参考：https://blog.51cto.com/wutengfei/2116663
创建和管理多个Pod--Deployment
Deployment 为 Pod 和 ReplicaSet 提供了一个声明式定义(declarative)方法，用来替代以前的ReplicationController 来方便的管理应用。
你只需要在 Deployment 中描述想要的目标状态是什么，Deployment controller 就会帮你将 Pod 和ReplicaSet 的实际状态改变到你的目标状态。你可以定义一个全新的 Deployment 来创建 ReplicaSet 或者删除已有的 Deployment 并创建一个新的来替换。
复制控制器（Replication Controller，RC）
RC是K8s集群中最早的保证Pod高可用的API对象。通过监控运行中的Pod来保证集群中运行指定数目的Pod副本。指定的数目可以是多个也可以是1个；少于指定数目，RC就会启动运行新的Pod副本；多于指定数目，RC就会杀死多余的Pod副本。即使在指定数目为1的情况下，通过RC运行Pod也比直接运行Pod更明智，因为RC也可以发挥它高可用的能力，保证永远有1个Pod在运行。RC是K8s较早期的技术概念，只适用于长期伺服型的业务类型，比如控制小机器人提供高可用的Web服务。
副本集（Replica Set，RS）
RS是新一代RC，提供同样的高可用能力，区别主要在于RS后来居上，能支持更多种类的匹配模式。副本集对象一般不单独使用，而是作为Deployment的理想状态参数使用。
Deployment典型的应用场景包括
（1）定义Deployment来创建Pod和ReplicaSet
（2）滚动升级和回滚应用；如果当前状态不稳定，回滚到之前的Deployment revision。每次回滚都会更新Deployment的revision。
（3）扩容和缩容，扩容Deployment以满足更高的负载。
（4）暂停和继续Deployment，暂停Deployment来应用PodTemplateSpec的多个修复，然后恢复上线。
实战Deployment
1.创建
定义deployment yaml，用于发布应用服务
[root@k8s-master ~]# pwd /root [root@k8s-master ~]# cat deployment-hello.yaml apiVersion: extensions/v1beta1 kind: Deployment metadata: name: hello spec: replicas: 4 template: metadata: labels: run: hello spec: containers: - name: hello image: tomcat:8 #确保node节点上有该镜像且可正常运行，注意是node节点机器上，不是master机器 imagePullPolicy: IfNotPresent ##Always,IfNotPresent,Never ports: - name: http containerPort: 8080
布署
[root@k8s-master ~]# kubectl create -f deployment-hello.yaml
deployment.extensions/hello created
查看部署
root@k8s-master ~]# kubectl get deployments #查看deploy是否成功 NAME READY UP-TO-DATE AVAILABLE AGE hello 0/4 4 0 55s
查看rs(副本集)
[root@k8s-master ~]# kubectl get rs ##因为deployment是三层架构，看rs是否成功,我们看自动创建4个rs.名称后面的字符串是 模板的哈希值。是不会发生变化的，最后pod的是随机值 NAME DESIRED CURRENT READY AGE hello-68df45bc79 4 4 0 8m29s
查看容器组
[root@k8s-master ~]# kubectl get pods #查看最后一层pod NAME READY STATUS RESTARTS AGE hello-7d46c7db4c-42bwl 0/1 ContainerCreating 0 67s hello-7d46c7db4c-rg9fq 0/1 ContainerCreating 0 67s hello-7d46c7db4c-tclps 0/1 ContainerCreating 0 67s hello-7d46c7db4c-whvsw 0/1 ContainerCreating 0
2.扩容
将容器组个数扩充至8个
##说明
如果集群支持 horizontal pod autoscaling 的话，还可以为Deployment设置自动扩展：
kubectl autoscale deployment nginx-test --min=10 --max=15 --cpu-percent=80 --namespace=test
[root@k8s-master ~]# kubectl get deployments NAME READY UP-TO-DATE AVAILABLE AGE hello 4/4 4 4 7h1m [root@k8s-master ~]# kubectl scale deployment hello --replicas 8 --namespace=default deployment.extensions/hello scaled [root@k8s-master ~]# kubectl get deployments NAME READY UP-TO-DATE AVAILABLE AGE hello 4/8 8 4 7h1m [root@k8s-master ~]# kubectl get rs NAME DESIRED CURRENT READY AGE hello-57b49c67cf 8 8 4 7h2m [root@k8s-master ~]# kubectl get pods ##正在扩容 NAME READY STATUS RESTARTS AGE hello-57b49c67cf-4r5wf 0/1 ContainerCreating 0 64s hello-57b49c67cf-6hh59 1/1 Running 0 7h2m hello-57b49c67cf-6ml78 1/1 Running 0 7h2m hello-57b49c67cf-8lgcl 0/1 ContainerCreating 0 64s hello-57b49c67cf-8xqgw 1/1 Running 0 7h2m hello-57b49c67cf-dg5lr 0/1 ContainerCreating 0 64s hello-57b49c67cf-pd826 1/1 Running 0 7h2m hello-57b49c67cf-rj28m 0/1 ContainerCreating 0 64s [root@k8s-master ~]#
3.缩容
将容器组缩减至2个
[root@k8s-master ~]# kubectl scale deployment hello --replicas 2 --namespace=default deployment.extensions/hello scaled [root@k8s-master ~]# kubectl get deployment --namespace=default NAME READY UP-TO-DATE AVAILABLE AGE hello 2/2 2 2 7h5m [root@k8s-master ~]# kubectl get rs --namespace=default NAME DESIRED CURRENT READY AGE hello-57b49c67cf 2 2 0 7h5m [root@k8s-master ~]# kubectl get pods --namespace=default ##正在缩容 NAME READY STATUS RESTARTS AGE hello-57b49c67cf-4r5wf 1/1 Running 0 5m29s hello-57b49c67cf-6hh59 1/1 Terminating 1 7h6m hello-57b49c67cf-6ml78 1/1 Terminating 1 7h6m hello-57b49c67cf-8lgcl 0/1 Terminating 0 5m29s hello-57b49c67cf-8xqgw 1/1 Terminating 1 7h6m hello-57b49c67cf-dg5lr 1/1 Running 0 5m29s hello-57b49c67cf-pd826 1/1 Terminating 1 7h6m hello-57b49c67cf-rj28m 0/1 Terminating 0 5m29s
[root@k8s-master ~]# kubectl get pods --namespace=default ##缩容完成
NAME READY STATUS RESTARTS AGE
hello-57b49c67cf-4r5wf 1/1 Running 0 7m7s
hello-57b49c67cf-dg5lr 1/1 Running 0 7m7s
4.滚动升级
将image: mytomcat:v8 升级为mytomcat:v9
kubectl edit deployment/hello修改 .spec.template.spec.containers[0].image
[root@k8s-master ~]# kubectl get deployment -o wide ##触发rollout NAME READY UP-TO-DATE AVAILABLE AGE CONTAINERS IMAGES SELECTOR hello 2/2 2 2 7h14m hello mytomcat:v8 run=hello [root@k8s-master ~]# docker tag mytomcat:v8 mytomcat:v9 [root@k8s-master ~]# docker images REPOSITORY TAG IMAGE ID CREATED SIZE mytomcat v8 f1332ae3f570 4 days ago 463MB mytomcat v9 f1332ae3f570 4 days ago 463MB
[root@k8s-master ~]# kubectl set image deployment/hello hello=mytomcat:v9 deployment.extensions/hello image updated
[root@k8s-master ~]# kubectl rollout status deployment/hello ##查看升级进度
Waiting for deployment "hello" rollout to finish: 1 old replicas are pending termination...
Waiting for deployment "hello" rollout to finish: 1 old replicas are pending termination...
Waiting for deployment "hello" rollout to finish: 1 old replicas are pending termination...
Waiting for deployment "hello" rollout to finish: 1 of 2 updated replicas are available...
deployment "hello" successfully rolled out
[root@k8s-master ~]# kubectl rollout status deployment/hello
deployment "hello" successfully rolled out
[root@k8s-master ~]# kubectl get deployments -o wide NAME READY UP-TO-DATE AVAILABLE AGE CONTAINERS IMAGES SELECTOR hello 1/2 2 1 7h46m hello mytomcat:v9 run=hello
查看历史版本
[root@k8s-master ~]# kubectl rollout history deployment/hello deployment.extensions/hello REVISION CHANGE-CAUSE 1 <none> 2 <none>
查看pod
[root@k8s-master ~]# kubectl get pods NAME READY STATUS RESTARTS AGE hello-57b49c67cf-4r5wf 1/1 Running 0 52m hello-5cd4456b66-2wc9q 0/1 ImagePullBackOff 0 8m10s hello-5cd4456b66-hzdns 0/1 ImagePullBackOff 0 8m9s
[root@k8s-master ~]# kubectl describe pod hello-5cd4456b66-2wc9q
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal Scheduled 10m default-scheduler Successfully assigned default/hello-5cd4456b66-2wc9q to k8s-node1
Normal Pulling 8m54s (x4 over 10m) kubelet, k8s-node1 Pulling image "mytomcat:v9"
Warning Failed 8m35s (x4 over 10m) kubelet, k8s-node1 Failed to pull image "mytomcat:v9": rpc error: code = Unknown desc = Error response from daemon: pull access denied for mytomcat, repository does not exist or may require 'docker login'
Warning Failed 8m35s (x4 over 10m) kubelet, k8s-node1 Error: ErrImagePull
Normal BackOff 8m8s (x7 over 10m) kubelet, k8s-node1 Back-off pulling image "mytomcat:v9"
Warning Failed 5m53s (x16 over 10m) kubelet, k8s-node1 Error: ImagePullBackOff
报ImagePullBackOff错，到节点k8s-node1主机上执行docker tag mytomcat:v8 mytomcat:v9
[root@k8s-master ~]# kubectl get pods -o wide NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES hello-5cd4456b66-2wc9q 1/1 Running 0 10m 10.244.1.30 k8s-node1 <none> <none> hello-5cd4456b66-hzdns 1/1 Running 0 10m 10.244.1.31 k8s-node1 <none> <none> [root@k8s-master ~]# kubectl get deployments NAME READY UP-TO-DATE AVAILABLE AGE hello 2/2 2 2 8h
[root@k8s-master ~]# kubectl rollout status deployment/hello
deployment "hello" successfully rolled out
[root@k8s-master ~]# kubectl get deployments -o wide NAME READY UP-TO-DATE AVAILABLE AGE CONTAINERS IMAGES SELECTOR hello 2/2 2 2 8h hello mytomcat:v9 run=hello [root@k8s-master ~]# kubectl rollout undo deployment/hello --namespace=default deployment.extensions/hello rolled back [root@k8s-master ~]# kubectl rollout status deployment/hello deployment "hello" successfully rolled out [root@k8s-master ~]# kubectl rollout history deployment/hello deployment.extensions/hello REVISION CHANGE-CAUSE 2 <none> 3 <none> [root@k8s-master ~]# kubectl get deployments -o wide NAME READY UP-TO-DATE AVAILABLE AGE CONTAINERS IMAGES SELECTOR hello 2/2 2 2 8h hello mytomcat:v8 run=hello [root@k8s-master ~]# kubectl get pods NAME READY STATUS RESTARTS AGE hello-57b49c67cf-gghms 1/1 Running 0 93s hello-57b49c67cf-l6fk2 1/1 Running 0 94s [root@k8s-master ~]# kubectl describe pod hello-57b49c67cf-gghms Image: mytomcat:v8
6.暂停或恢复
你可以在触发一次或多次更新前暂停一个Deployment，然后再恢复它。这样你就能多次暂停和恢复Deployment，在此期间进行一些修复工作，而不会出发不必要的rollout。
kubectl rollout pause deployment/nginx-deployment
kubectl set image deploy/hello hello=mytomcat:v9
##kubectl edit deployment/hello
kubectl set resources deployment hello -c=hello --limits=cpu=200m,memory=512Mi ##命令详解 -c为容器名限定kubectl get deployments -o wide中的containers
kubectl rollout resume deploy hello