apps/nginx-deployment nginx=nginx:1.9.1 --record=true
     Containers:
      nginx:
       Image:	nginx:1.9.1
       Port:	80/TCP
       Host Port:	0/TCP
       Environment:	<none>
       Mounts:	<none>
     Volumes:	<none>
   

回滚到前一个 revision（版本）

从上面可以得知前一个版本(v2)是可用版本,那么我们就将其回滚到前一个版本（v2）

1、将当前版本回滚到前一个版本。

kubectl rollout undo deployment.v1.apps/nginx-deployment

输出结果：

deployment.apps/nginx-deployment rolled back

或者，您也可以使用 --to-revision 选项回滚到前面的某一个指定版本

kubectl rollout undo deployment.v1.apps/nginx-deployment --to-revision=2

输出结果

deployment.apps/nginx-deployment rolled back

此时，Deployment 已经被回滚到前一个稳定版本。您可以看到 Deployment Controller 为该 Deployment 产生了 DeploymentRollback event。

2. 检查该回滚是否成功，Deployment 是否按预期的运行

   kubectl get deployment nginx-deployment
   

输出结果：

NAME               READY   UP-TO-DATE   AVAILABLE   AGE
nginx-deployment   3/3     3            3           22h

3. 查看 Deployment 的详情

   kubectl describe deployment nginx-deployment
   

   输出结果：

   Name:                   nginx-deployment
   Namespace:              default
   CreationTimestamp:      Mon, 22 May 2023 19:51:32 +0800
   Labels:                 app=nginx
   Annotations:            deployment.kubernetes.io/revision: 6
                           kubernetes.io/change-cause:
                             kubectl deployment.apps/nginx-deployment set image deployment.v1.apps/nginx-deployment nginx=nginx:1.9.1 --record=true
   Selector:               app=nginx
   Replicas:               3 desired | 3 updated | 3 total | 3 available | 0 unavailable
   StrategyType:           RollingUpdate
   MinReadySeconds:        0
   RollingUpdateStrategy:  25% max unavailable, 25% max surge
   Pod Template:
     Labels:  app=nginx
     Containers:
      nginx:
       Image:        nginx:1.9.1
       Port:         80/TCP
       Host Port:    0/TCP
       Environment:  <none>
       Mounts:       <none>
     Volumes:        <none>
   Conditions:
     Type           Status  Reason
     ----           ------  ------
     Available      True    MinimumReplicasAvailable
     Progressing    True    NewReplicaSetAvailable
   OldReplicaSets:  <none>
   NewReplicaSet:   nginx-deployment-8bf4959b6 (3/3 replicas created)
   Events:
     Type    Reason             Age   From                   Message
     ----    ------             ----  ----                   -------
     Normal  ScalingReplicaSet  51m   deployment-controller  Scaled up replica set nginx-deployment-6549cf8856 to 1
     Normal  ScalingReplicaSet  108s  deployment-controller  Scaled down replica set nginx-deployment-6549cf8856 to 0
     Normal  ScalingReplicaSet  23s   deployment-controller  Scaled up replica set nginx-deployment-84df99548d to 1
     Normal  ScalingReplicaSet  15s   deployment-controller  Scaled down replica set nginx-deployment-84df99548d to 0