一、环境准备

我们紧接上一节的环境，进行下面的操作，如果不清楚的，可以先查看上一篇博文。

滚动更新是一次只更新一小部分副本，成功后，再更新更多的副本，最终完成所有副本的更新。滚动更新的最大的好处是零停机，整个更新过程始终有副本在运行，从而保证了业务的连续性。

二、更新

我们查看一下上一节的配置文件mytest-deploy.yaml。

apiVersion: extensions/v1beta1kind: Deploymentmetadata: ?name: mytestspec: ?replicas: 3 ?template: ???metadata: ?????labels: ???????run: mytest ???spec: ?????containers: ?????- name: mytest ???????image: wangzan18/mytest:v1 ???????ports: ???????- containerPort: 80

我们看到设定的镜像版本为v1，我们查看一下创建的ReplicaSet。

[root@master ~]# kubectl get rs -o wideNAME ??????????????DESIRED ??CURRENT ??READY ??AGE ??CONTAINERS ??IMAGES ???????????????SELECTORmytest-88d46bf99 ??3 ????????3 ????????3 ??????68m ??mytest ??????wangzan18/mytest:v1 ??pod-template-hash=88d46bf99,run=mytest

目前运行的 Pod 由 ReplicaSet mytest-88d46bf99进行控制，我们将配置文件中的 v1 替换为 v2，再次应用。

[root@master ~]# kubectl apply -f mytest-deploy.yaml deployment.extensions/mytest configured[root@master ~]# kubectl get podsNAME ????????????????????READY ??STATUS ???????RESTARTS ??AGEmytest-56c55b4c6-6gjxc ??1/1 ????Running ??????0 ?????????24smytest-56c55b4c6-f5trx ??1/1 ????Running ??????0 ?????????18smytest-56c55b4c6-sh5wd ??1/1 ????Running ??????0 ?????????24smytest-88d46bf99-48f6n ??1/1 ????Terminating ??0 ?????????70mmytest-88d46bf99-mv6cf ??1/1 ????Terminating ??0 ?????????70mmytest-88d46bf99-p9w79 ??1/1 ????Terminating ??0 ?????????70m

[root@master ~]# kubectl get deploy -o wideNAME ????READY ??UP-TO-DATE ??AVAILABLE ??AGE ??CONTAINERS ??IMAGES ???????????????SELECTORmytest ??3/3 ????3 ???????????3 ??????????72m ??mytest ??????wangzan18/mytest:v2 ??run=mytest[root@master ~]# kubectl get rs -o wideNAME ??????????????DESIRED ??CURRENT ??READY ??AGE ??CONTAINERS ??IMAGES ???????????????SELECTORmytest-56c55b4c6 ??3 ????????3 ????????3 ??????59s ??mytest ??????wangzan18/mytest:v2 ??pod-template-hash=56c55b4c6,run=mytestmytest-88d46bf99 ??0 ????????0 ????????0 ??????71m ??mytest ??????wangzan18/mytest:v1 ??pod-template-hash=88d46bf99,run=mytest

我们可以看到 Deployment 的镜像更新为wangzan18/mytest:v2，创建了新的 ReplicaSet mytest-56c55b4c6，并管理了三个新的 Pod。

[root@master ~]# kubectl describe deployment mytestEvents: ?Type ???Reason ????????????Age ??From ??????????????????Message ?---- ???------ ????????????---- ?---- ??????????????????------- ?Normal ?ScalingReplicaSet ?4m9s ?deployment-controller ?Scaled up replica set mytest-56c55b4c6 to 1 ?Normal ?ScalingReplicaSet ?4m9s ?deployment-controller ?Scaled down replica set mytest-88d46bf99 to 2 ?Normal ?ScalingReplicaSet ?4m9s ?deployment-controller ?Scaled up replica set mytest-56c55b4c6 to 2 ?Normal ?ScalingReplicaSet ?4m3s ?deployment-controller ?Scaled down replica set mytest-88d46bf99 to 1 ?Normal ?ScalingReplicaSet ?4m3s ?deployment-controller ?Scaled up replica set mytest-56c55b4c6 to 3 ?Normal ?ScalingReplicaSet ?4m3s ?deployment-controller ?Scaled down replica set mytest-88d46bf99 to 0

每次只更新替换一个 Pod。

三、回滚

kubectl apply 每次更新应用时 Kubernetes 都会记录下当前的配置，保存为一个 revision（版次），这样就可以回滚到某个特定 revision。

默认配置下，Kubernetes 只会保留最近的几个 revision，可以在 Deployment 配置文件中通过 revisionHistoryLimit 属性增加 revision 数量。

通过命令kubectl rollout undo我们可以回滚到上一个版本。

[root@master ~]# kubectl rollout undo deploy mytestdeployment.extensions/mytest rolled back

[root@master ~]# kubectl get rs -o wideNAME ??????????????DESIRED ??CURRENT ??READY ??AGE ??CONTAINERS ??IMAGES ???????????????SELECTORmytest-56c55b4c6 ??0 ????????0 ????????0 ??????12m ??mytest ??????wangzan18/mytest:v2 ??pod-template-hash=56c55b4c6,run=mytestmytest-88d46bf99 ??3 ????????3 ????????3 ??????82m ??mytest ??????wangzan18/mytest:v1 ??pod-template-hash=88d46bf99,run=mytest

可以看到三个 Pod 重新由镜像为wangzan18/mytest:v1的ReplicaSet进行接管。

Kubernetes Pod应用的滚动更新(八)

原文地址：https://www.cnblogs.com/wzlinux/p/10159314.html