k8s-pod版本更新
pod版本更新⭐️⭐️
在实际应用中，升级是一个常见的场景，Deployment能够很方便的支撑应用升级。
Deployment可以设置不同的升级策略，有如下两种。
RollingUpdate：滚动升级，即逐步创建新Pod再删除旧Pod，为默认策略。
Recreate：替换升级，即先把当前Pod删掉再重新创建Pod。
Deployment的升级可以是声明式的，也就是说只需要修改Deployment的YAML定义即可，
比如使用kubectl edit命令将上面Deployment中的镜像修改为nginx:alpine
修改完成后再查询ReplicaSet和Pod，发现创建了一个新的ReplicaSet，Pod也重新创建了。
# 注意，是修改Deployment控制器，实现对pod的更新
[root@k8s-master-10 ~]#kubectl edit deployments.apps nginx-deployment
34 spec:
35 containers:
36 - image: nginx:latest
deployment和RS和pod关系
查看deployment更新事件
[root@k8s-master-10 ~]#kubectl -n yuchaoit describe deployments.apps nginx-deployment
# 创建了新的RS控制器
NewReplicaSet: nginx-deployment-5b78bc88d4 (2/2 replicas created)
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal ScalingReplicaSet 34m deployment-controller Scaled up replica set nginx-deployment-64bcc7b77f to 2
Normal ScalingReplicaSet 4m53s deployment-controller Scaled up replica set nginx-deployment-5b78bc88d4 to 1
Normal ScalingReplicaSet 4m32s deployment-controller Scaled down replica set nginx-deployment-64bcc7b77f to 1
Normal ScalingReplicaSet 4m32s deployment-controller Scaled up replica set nginx-deployment-5b78bc88d4 to 2
Normal ScalingReplicaSet 4m12s deployment-controller Scaled down replica set nginx-deployment-64bcc7b77f to 0
查看rs更新事件
# 旧的RS，已经不负责pod管理
[root@k8s-master-10 ~]#kubectl -n yuchaoit describe rs nginx-deployment-64bcc7b77f
# 新的RS，创建新的pod
[root@k8s-master-10 ~]#kubectl -n yuchaoit describe rs nginx-deployment-5b78bc88d4
查看pod更新事件
此时的pod，也是基于新的RS控制器创建了
[root@k8s-master-10 ~]#kubectl -n yuchaoit describe pod nginx-deployment-5b78bc88d4-
# pod和RS控制器关系
Controlled By: ReplicaSet/nginx-deployment-5b78bc88d4
deployment更新镜像原理图
默认是滚动更新，逐步创建新pod、然后删除旧pod.
# 命令
[root@k8s-master-10 ~]#kubectl -n yuchaoit describe deployments.apps nginx-deployment
# 滚动更新事件
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal ScalingReplicaSet 42m deployment-controller Scaled up replica set nginx-deployment-64bcc7b77f to 2
Normal ScalingReplicaSet 12m deployment-controller Scaled up replica set nginx-deployment-5b78bc88d4 to 1
Normal ScalingReplicaSet 12m deployment-controller Scaled down replica set nginx-deployment-64bcc7b77f to 1
Normal ScalingReplicaSet 12m deployment-controller Scaled up replica set nginx-deployment-5b78bc88d4 to 2
Normal ScalingReplicaSet 11m deployment-controller Scaled down replica set nginx-deployment-64bcc7b77f to 0
查看滚动更新状态
[root@k8s-master-10 ~]# kubectl -n yuchaoit rollout status deployment nginx-deployment
deployment "nginx-deployment" successfully rolled out
蓝绿更新原理
其他更新镜像命令
1.直接修改yaml文件，修改镜像（镜像就是容器运行的交付标准）
# 语法
# 结尾跟上容器名，修改镜像即可
[root@k8s-master-10 ~]#kubectl set image -f deployment-nginx.yml nginx-containers=nginx:1.21.5
deployment.apps/nginx-deployment image updated
[root@k8s-master-10 ~]#
2.还有可以直接选择控制器，进行镜像修改
[root@k8s-master-10 ~]#kubectl -n yuchaoit set image deployment nginx-deployment nginx-containers=nginx:1.15.0
deployment.apps/nginx-deployment image updated
pod回滚上一个版本
回滚也称为回退，即当发现升级出现问题时，让应用回到老的版本。
Deployment可以非常方便的回滚到老版本。
例如上面升级的新版镜像有问题，可以执行kubectl rollout undo命令进行回滚。
Deployment之所以能如此容易的做到回滚，是因为Deployment是通过ReplicaSet控制Pod的
升级后之前ReplicaSet都一直存在，Deployment回滚做的就是使用之前的ReplicaSet再次把Pod创建出来。
Deployment中保存ReplicaSet的数量可以使用revisionHistoryLimit参数限制，默认值为10。
[root@k8s-master-10 ~]#kubectl -n yuchaoit describe deployments.apps nginx-deployment
[root@k8s-master-10 ~]#kubectl -n yuchaoit rollout undo deployment nginx-deployment
deployment.apps/nginx-deployment rolled back
# 查看回滚上一个控制器的版本
[root@k8s-master-10 ~]#kubectl -n yuchaoit describe deployments.apps nginx-deployment
# 最新pod信息
[root@k8s-master-10 ~]#kubectl -n yuchaoit get po -owide
# 查看RS，很明显，回到上一个版本了
[root@k8s-master-10 ~]#kubectl -n yuchaoit get rs -owide
NAME DESIRED CURRENT READY AGE CONTAINERS IMAGES SELECTOR
nginx-deployment-5b48487b68 0 0 0 6m22s nginx-containers nginx:1.21.5 app=nginx,pod-template-hash=5b48487b68
nginx-deployment-5b78bc88d4 2 2 2 45m nginx-containers nginx:latest app=nginx,pod-template-hash=5b78bc88d4
nginx-deployment-64bcc7b77f 0 0 0 74m nginx-containers nginx:1.14.0 app=nginx,pod-template-hash=64bcc7b77f
[root@k8s-master-10 ~]#
pod指定版本记录回滚⭐️⭐️
全流程⭐️
# 清理旧环境，删除Deployment，也会删除对应的RS控制器记录
[root@k8s-master-10 ~]#kubectl delete -f deployment-nginx.yml
deployment.apps "nginx-deployment" deleted
# 修改镜像版本
[root@k8s-master-10 ~]#grep image deployment-nginx.yml
- image: nginx:1.14.0
imagePullPolicy: IfNotPresent
# 创建第一版pod，记录版本，添加参数 --record
[root@k8s-master-10 ~]#kubectl create -f deployment-nginx.yml --record
deployment.apps/nginx-deployment created
# 测试版本
[root@k8s-master-10 ~]#curl -s 10.2.2.32 -I |grep Server
Server: nginx/1.14.0
# 更新第二版 1.15.0
[root@k8s-master-10 ~]#kubectl -n yuchaoit set image deployment nginx-deployment nginx-containers=nginx:1.15.0
deployment.apps/nginx-deployment image updated
# 验证
[root@k8s-master-10 ~]#curl -s 10.2.1.24 -I |grep Server
Server: nginx/1.15.0
# 更新第三版1.16.0
kubectl -n yuchaoit set image deployment nginx-deployment nginx-containers=nginx:1.16.0
# 验证
[root@k8s-master-10 ~]#curl 10.2.1.25 -s -I |grep Server
Server: nginx/1.16.0
# 查看控制器所有的历史版本
[root@k8s-master-10 ~]#kubectl rollout history deployment nginx-deployment -n yuchaoit
deployment.apps/nginx-deployment
REVISION CHANGE-CAUSE
1 kubectl create --filename=deployment-nginx.yml --record=true
2 kubectl create --filename=deployment-nginx.yml --record=true
3 kubectl create --filename=deployment-nginx.yml --record=true
# 查看指定历史版本的信息
[root@k8s-master-10 ~]#kubectl rollout history deployment nginx-deployment -n yuchaoit --revision=1
# 回到指定版本
[root@k8s-master-10 ~]#kubectl rollout undo deployment nginx-deployment -n yuchaoit --to-revision=1
deployment.apps/nginx-deployment rolled back
[root@k8s-master-10 ~]#kubectl -n yuchaoit get po -owide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
nginx-deployment-64bcc7b77f-6xb95 1/1 Running 0 5s 10.2.1.26 k8s-node-11 <none> <none>
nginx-deployment-64bcc7b77f-jxwzd 1/1 Running 0 5s 10.2.2.35 k8s-node-12 <none> <none>
[root@k8s-master-10 ~]#curl -s -I 10.2.2.35 |grep Server
Server: nginx/1.14.0
[root@k8s-master-10 ~]#
扩缩容deployment
[root@k8s-master-10 ~]#kubectl -n yuchaoit scale deployment nginx-deployment --replicas=5
deployment.apps/nginx-deployment scaled
[root@k8s-master-10 ~]#kubectl -n yuchaoit get po -owide -w
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
nginx-deployment-64bcc7b77f-6xb95 1/1 Running 0 72s 10.2.1.26 k8s-node-11 <none> <none>
nginx-deployment-64bcc7b77f-fgbdq 1/1 Running 0 3s 10.2.1.27 k8s-node-11 <none> <none>
nginx-deployment-64bcc7b77f-jbf9d 1/1 Running 0 3s 10.2.2.36 k8s-node-12 <none> <none>
nginx-deployment-64bcc7b77f-jxwzd 1/1 Running 0 72s 10.2.2.35 k8s-node-12 <none> <none>
nginx-deployment-64bcc7b77f-lrxrx 1/1 Running 0 3s 10.2.2.37 k8s-node-12 <none> <none>
# 缩容
[root@k8s-master-10 ~]#kubectl -n yuchaoit scale deployment nginx-deployment --replicas=1
deployment.apps/nginx-deployment scaled
[root@k8s-master-10 ~]#kubectl -n yuchaoit get po -owide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
nginx-deployment-64bcc7b77f-6xb95 1/1 Running 0 108s 10.2.1.26 k8s-node-11 <none> <none>
[root@k8s-master-10 ~]#
