测试开发技术网站

# 一、版本升级过程的变化原理
编辑pc_deployment.yaml文件，内容如下：
```yaml
apiVersion: v1
kind: Namespace
metadata:
  name: dev

---

apiVersion: apps/v1
kind: Deployment
metadata:
  name: pc-deployment
  namespace: dev
spec:
  replicas: 3
  strategy:
    type: RollingUpdate
    rollingUpdate:
      maxUnavailable: 25%
      maxSurge: 25%
  selector:
    matchLabels:
      app: nginx-pod
  template:
    metadata:
      labels:
        app: nginx-pod
    spec:
      containers:
      - name: nginx
        image: nginx:1.17.1
```
使用如下命令创建资源，注意加--record参数，用于记录创建记录信息
```bash
[root@master pod_controller]# kubectl apply -f pc_deployment.yaml --record=true
namespace/dev created
deployment.apps/pc-deployment created
[root@master pod_controller]#
```
然后编辑yaml文件，将镜像版本升级到1.17.2，再次使用上述命令
然后编辑yaml文件，将镜像版本升级到1.17.3，再次使用上述命令
然后编辑yaml文件，将镜像版本升级到1.17.4，再次使用上述命令
此时，通过如下命令可以查看到，有四个ReplicaSet
```bash
[root@master pod_controller]# kubectl get rs -n dev
NAME                       DESIRED   CURRENT   READY   AGE
pc-deployment-5d9c9b97bb   0         0         0       4m9s
pc-deployment-6b575b9c45   3         3         3       47s
pc-deployment-76fd8c7f84   0         0         0       88s
pc-deployment-7c7477c7ff   0         0         0       118s
```
而通过监控ReplicaSet的变化过程如下看可以发现，在每次镜像升级的过程中，其实本质是每次都创建一个新的ReplicaSet，然后依次在新的ReplicaSet上创建pod
```bash
[root@master ~]# kubectl get rs -n dev -w
NAME                       DESIRED   CURRENT   READY   AGE
pc-deployment-5d9c9b97bb   3         3         3       2m7s
pc-deployment-7c7477c7ff   1         0         0       0s
pc-deployment-7c7477c7ff   1         0         0       0s
pc-deployment-7c7477c7ff   1         1         0       0s
pc-deployment-7c7477c7ff   1         1         1       1s
pc-deployment-5d9c9b97bb   2         3         3       2m12s
pc-deployment-5d9c9b97bb   2         3         3       2m12s
pc-deployment-7c7477c7ff   2         1         1       1s
pc-deployment-5d9c9b97bb   2         2         2       2m12s
pc-deployment-7c7477c7ff   2         1         1       1s
pc-deployment-7c7477c7ff   2         2         1       1s
pc-deployment-7c7477c7ff   2         2         2       2s
pc-deployment-5d9c9b97bb   1         2         2       2m13s
pc-deployment-5d9c9b97bb   1         2         2       2m13s
pc-deployment-7c7477c7ff   3         2         2       2s
pc-deployment-5d9c9b97bb   1         1         1       2m13s
pc-deployment-7c7477c7ff   3         2         2       2s
pc-deployment-7c7477c7ff   3         3         2       2s
pc-deployment-7c7477c7ff   3         3         3       24s
pc-deployment-5d9c9b97bb   0         1         1       2m35s
pc-deployment-5d9c9b97bb   0         1         1       2m35s
pc-deployment-5d9c9b97bb   0         0         0       2m35s
pc-deployment-76fd8c7f84   1         0         0       0s
pc-deployment-76fd8c7f84   1         0         0       0s
pc-deployment-76fd8c7f84   1         1         0       0s
pc-deployment-76fd8c7f84   1         1         1       1s
pc-deployment-7c7477c7ff   2         3         3       31s
pc-deployment-7c7477c7ff   2         3         3       31s
pc-deployment-76fd8c7f84   2         1         1       1s
pc-deployment-76fd8c7f84   2         1         1       1s
pc-deployment-7c7477c7ff   2         2         2       31s
pc-deployment-76fd8c7f84   2         2         1       1s
pc-deployment-76fd8c7f84   2         2         2       2s
pc-deployment-7c7477c7ff   1         2         2       32s
pc-deployment-7c7477c7ff   1         2         2       32s
pc-deployment-76fd8c7f84   3         2         2       2s
pc-deployment-7c7477c7ff   1         1         1       32s
pc-deployment-76fd8c7f84   3         2         2       2s
pc-deployment-76fd8c7f84   3         3         2       2s
pc-deployment-76fd8c7f84   3         3         3       3s
pc-deployment-7c7477c7ff   0         1         1       33s
pc-deployment-7c7477c7ff   0         1         1       33s
pc-deployment-7c7477c7ff   0         0         0       33s
pc-deployment-6b575b9c45   1         0         0       0s
pc-deployment-6b575b9c45   1         0         0       0s
pc-deployment-6b575b9c45   1         1         0       0s
pc-deployment-6b575b9c45   1         1         1       22s
pc-deployment-76fd8c7f84   2         3         3       63s
pc-deployment-6b575b9c45   2         1         1       22s
pc-deployment-76fd8c7f84   2         3         3       63s
pc-deployment-6b575b9c45   2         1         1       22s
pc-deployment-6b575b9c45   2         2         1       22s
pc-deployment-76fd8c7f84   2         2         2       63s
pc-deployment-6b575b9c45   2         2         2       45s
pc-deployment-76fd8c7f84   1         2         2       86s
pc-deployment-6b575b9c45   3         2         2       45s
pc-deployment-76fd8c7f84   1         2         2       86s
pc-deployment-76fd8c7f84   1         1         1       86s
pc-deployment-6b575b9c45   3         2         2       45s
pc-deployment-6b575b9c45   3         3         2       45s
pc-deployment-6b575b9c45   3         3         3       46s
pc-deployment-76fd8c7f84   0         1         1       87s
pc-deployment-76fd8c7f84   0         1         1       87s
pc-deployment-76fd8c7f84   0         0         0       87s
```
# 二、版本回退
## 2.1 版本回退简介
deployment控制器支持版本回退，因为在每次升级的时候deployment都是创建了一个新的ReplicaSet，依次，版本回退就是回退到旧的ReplicaSet上即可

deployment版本回退相关的命令是：kubectl rollout 命令，后跟如下选项
* status      显示当前升级状态
* history     显示升级历史记录
* pause      暂停版本升级过程
* resume    继续已经暂停的版本升级过程
* restart     重启版本升级过程
* undo       回滚到上一级版本（可以使用 --to-revision回滚到指定的版本）

## 2.2 版本回退
如下即查看当前升级状态
```bash
[root@master pod_controller]# kubectl rollout status deployment pc-deployment -n dev
deployment "pc-deployment" successfully rolled out
[root@master pod_controller]#
```
如下命令可以查看到升级历史记录
```bash
[root@master pod_controller]# kubectl rollout history deployment pc-deployment -n dev
deployment.apps/pc-deployment
REVISION  CHANGE-CAUSE
1         kubectl apply --filename=pc_deployment.yaml --record=true
2         kubectl apply --filename=pc_deployment.yaml --record=true
3         kubectl apply --filename=pc_deployment.yaml --record=true
4         kubectl apply --filename=pc_deployment.yaml --record=true

[root@master pod_controller]#
```
通过如下命令即可回退到上个版本
```bash
[root@master pod_controller]# kubectl rollout undo deployment pc-deployment -n dev
deployment.apps/pc-deployment rolled back
[root@master pod_controller]#
```
此时查看历史版本如下，已经没有3了，而是多了一个5
```bash
[root@master pod_controller]# kubectl rollout history deployment pc-deployment -n dev
deployment.apps/pc-deployment
REVISION  CHANGE-CAUSE
1         kubectl apply --filename=pc_deployment.yaml --record=true
2         kubectl apply --filename=pc_deployment.yaml --record=true
3         kubectl apply --filename=pc_deployment.yaml --record=true
4         kubectl apply --filename=pc_deployment.yaml --record=true

[root@master pod_controller]#
```
若直接回退到第一个版本，则使用如下命令即可
```bash
[root@master pod_controller]# kubectl rollout undo deployment pc-deployment --to-revision=1 -n dev
deployment.apps/pc-deployment rolled back
[root@master pod_controller]#
```
此时再次查看历史记录如下，此时1已经没有了，而多出来一个6
```bash
[root@master pod_controller]# kubectl rollout history deployment pc-deployment -n dev
deployment.apps/pc-deployment
REVISION  CHANGE-CAUSE
2         kubectl apply --filename=pc_deployment.yaml --record=true
4         kubectl apply --filename=pc_deployment.yaml --record=true
5         kubectl apply --filename=pc_deployment.yaml --record=true
6         kubectl apply --filename=pc_deployment.yaml --record=true

[root@master pod_controller]#
```

测试开发技术网站

博客