Deployment的典型用例：
1. 使用Deployment来启动（上线/部署）一个Pod或者ReplicaSet
2. 检查一个Deployment是否成功执行
3. 更新Deployment来重新创建相应的Pods（例如，需要使用一个新的Image）
4. 如果现有的Deployment不稳定，那么回滚到一个早期的稳定的Deployment版本
5. 暂停或者恢复一个Deployment

deployment 部署过程

创建一个deployment

1. yaml文件

   apiVersion: extensions/v1beta1
   kind: Deployment
   metadata:
     name: nginx-deployment
   spec:
     replicas: 3
     template:
       metadata:
         labels:
            app: nginx
       spec:
          containers:
            - name: nginx
              image: docker.io/nginx
              ports:
               - containerPort: 80

2. 创建deployment

   kubectl create -f nginx-deployment.yaml  —record

   —record选项是为了记录当前执行的命令所创建/更新的资源，即使用—record选型执行的deployment都会被记录下来，
   用于以后查看每次deployment的细节，查看所有Deployment的历史，或者回滚到某一历史版本。

3. 查看deployment的状态

   [root@controller01 deployment]# kubectl get deployments nginx-deployment
   NAME               DESIRED   CURRENT   UP-TO-DATE   AVAILABLE   AGE
   nginx-deployment   3         3         3            3           29s

4. 查看pod的label
   
[root@controller01 deployment]# kubectl get pods --show-labels | grep ^nginx
nginx-deployment-4087004473-8lvh9 0/1 ContainerCreating 0 6m app=nginx,pod-template-hash=4087004473
nginx-deployment-4087004473-d8fwd 0/1 ContainerCreating 0 6m app=nginx,pod-template-hash=4087004473
nginx-deployment-4087004473-xwfkz 0/1 ContainerCreating 0 6m app=nginx,pod-template-hash=4087004473

5. 查看deployment的状态，生成yaml文件
   
[root@controller01 deployment]# kubectl get deployment nginx-deployment -o yaml


升级操作

1. yaml文件

   apiVersion: extensions/v1beta1
   kind: Deployment
   metadata:
   name: nginx-deployment
   spec:
   replicas: 3
   template:
   metadata:
    labels:
        app: nginx
   spec:
      containers:
        - name: nginx
          image: docker.io/bitnami/nginx
          ports:
           - containerPort: 80

     区别：更改了nginx的image，相当于nginx的升级
   

2. 操作

   kubectl apply -f update-nginx-deployment.yaml 
   

3. 检验升级是否成功

   kubectl get deployments nginx-deployment  -o yaml
   

   可以看到image已经修改

   [root@controller01 deployment]# kubectl get po | grep ^nginx
   nginx-deployment-3905994866-f8bjp   1/1       Running            0          1h
   nginx-deployment-3905994866-j7w0d   1/1       Running            0          36m
   nginx-deployment-3905994866-qc5f3   1/1       Running            0          1h
   

   可以看到服务已经升级结束

   服务升级失败

   [root@controller01 deployment]# kubectl get po | grep ^nginx
   nginx-deployment-3446456637-vdj97   0/1       ImagePullBackOff   0          30m
   nginx-deployment-3446456637-zllkx   0/1       ImagePullBackOff   0          30m
   nginx-deployment-3905994866-f8bjp   1/1       Running            0          51m
   nginx-deployment-3905994866-qc5f3   1/1       Running            0          51m

我们发现最新一次Deployment出现了问题，没有找到相应的镜像，部署失败。对于这种情形，我们选择将Deployment恢复到上一版本，
注意，这里不是通过恢复之前的RC或者Pods或者RS，而是回滚Deployment。

失败回滚

1. 查看可以回滚的版本

   [root@controller01 deployment]# kubectl rollout history deployment nginx-deployment
   deployments "nginx-deployment"
   REVISION    CHANGE-CAUSE
   1       <none>
   2       <none>
   3       <none>

   可以看到一共有三个版本，由于前面没有加 --record，没有看到版本信息

2. 回滚到指定版本

   [root@controller01 deployment]# kubectl rollout undo deployment nginx-deployment --to-revision=2
   deployment "nginx-deployment" rolled back

3. 查看运行是否正常

   [root@controller01 deployment]# kubectl get rs | grep ^nginx
   nginx-deployment-3446456637   0         0         0         41m
   nginx-deployment-3905994866   3         3         3         1h
   nginx-deployment-67353951     0         0         0         5h

暂停服务和重启服务


kubectl rollout pause deployment nginx-deployment
kubectl rollout resume deployment/nginx-deployment


deployment 和 Rs关系

Replica Set（简称RS）是k8s新一代的Pod controller。与RC相比仅有selector存在差异，RS支持了set-based selector（可以使用in、notin、key存在、key不存在四种方式来选择满足条件的label集合）。Deployment是基于RS实现的。