Kubernetes之deployment详解
1.deployment介绍
为了更好地解决服务编排的问题，k8s在V1.2版本开始，引入了deployment控制器，值得一提的是，这种控制器并不直接管理pod，
而是通过管理replicaset来间接管理pod，即：deployment管理replicaset，replicaset管理pod。所以deployment比replicaset的功能更强大。
deployment的主要功能有下面几个：
- 支持replicaset的所有功能
- 支持版本的滚动更新和版本回退
- 支持发布的停止、继续
deployment的资源清单文件
apiVersion: apps/v1 #版本号 kind: Deployment #类型 metadata: #元数据 name: #rs名称 namespace: #所属命名空间 labels: #标签 controller: deploy spec: #详情描述 replicas: #副本数量 revisionHistoryLimit: #保留历史版本，默认是10 paused: #暂停部署，默认是false progressDeadlineSeconds: #部署超时时间(s)，默认是600 strategy: #策略 type: RollingUpdates #滚动更新策略 rollingUpdate: #滚动更新 maxSurge: #最大额外可以存在的副本数，可以为百分比，也可以为整数 maxUnavaliable: #最大不可用状态的pod的最大值，可以为百分比，也可以为整数 selector: #选择器，通过它指定该控制器管理哪些pod matchLabels: #Labels匹配规则 app: nginx-pod matchExpressions: #Expression匹配规则 - {key: app, operator: In, values: [nginx-pod]} template: #模板，当副本数量不足时，会根据下面的模板创建pod副本 metadata: labels: app: nginx-pod spec: containers: - name: nginx image: nginx:1.17.1 ports: - containerPort: 80
2.deployment创建
1）nginx-deployment.yaml，内容如下：
[root@k8s-master yaml]# cat nginx-deployment.yaml apiVersion: apps/v1 kind: Deployment metadata: name: nginx-deployment namespace: default spec: replicas: 3 selector: matchLabels: app: nginx-pod template: metadata: labels: app: nginx-pod spec: containers: - name: nginx image: nginx:1.17.1
2）执行创建
[root@k8s-master yaml]# kubectl apply -f nginx-deployment.yaml [root@k8s-master yaml]# kubectl get deploy,rs,po -o wide NAME READY UP-TO-DATE AVAILABLE AGE CONTAINERS IMAGES SELECTOR deployment.apps/nginx-deployment 3/3 3 3 41s nginx nginx:1.17.1 app=nginx-pod NAME DESIRED CURRENT READY AGE CONTAINERS IMAGES SELECTOR replicaset.apps/nginx-deployment-5d9c9b97bb 3 3 3 41s nginx nginx:1.17.1 app=nginx-pod,pod-template-hash=5d9c9b97bb NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES pod/nginx-deployment-5d9c9b97bb-4p2hx 1/1 Running 0 41s 10.244.36.82 k8s-node1 <none> <none> pod/nginx-deployment-5d9c9b97bb-p58lf 1/1 Running 0 41s 10.244.169.133 k8s-node2 <none> <none> pod/nginx-deployment-5d9c9b97bb-sm9zz 1/1 Running 0 41s 10.244.36.83 k8s-node1 <none> <none>
注：查看deployment控制的rs和pod，发现rs是在deployment之后加了一段字符串，而pod是在rs之后加了一段字符串
即：deployment管理replicaset，replicaset管理pod。所以deployment比replicaset的功能更强大
3.deployment扩缩容
Deployment扩容缩容的本质其实就是改变ReplicaSet的数量来控制Pod的数量，增加就是扩容，缩小就是缩容。
- 方式一：kubectl scale deploy deploy名称 --replicas=pod数量 -n 命名空间
- 方式二：kubectl edit deploy deploy名字 -n 命名空间
注：以上两种方式均可以实现扩缩容，根据实际情况选择即可
1）扩容示例：假设我们要将nginx从3个pod扩展到5个pod
[root@k8s-master yaml]# kubectl scale deploy nginx-deployment --replicas=5 -n default deployment.apps/nginx-deployment scaled [root@k8s-master yaml]# kubectl get deploy,rs,po NAME READY UP-TO-DATE AVAILABLE AGE deployment.apps/nginx-deployment 5/5 5 5 19m NAME DESIRED CURRENT READY AGE replicaset.apps/nginx-deployment-5d9c9b97bb 5 5 5 19m NAME READY STATUS RESTARTS AGE pod/nginx-deployment-5d9c9b97bb-4p2hx 1/1 Running 0 19m pod/nginx-deployment-5d9c9b97bb-bjnbs 1/1 Running 0 29s pod/nginx-deployment-5d9c9b97bb-p58lf 1/1 Running 0 19m pod/nginx-deployment-5d9c9b97bb-sm9zz 1/1 Running 0 19m pod/nginx-deployment-5d9c9b97bb-x92nj 1/1 Running 0 29s
2）缩容示例：假设我们将nginx从5个缩容到3个
[root@k8s-master yaml]# kubectl edit deploy nginx-deployment -n default deployment.apps/nginx-deployment edited #找到replicas，将其数量改为3 spec: progressDeadlineSeconds: 600 replicas: 3 [root@k8s-master yaml]# kubectl get deploy,rs,po NAME READY UP-TO-DATE AVAILABLE AGE deployment.apps/nginx-deployment 3/3 3 3 29m NAME DESIRED CURRENT READY AGE replicaset.apps/nginx-deployment-5d9c9b97bb 3 3 3 29m NAME READY STATUS RESTARTS AGE pod/nginx-deployment-5d9c9b97bb-4p2hx 1/1 Running 0 29m pod/nginx-deployment-5d9c9b97bb-bjnbs 1/1 Running 0 9m47s pod/nginx-deployment-5d9c9b97bb-p58lf 1/1 Running 0 29m
4.deployment更新和回滚
deployment支持两种镜像更新策略：重建更新和滚动更新（默认），可以通过strategy选项进行配置
strategy：指定新的pod替换旧的pod的策略，支持两个属性：
type：指定策略类型，支持两种策略
Recreate：在创建出新的pod之前会先杀掉所有已存在的pod
RollingUpdate：滚动更新，就是杀死一部分，就启动一部分，在更新过程中，存在两个版本pod
rollingUpdate：当type为RollingUpdate时生效，用于为RollingUpdate设置参数，支持两个属性
maxUnavailable：用来指定在升级过程中不可用pod的最大数量，默认为25%
maxSurge：用来指定在升级过程中可以超过期望的pod的最大数量，默认为25%
1）重建更新
编辑pc-deployment.yaml，在spec节点下添加更新策略
spec:
strategy: #策略
type: Recreate #重建更新策略
创建deploy进行验证
#首先记录原本的pod名 [root@k8s-master yaml]# kubectl get pods NAME READY STATUS RESTARTS AGE nginx-deployment-5d9c9b97bb-4p2hx 1/1 Running 0 40m nginx-deployment-5d9c9b97bb-bjnbs 1/1 Running 0 21m nginx-deployment-5d9c9b97bb-p58lf 1/1 Running 0 40m #更改pod镜像 [root@k8s-master yaml]# kubectl set image deploy nginx-deployment nginx=nginx:1.17.2 deployment.apps/nginx-deployment image updated #再次查看镜像(先是中断rs控制器，再是拉取nginx新镜像，最后根据新镜像创建pod) [root@k8s-master yaml]# kubectl get pods NAME READY STATUS RESTARTS AGE nginx-deployment-5d9c9b97bb-4p2hx 0/1 Terminating 0 42m [root@k8s-master yaml]# kubectl get pods NAME READY STATUS RESTARTS AGE nginx-deployment-7c7477c7ff-5xmxl 0/1 ContainerCreating 0 0s nginx-deployment-7c7477c7ff-xjdcg 0/1 ContainerCreating 0 0s nginx-deployment-7c7477c7ff-xtxh7 0/1 ContainerCreating 0 0s [root@k8s-master yaml]# kubectl get pods NAME READY STATUS RESTARTS AGE nginx-deployment-7c7477c7ff-5xmxl 1/1 Running 0 63s nginx-deployment-7c7477c7ff-xjdcg 1/1 Running 0 63s nginx-deployment-7c7477c7ff-xtxh7 1/1 Running 0 63s
注：Recreate 的升级策略就是完全删除再重建，这样会导致服务一段时间不可用
2）滚动更新
编辑pc-deployment.yaml，在spec节点下添加滚动更新策略（也可以把strategy去掉，因为默认滚动更新策略）
strategy: type: RollingUpdate #滚动更新策略 rollingUpdate: maxUnavailable: 25% #用来指定在升级过程中不可用pod的最大数量，默认为25% maxSurge: 25% #用来指定在升级过程中可以超过期望的pod的最大数量，默认为25%
创建deploy进行验证
滚动更新前(nginx为1.17.2) [root@k8s-master yaml]# kubectl get deploy,rs,pod -o wide NAME READY UP-TO-DATE AVAILABLE AGE CONTAINERS IMAGES SELECTOR deployment.apps/nginx-deployment 3/3 3 3 6m4s nginx nginx:1.17.2 app=nginx-pod NAME DESIRED CURRENT READY AGE CONTAINERS IMAGES SELECTOR replicaset.apps/nginx-deployment-5d9c9b97bb 0 0 0 6m4s nginx nginx:1.17.1 app=nginx-pod,pod-template-hash=5d9c9b97bb replicaset.apps/nginx-deployment-7c7477c7ff 3 3 3 87s nginx nginx:1.17.2 app=nginx-pod,pod-template-hash=7c7477c7ff NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES pod/nginx-deployment-7c7477c7ff-769lc 1/1 Running 0 84s 10.244.36.94 k8s-node1 <none> <none> pod/nginx-deployment-7c7477c7ff-9jvdj 1/1 Running 0 87s 10.244.36.93 k8s-node1 <none> <none> pod/nginx-deployment-7c7477c7ff-m9b2r 1/1 Running 0 85s 10.244.169.139 k8s-node2 <none> <none> 更新镜像 [root@k8s-master yaml]# kubectl set image deploy nginx-deployment nginx=nginx:1.17.3 deployment.apps/nginx-deployment image updated [root@k8s-master yaml]# kubectl get deploy,rs,pod -o wide NAME READY UP-TO-DATE AVAILABLE AGE CONTAINERS IMAGES SELECTOR deployment.apps/nginx-deployment 3/3 3 3 8m43s nginx nginx:1.17.3 app=nginx-pod NAME DESIRED CURRENT READY AGE CONTAINERS IMAGES SELECTOR replicaset.apps/nginx-deployment-5d9c9b97bb 0 0 0 8m43s nginx nginx:1.17.1 app=nginx-pod,pod-template-hash=5d9c9b97bb replicaset.apps/nginx-deployment-76fd8c7f84 3 3 2 3s nginx nginx:1.17.3 app=nginx-pod,pod-template-hash=76fd8c7f84 replicaset.apps/nginx-deployment-7c7477c7ff 1 1 1 4m6s nginx nginx:1.17.2 app=nginx-pod,pod-template-hash=7c7477c7ff NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES pod/nginx-deployment-76fd8c7f84-2ttfm 1/1 Running 0 2s 10.244.169.140 k8s-node2 <none> <none> pod/nginx-deployment-76fd8c7f84-4g5wq 0/1 ContainerCreating 0 1s <none> k8s-node1 <none> <none> pod/nginx-deployment-76fd8c7f84-694sl 1/1 Running 0 3s 10.244.36.95 k8s-node1 <none> <none> pod/nginx-deployment-7c7477c7ff-769lc 0/1 Terminating 0 4m3s 10.244.36.94 k8s-node1 <none> <none> pod/nginx-deployment-7c7477c7ff-9jvdj 1/1 Running 0 4m6s 10.244.36.93 k8s-node1 <none> <none> pod/nginx-deployment-7c7477c7ff-m9b2r 1/1 Terminating 0 4m4s 10.244.169.139 k8s-node2 <none> <none> 查看更新后镜像版本（nginx为1.17.3） [root@k8s-master yaml]# kubectl get deploy,rs,pod -o wide NAME READY UP-TO-DATE AVAILABLE AGE CONTAINERS IMAGES SELECTOR deployment.apps/nginx-deployment 3/3 3 3 10m nginx nginx:1.17.3 app=nginx-pod NAME DESIRED CURRENT READY AGE CONTAINERS IMAGES SELECTOR replicaset.apps/nginx-deployment-5d9c9b97bb 0 0 0 10m nginx nginx:1.17.1 app=nginx-pod,pod-template-hash=5d9c9b97bb replicaset.apps/nginx-deployment-76fd8c7f84 3 3 3 118s nginx nginx:1.17.3 app=nginx-pod,pod-template-hash=76fd8c7f84 replicaset.apps/nginx-deployment-7c7477c7ff 0 0 0 6m1s nginx nginx:1.17.2 app=nginx-pod,pod-template-hash=7c7477c7ff NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES pod/nginx-deployment-76fd8c7f84-2ttfm 1/1 Running 0 117s 10.244.169.140 k8s-node2 <none> <none> pod/nginx-deployment-76fd8c7f84-4g5wq 1/1 Running 0 116s 10.244.36.96 k8s-node1 <none> <none> pod/nginx-deployment-76fd8c7f84-694sl 1/1 Running 0 118s 10.244.36.95 k8s-node1 <none> <none>
3）滚动更新的过程
5.deployment回滚
deployment支持版本升级过程中的暂停，继续功能以及版本回退等诸多功能，下面具体来看
kubectl rollout：版本升级相关功能，支持下面的选项：
- 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回滚到指定版本）
#当前nginx版本 [root@k8s-master yaml]# kubectl get deploy -o wide NAME READY UP-TO-DATE AVAILABLE AGE CONTAINERS IMAGES SELECTOR nginx-deployment 3/3 3 3 24m nginx nginx:1.17.3 app=nginx-pod #查看升级状态 [root@k8s-master yaml]# kubectl rollout status deploy nginx-deployment deployment "nginx-deployment" successfully rolled out #查看升级历史 [root@k8s-master yaml]# kubectl rollout history deploy nginx-deployment deployment.apps/nginx-deployment REVISION CHANGE-CAUSE 1 <none> 2 <none> 3 <none> #查看版本详情 [root@k8s-master yaml]# kubectl rollout history deployment nginx-deployment --revision=2 deployment.apps/nginx-deployment with revision #2 Pod Template: Labels: app=nginx-pod pod-template-hash=7c7477c7ff Containers: nginx: Image: nginx:1.17.2 Port: <none> Host Port: <none> Environment: <none> Mounts: <none> Volumes: <none> #版本回滚这里使用--to-revision=2回滚到2版本，如果省略这个选项，则会回退到上个版本 [root@k8s-master yaml]# kubectl rollout undo deployment nginx-deployment --to-revision=2 deployment.apps/nginx-deployment rolled back #查看当前版本（nginx为1.17.2）说明回退成功 [root@k8s-master yaml]# kubectl get deployments -o wide NAME READY UP-TO-DATE AVAILABLE AGE CONTAINERS IMAGES SELECTOR nginx-deployment 3/3 3 3 27m nginx nginx:1.17.2 app=nginx-pod
6.金丝雀发布（灰度发布）
deployment支持更新过程中的控制，如"暂停（pause）"或"继续（resume）"更新操作，观察新服务实际运行状况再决定操作
金丝雀发布（可以理解为灰度发布）：比如有一批新的pod资源创建完成后立即暂停更新过程，此时，仅存在一部分新版本的应用，主体部分还是旧的版本。然后，再筛选一小部分的用户请求路由到新的pod应用，继续观察能否稳定地按期望的方式运行。确定没问题之后再继续完成余下的pod资源滚动更新，否则立即回滚更新操作。这就是所谓的金丝雀发布。
#更新deployment版本，并配置暂停deployment [root@k8s-master yaml]# kubectl set image deploy nginx-deployment nginx=nginx:1.17.3 && kubectl rollout pause deploy nginx-deployment deployment.apps/nginx-deployment image updated deployment.apps/nginx-deployment paused #查看rs，发现老版本rs没有减少，新版本rs增加一个，状态为更新中.... [root@k8s-master yaml]# kubectl get rs,deploy NAME DESIRED CURRENT READY AGE replicaset.apps/nginx-deployment-5d9c9b97bb 0 0 0 39m replicaset.apps/nginx-deployment-76fd8c7f84 1 1 1 30m replicaset.apps/nginx-deployment-7c7477c7ff 3 3 3 34m NAME READY UP-TO-DATE AVAILABLE AGE deployment.apps/nginx-deployment 4/3 1 4 39m [root@k8s-master yaml]# kubectl rollout status deployment nginx-deployment Waiting for deployment "nginx-deployment" rollout to finish: 1 out of 3 new replicas have been updated... #继续deploy的更新 [root@k8s-master yaml]# kubectl rollout resume deployment nginx-deployment deployment.apps/nginx-deployment resumed [root@k8s-master yaml]# kubectl rollout status deployment nginx-deployment Waiting for deployment "nginx-deployment" rollout to finish: 1 old replicas are pending termination... Waiting for deployment "nginx-deployment" rollout to finish: 1 old replicas are pending termination... deployment "nginx-deployment" successfully rolled out #查看rs更新结果，发现老版本均停止，新版本已经创建好 [root@k8s-master yaml]# kubectl get rs,deploy NAME DESIRED CURRENT READY AGE replicaset.apps/nginx-deployment-5d9c9b97bb 0 0 0 41m replicaset.apps/nginx-deployment-76fd8c7f84 3 3 3 32m replicaset.apps/nginx-deployment-7c7477c7ff 0 0 0 36m NAME READY UP-TO-DATE AVAILABLE AGE deployment.apps/nginx-deployment 3/3 3 3 41m
作者：杰宏唯一
本文版权归作者和博客园共有，欢迎转载，但未经作者同意必须保留此段声明，且在文章页面明显位置给出原文连接，否则保留追究法律责任的权利.
【推荐】国内首个AI IDE，深度理解中文开发场景，立即下载体验Trae
【推荐】编程新体验，更懂你的AI，立即体验豆包MarsCode编程助手
【推荐】抖音旗下AI助手豆包，你的智能百科全书，全免费不限次数
【推荐】轻量又高性能的 SSH 工具 IShell：AI 加持，快人一步
· Linux系列：如何用heaptrack跟踪.NET程序的非托管内存泄露
· 开发者必知的日志记录最佳实践
· SQL Server 2025 AI相关能力初探
· Linux系列：如何用 C#调用 C方法造成内存泄露
· AI与.NET技术实操系列（二）：开始使用ML.NET
· 被坑几百块钱后，我竟然真的恢复了删除的微信聊天记录！
· 没有Manus邀请码？试试免邀请码的MGX或者开源的OpenManus吧
· 【自荐】一款简洁、开源的在线白板工具 Drawnix
· 园子的第一款AI主题卫衣上架——"HELLO! HOW CAN I ASSIST YOU TODAY
· Docker 太简单，K8s 太复杂？w7panel 让容器管理更轻松！