K8S-POD的升级与回滚

最新推荐文章于 2024-08-20 13:40:02 发布

〰振振 ༽

最新推荐文章于 2024-08-20 13:40:02 发布

阅读量289

点赞数

分类专栏： K8S 文章标签： kubernetes docker 运维

本文链接：https://blog.csdn.net/weixin_44532734/article/details/128471380

版权

K8S 专栏收录该内容

3 篇文章 0 订阅

订阅专栏

Pod的升级与回滚

一、 Deployment的升级与回滚

Deployment的升级

①kubectl set image deployment/nginx-deployment nginx=nginx:1.9.1

②kubectl edit deployment nginx-deployment进入编辑版本

镜像名一旦发生修改，则将触发系统完成Deployment所有运行Pod的滚动升级操作，可以使用kubectl rollout status命令查看Deployment的更新过程

[root@k8s-master k8s]# kubectl set image  deployment/nginx-deployment nginx=nginx:1.9.9
deployment.apps/nginx-deployment image updated
[root@k8s-master k8s]# kubectl rollout status deployment/nginx-deployment
Waiting for deployment "nginx-deployment" rollout to finish: 1 out of 2 new replicas have been updated...
Waiting for deployment "nginx-deployment" rollout to finish: 1 out of 2 new replicas have been updated...
Waiting for deployment "nginx-deployment" rollout to finish: 1 out of 2 new replicas have been updated...
Waiting for deployment "nginx-deployment" rollout to finish: 1 old replicas are pending termination...
Waiting for deployment "nginx-deployment" rollout to finish: 1 old replicas are pending termination...
deployment "nginx-deployment" successfully rolled out

[root@k8s-master k8s]# kubectl get pods|grep nginx|grep -i running
nginx-deployment-595cf8f5f8-52hqs   1/1     Running       0          2m1s
nginx-deployment-595cf8f5f8-vrc2s   1/1     Running       0          2m30s
[root@k8s-master k8s]# kubectl describe pod/nginx-deployment-595cf8f5f8-52hqs|grep image
  Normal  Pulled     2m8s  kubelet            Container image "nginx:1.9.9" already present on machine

运行过程示意图

[root@k8s-master k8s]# kubectl describe deployment nginx-deployment
Name:                   nginx-deployment
Namespace:              default
CreationTimestamp:      Wed, 14 Dec 2022 11:27:11 -0800
Labels:                 <none>
Annotations:            deployment.kubernetes.io/revision: 3
Selector:               app=nginx
Replicas:               2 desired | 2 updated | 2 total | 2 available | 0 unavailable
StrategyType:           RollingUpdate
MinReadySeconds:        0
RollingUpdateStrategy:  25% max unavailable, 25% max surge
Pod Template:
  Labels:  app=nginx
  Containers:
   nginx:
    Image:        nginx:1.9.9
    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-595cf8f5f8 (2/2 replicas created)
Events:
  Type    Reason             Age   From                   Message
  ----    ------             ----  ----                   -------
  Normal  ScalingReplicaSet  23m   deployment-controller  Scaled up replica set nginx-deployment-69c44dfb78 to 1
  Normal  ScalingReplicaSet  22m   deployment-controller  Scaled down replica set nginx-deployment-5d59d67564 to 1
  Normal  ScalingReplicaSet  22m   deployment-controller  Scaled up replica set nginx-deployment-69c44dfb78 to 2
  Normal  ScalingReplicaSet  22m   deployment-controller  Scaled down replica set nginx-deployment-5d59d67564 to 0
  Normal  ScalingReplicaSet  17m   deployment-controller  Scaled up replica set nginx-deployment-595cf8f5f8 to 1
  Normal  ScalingReplicaSet  17m   deployment-controller  Scaled down replica set nginx-deployment-69c44dfb78 to 1
  Normal  ScalingReplicaSet  17m   deployment-controller  Scaled up replica set nginx-deployment-595cf8f5f8 to 2
  Normal  ScalingReplicaSet  17m   deployment-controller  Scaled down replica set nginx-deployment-69c44dfb78 to 0 # 这里经过2次升级。

通过查看ReplicaSet信息也可以看出：

[root@k8s-master k8s]# kubectl get rs
NAME                          DESIRED   CURRENT   READY   AGE
nginx-deployment-595cf8f5f8   2         2         2       22m
nginx-deployment-5d59d67564   0         0         0       3d21h
nginx-deployment-69c44dfb78   0         0         0       28m

之所以不删除，是为了后面回滚。

Deployment的回滚

查看Deployment的历史部署记录

[root@k8s-master k8s]# kubectl rollout history deployment/nginx-deployment
deployment.apps/nginx-deployment
REVISION  CHANGE-CAUSE
1         <none>
2         <none>
3         <none>

当部署升级的时候，写错版本即一个不存在镜像时，就会一直拉取不到，如下

[root@k8s-master k8s]# kubectl set image  deployment/nginx-deployment nginx=nginx:1.9.99
deployment.apps/nginx-deployment image updated
[root@k8s-master k8s]# kubectl get rs
NAME                          DESIRED   CURRENT   READY   AGE
nginx-deployment-595cf8f5f8   2         2         2       27m
nginx-deployment-5d59d67564   0         0         0       3d21h
nginx-deployment-69c44dfb78   0         0         0       33m
nginx-deployment-74cdcd844f   1         1         0       5s
[root@k8s-master k8s]# kubectl get po|grep nginx
nginx-deployment-595cf8f5f8-52hqs   1/1     Running        0          27m
nginx-deployment-595cf8f5f8-vrc2s   1/1     Running        0          27m
nginx-deployment-5d59d67564-ccxv7   1/1     Terminating    0          3d21h
nginx-deployment-5d59d67564-sfqjc   1/1     Terminating    0          3d21h
nginx-deployment-74cdcd844f-mmc4s   0/1     ErrImagePull   0          27s

回滚

[root@k8s-master k8s]# kubectl rollout undo deployment/nginx-deployment
deployment.apps/nginx-deployment rolled back
[root@k8s-master k8s]# kubectl get rs
NAME                          DESIRED   CURRENT   READY   AGE
nginx-deployment-595cf8f5f8   2         2         2       31m
nginx-deployment-5d59d67564   0         0         0       3d21h
nginx-deployment-69c44dfb78   0         0         0       37m
nginx-deployment-74cdcd844f   0         0         0       4m5s
[root@k8s-master k8s]# kubectl describe pod nginx-deployment-595cf8f5f8-52hqs|grep -i image
    Image:          nginx:1.9.9  # 版本回到了1.9.9版本
    Image ID:       docker-pullable://nginx@sha256:c56e8eaec9ff118688c857e263f744039f5ba9fe3b30efcaeef5342f27534a01
  Normal  Pulled     32m   kubelet            Container image "nginx:1.9.9" already present on machine

对于复杂的更新操作，可以先配置暂停更新，然后一次性进行更新操作

kubectl rollout pause deployment/nginx-deployment  # 暂停Deployment的更新操作
kubectl set image deploy nginx-deployment nginx=nginx:1.9.1   # 修改镜像信息
kubectl set resources deployment nginx-deployment -c=nginx --limits=cpu=200m,memory=512Mi  # 修改资源限制
kubectl rollout resume deploy nginx-deployment    # 恢复Deployment的更新
kubectl describe deploy nginx-deployment   # 检查Deployment的更新情况

DaemonSet与StatefulSet的更新策略

DaemonSet

1.6版本开始支持滚动更新。如如下两种策略

OnDelete ：默认升级策略，在创建好新的DaemonSet配置之后，新的Pod不会被自动创建，用户需要手动删除旧版本的Pod，才出发新建操作。

RollingUpdate：旧版本的POD 将被自动杀掉，然后自动创建新版的DaemonSet Pod。与Deployment 不同为不支持查看和管理DaemonSet的更新记录；回滚操作是通过再次提交旧版本配置而不是 rollback命令实现

StatefulSet

OnDelete ：默认升级策略，在创建好新的StatefulSetSet配置之后，新的Pod不会被自动创建，用户需要手动删除旧版本的Pod，才出发新建操作。

RollingUpdate： StatefulSet 控制器将删除并重新创建 StatefulSet 中的每个 Pod。它将按照与 Pod 终止相同的顺序进行（从最大的序数到最小的），依次更新每个 Pod。Kubernetes 控制平面会等到更新的 Pod 运行并准备好，然后再更新其前任。如果您已设置.spec.minReadySeconds（请参阅“最小就绪秒数”），则控制平面会在 Pod 准备就绪后额外等待该时间，然后再继续。

partitioned：分区滚动更新。如果指定了分区，则在更新 StatefulSet 时，将更新序数大于或等于该分区的所有 Pod .spec.template。所有序号小于分区的 Pod 都不会更新，即使删除了，也会在之前的版本中重新创建。如果 StatefulSet 的.spec.updateStrategy.rollingUpdate.partition大于其.spec.replicas，则对其的更新.spec.template将不会传播到其 Pod。在大多数情况下，您不需要使用分区，但如果您想要暂存更新、roll out Canary 或执行分阶段roll out，它们会很有用

回滚更新：一旦更新配置有误，即使修改了配置，k8s也不会继续更新，而是等待异常POD恢复至正常。这时我们需要手动删除异常POD