Kubernetes K8S之资源控制器RC、RS、Deployment详解
Kubernetes的资源控制器ReplicationController（RC）、ReplicaSet（RS）、Deployment（Deploy）详解与示例
主机配置规划
| 服务器名称(hostname) | 系统版本 | 配置 | 内网IP | 外网IP(模拟) |
|---|---|---|---|---|
| k8s-master | CentOS7.7 | 2C/4G/20G | 172.16.1.110 | 10.0.0.110 |
| k8s-node01 | CentOS7.7 | 2C/4G/20G | 172.16.1.111 | 10.0.0.111 |
| k8s-node02 | CentOS7.7 | 2C/4G/20G | 172.16.1.112 | 10.0.0.112 |
什么是控制器
kubernetes中内建了很多controller（控制器），这些相当于一个状态机，用来控制pod的具体状态和行为。
部分控制器类型如下：
- ReplicationController 和 ReplicaSet
- Deployment
- DaemonSet
- StatefulSet
- Job/CronJob
- HorizontalPodAutoscaler
ReplicationController 和 ReplicaSet
ReplicationController (RC)用来确保容器应用的副本数始终保持在用户定义的副本数，即如果有容器异常退出，会自动创建新的pod来替代；而异常多出来的容器也会自动回收。
在新版的Kubernetes中建议使用ReplicaSet (RS)来取代ReplicationController。ReplicaSet跟ReplicationController没有本质的不同，只是名字不一样，但ReplicaSet支持集合式selector。
虽然 ReplicaSets 可以独立使用，但如今它主要被Deployments 用作协调 Pod 的创建、删除和更新的机制。当使用 Deployment 时，你不必担心还要管理它们创建的 ReplicaSet，Deployment 会拥有并管理它们的 ReplicaSet。
ReplicaSet 是下一代的 Replication Controller。 ReplicaSet 和 Replication Controller 的唯一区别是选择器的支持。ReplicaSet 支持新的基于集合的选择器需求，这在标签用户指南中有描述。而 Replication Controller 仅支持基于相等选择器的需求。
Deployments
Deployment 控制器为 Pods和 ReplicaSets提供描述性的更新方式。用来替代以前的ReplicationController以方便管理应用。
典型的应用场景包括：
- 定义Deployment来创建Pod和ReplicaSet
- 滚动升级和回滚应用
- 扩容和缩容
- 暂停和继续Deployment
ReplicaSet示例
yaml文件
1 [root@k8s-master controller]# pwd 2 /root/k8s_practice/controller 3 [root@k8s-master controller]# cat ReplicaSet-01.yaml 4 apiVersion: apps/v1 5 kind: ReplicaSet 6 metadata: 7 name: frontend 8 spec: 9 replicas: 3 10 selector: 11 matchLabels: 12 tier: frontend 13 template: 14 metadata: 15 labels: 16 tier: frontend 17 spec: 18 containers: 19 - name: nginx 20 image: registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17 21 imagePullPolicy: IfNotPresent 22 ports: 23 - name: httpd 24 containerPort: 80
创建ReplicaSet，并查看rs状态与详情
1 [root@k8s-master controller]# kubectl apply -f ReplicaSet-01.yaml 2 replicaset.apps/frontend created 3 [root@k8s-master controller]# kubectl get rs -o wide # 查看状态 4 NAME DESIRED CURRENT READY AGE CONTAINERS IMAGES SELECTOR 5 frontend 3 3 3 2m12s nginx registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17 tier=frontend 6 [root@k8s-master controller]# 7 [root@k8s-master controller]# kubectl describe rs frontend # 查看详情 8 Name: frontend 9 Namespace: default 10 Selector: tier=frontend 11 Labels: <none> 12 Annotations: kubectl.kubernetes.io/last-applied-configuration: 13 {"apiVersion":"apps/v1","kind":"ReplicaSet","metadata":{"annotations":{},"name":"frontend","namespace":"default"},"spec":{"replicas":3,"se... 14 Replicas: 3 current / 3 desired 15 Pods Status: 3 Running / 0 Waiting / 0 Succeeded / 0 Failed 16 Pod Template: 17 Labels: tier=frontend 18 Containers: 19 nginx: 20 Image: registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17 21 Port: 80/TCP 22 Host Port: 0/TCP 23 Environment: <none> 24 Mounts: <none> 25 Volumes: <none> 26 Events: 27 Type Reason Age From Message 28 ---- ------ ---- ---- ------- 29 Normal SuccessfulCreate 10m replicaset-controller Created pod: frontend-kltwp 30 Normal SuccessfulCreate 10m replicaset-controller Created pod: frontend-76dbn 31 Normal SuccessfulCreate 10m replicaset-controller Created pod: frontend-jk8td
查看pod状态信息
1 [root@k8s-master controller]# kubectl get pod -o wide --show-labels 2 NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES LABELS 3 frontend-76dbn 1/1 Running 0 5m15s 10.244.4.31 k8s-node01 <none> <none> tier=frontend 4 frontend-jk8td 1/1 Running 0 5m15s 10.244.2.35 k8s-node02 <none> <none> tier=frontend 5 frontend-kltwp 1/1 Running 0 5m15s 10.244.2.34 k8s-node02 <none> <none> tier=frontend
删除一个pod，然后再次查看
1 [root@k8s-master controller]# kubectl delete pod frontend-kltwp 2 pod "frontend-kltwp" deleted 3 [root@k8s-master controller]# 4 [root@k8s-master controller]# kubectl get pod -o wide --show-labels # 可见重新创建了一个pod 5 NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES LABELS 6 frontend-76dbn 1/1 Running 0 7m27s 10.244.4.31 k8s-node01 <none> <none> tier=frontend 7 frontend-jk8td 1/1 Running 0 7m27s 10.244.2.35 k8s-node02 <none> <none> tier=frontend 8 frontend-mf79k 1/1 Running 0 16s 10.244.4.32 k8s-node01 <none> <none> tier=frontend
由上可见，rs又新建了一个pod，保证了pod数总是为3.
Deployment示例
创建 Deployment
yaml文件
1 [root@k8s-master controller]# pwd 2 /root/k8s_practice/controller 3 [root@k8s-master controller]# cat nginx-deployment-1.17.1.yaml 4 apiVersion: apps/v1 5 kind: Deployment 6 metadata: 7 name: nginx-deployment 8 labels: 9 app: nginx 10 spec: 11 replicas: 3 12 # 重点关注该字段 13 selector: 14 matchLabels: 15 app: nginx 16 template: 17 metadata: 18 labels: 19 app: nginx 20 spec: 21 containers: 22 - name: nginx 23 image: registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.1 24 ports: 25 - containerPort: 80 26 [root@k8s-master controller]# 27 [root@k8s-master controller]# cat nginx-deployment-1.17.5.yaml 28 apiVersion: apps/v1 29 kind: Deployment 30 metadata: 31 name: nginx-deployment 32 labels: 33 app: nginx 34 spec: 35 replicas: 3 36 # 重点关注该字段 37 selector: 38 matchLabels: 39 app: nginx 40 template: 41 metadata: 42 labels: 43 app: nginx 44 spec: 45 containers: 46 - name: nginx 47 image: registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.5 48 ports: 49 - containerPort: 80
selector 字段定义 Deployment 如何查找要管理的 Pods。 在这种情况下，会选择在 template（Pod）模板中定义的标签labels（app: nginx）。但更复杂的选择规则是可能的，只要 template （Pod） 模板本身满足规则。
刚启动时状态说明
启动deployment，并查看状态
1 [root@k8s-master controller]# kubectl apply -f nginx-deployment-1.17.1.yaml --record 2 deployment.apps/nginx-deployment created 3 [root@k8s-master controller]# 4 [root@k8s-master controller]# kubectl get deployment -o wide 5 NAME READY UP-TO-DATE AVAILABLE AGE CONTAINERS IMAGES SELECTOR 6 nginx-deployment 2/3 3 2 10s nginx registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.1 app=nginx 7 8 # --record 参数可以记录命令，通过 kubectl rollout history deployment/nginx-deployment 可查询
参数说明：
- NAME：列出集群中 Deployments 的名称
- READY：已就绪副本数/期望副本数
- UP-TO-DATE：显示已更新和正在更新中的副本数
- AVAILABLE：显示应用程序可供用户使用的副本数
- AGE：显示运行的时间
查看ReplicaSet状态
1 [root@k8s-master controller]# kubectl get rs -o wide 2 NAME DESIRED CURRENT READY AGE CONTAINERS IMAGES SELECTOR 3 nginx-deployment-76b9d6bcf5 3 3 2 17s nginx registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.1 app=nginx,pod-template-hash=76b9d6bcf5
参数说明：
- NAME：列出集群中 ReplicaSet的名称
- DESIRED：期望副本数
- CURRENT：当前副本数
- READY：已就绪副本数
- AGE：运行时间
查看pod状态
1 [root@k8s-master controller]# kubectl get pod -o wide 2 NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES 3 nginx-deployment-76b9d6bcf5-ngpg5 1/1 Running 0 26s 10.244.2.43 k8s-node02 <none> <none> 4 nginx-deployment-76b9d6bcf5-rw827 1/1 Running 0 26s 10.244.2.44 k8s-node02 <none> <none> 5 nginx-deployment-76b9d6bcf5-ttf4j 0/1 ContainerCreating 0 26s <none> k8s-node01 <none> <none>
过一会儿状态说明
1 [root@k8s-master controller]# kubectl get deployment -o wide --show-labels 2 NAME READY UP-TO-DATE AVAILABLE AGE CONTAINERS IMAGES SELECTOR LABELS 3 nginx-deployment 3/3 3 3 23m nginx registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.1 app=nginx app=nginx 4 [root@k8s-master controller]# 5 [root@k8s-master controller]# kubectl get rs -o wide --show-labels 6 NAME DESIRED CURRENT READY AGE CONTAINERS IMAGES SELECTOR LABELS 7 nginx-deployment-76b9d6bcf5 3 3 3 23m nginx registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.1 app=nginx,pod-template-hash=76b9d6bcf5 app=nginx,pod-template-hash=76b9d6bcf5 8 [root@k8s-master controller]# 9 [root@k8s-master controller]# kubectl get pod -o wide --show-labels 10 NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES LABELS 11 nginx-deployment-76b9d6bcf5-ngpg5 1/1 Running 0 23m 10.244.2.43 k8s-node02 <none> <none> app=nginx,pod-template-hash=76b9d6bcf5 12 nginx-deployment-76b9d6bcf5-rw827 1/1 Running 0 23m 10.244.2.44 k8s-node02 <none> <none> app=nginx,pod-template-hash=76b9d6bcf5 13 nginx-deployment-76b9d6bcf5-ttf4j 1/1 Running 0 23m 10.244.4.37 k8s-node01 <none> <none> app=nginx,pod-template-hash=76b9d6bcf5
重点说明
1、ReplicaSet 的名称始终被格式化为[DEPLOYMENT-NAME]-[RANDOM-STRING]。随机字符串是随机生成，并使用 pod-template-hash 作为选择器和标签。
2、Deployment 控制器将 pod-template-hash 标签添加到 Deployment 创建或使用的每个 ReplicaSet 。此标签可确保 Deployment 的子 ReplicaSets 不重叠。因此不可修改。
3、注意Deployment、ReplicaSet和Pod三者的名称关系
更新 Deployment
Deployment 可确保在更新时仅关闭一定数量的 Pods。默认情况下，它确保至少 75%所需 Pods 运行（25%最大不可用）。
Deployment 更新过程中还确保仅创建一定数量的 Pods 且高于期望的 Pods 数。默认情况下，它可确保最多增加 25% 期望 Pods 数（25%最大增量）。
备注：实际操作中如果更新Deployment，那么最好通过yaml文件更新，这样回滚到任何版本都非常便捷，而且更容易追述；而不是通过命令行。
如下Deployment示例，由于只有3个副本。因此更新时不会先删除旧的pod，而是先新建一个pod。新pod运行时，才会删除对应老的pod。一切的前提都是为了满足上述的条件。
需求：更新 nginx Pods，从当前的1.17.1版本改为1.17.5版本。
1 # 方式一 2 kubectl edit deployment/nginx-deployment # 然后修改 image 镜像信息 【不推荐】 3 # 上述方法不会记录命令，通过kubectl rollout history deployment/nginx-deployment 无法查询 4 5 # 方式二如下【可使用】： 6 kubectl set image deployment/nginx-deployment nginx=registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.5 --record 7 8 # 方式三如下【推荐★★★★★】 9 kubectl apply -f nginx-deployment-1.17.5.yaml --record 10 11 # --record 参数可以记录命令，通过 kubectl rollout history deployment/nginx-deployment 可查询
要查看更新状态
1 [root@k8s-master controller]# kubectl rollout status deployment/nginx-deployment 2 # 如没有更新完成，则显示更新过程直到更新成功 3 Waiting for deployment "nginx-deployment" rollout to finish: 1 out of 3 new replicas have been updated... 4 Waiting for deployment "nginx-deployment" rollout to finish: 1 out of 3 new replicas have been updated... 5 Waiting for deployment "nginx-deployment" rollout to finish: 2 out of 3 new replicas have been updated... 6 Waiting for deployment "nginx-deployment" rollout to finish: 2 out of 3 new replicas have been updated... 7 Waiting for deployment "nginx-deployment" rollout to finish: 2 out of 3 new replicas have been updated... 8 Waiting for deployment "nginx-deployment" rollout to finish: 1 old replicas are pending termination... 9 Waiting for deployment "nginx-deployment" rollout to finish: 1 old replicas are pending termination... 10 deployment "nginx-deployment" successfully rolled out 11 # 如已更新完毕，直接显示更新成功 12 deployment "nginx-deployment" successfully rolled out
更新中的Deployment、ReplicaSet、Pod信息
1 [root@k8s-master controller]# kubectl get deployment -o wide --show-labels 2 NAME READY UP-TO-DATE AVAILABLE AGE CONTAINERS IMAGES SELECTOR LABELS 3 nginx-deployment 3/3 1 3 12m nginx registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.5 app=nginx app=nginx 4 [root@k8s-master controller]# 5 [root@k8s-master controller]# kubectl get rs -o wide --show-labels 6 NAME DESIRED CURRENT READY AGE CONTAINERS IMAGES SELECTOR LABELS 7 nginx-deployment-56d78686f5 1 1 0 23s nginx registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.5 app=nginx,pod-template-hash=56d78686f5 app=nginx,pod-template-hash=56d78686f5 8 nginx-deployment-76b9d6bcf5 3 3 3 12m nginx registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.1 app=nginx,pod-template-hash=76b9d6bcf5 app=nginx,pod-template-hash=76b9d6bcf5 9 [root@k8s-master controller]# 10 [root@k8s-master controller]# kubectl get pod -o wide --show-labels 11 NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES LABELS 12 nginx-deployment-56d78686f5-4kn4c 0/1 ContainerCreating 0 30s <none> k8s-node02 <none> <none> app=nginx,pod-template-hash=56d78686f5 13 nginx-deployment-76b9d6bcf5-7lcr9 1/1 Running 0 12m 10.244.4.41 k8s-node01 <none> <none> app=nginx,pod-template-hash=76b9d6bcf5 14 nginx-deployment-76b9d6bcf5-jbb5h 1/1 Running 0 12m 10.244.2.48 k8s-node02 <none> <none> app=nginx,pod-template-hash=76b9d6bcf5 15 nginx-deployment-76b9d6bcf5-rt4m7 1/1 Running 0 12m 10.244.4.42 k8s-node01 <none> <none> app=nginx,pod-template-hash=76b9d6bcf5
更新成功后的Deployment、ReplicaSet、Pod信息
1 [root@k8s-master controller]# kubectl get deployment -o wide --show-labels 2 NAME READY UP-TO-DATE AVAILABLE AGE CONTAINERS IMAGES SELECTOR LABELS 3 nginx-deployment 3/3 3 3 15m nginx registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.5 app=nginx app=nginx 4 [root@k8s-master controller]# 5 [root@k8s-master controller]# kubectl get rs -o wide --show-labels 6 NAME DESIRED CURRENT READY AGE CONTAINERS IMAGES SELECTOR LABELS 7 nginx-deployment-56d78686f5 3 3 3 3m23s nginx registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.5 app=nginx,pod-template-hash=56d78686f5 app=nginx,pod-template-hash=56d78686f5 8 nginx-deployment-76b9d6bcf5 0 0 0 15m nginx registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.1 app=nginx,pod-template-hash=76b9d6bcf5 app=nginx,pod-template-hash=76b9d6bcf5 9 [root@k8s-master controller]# 10 [root@k8s-master controller]# kubectl get pod -o wide --show-labels 11 NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES LABELS 12 nginx-deployment-56d78686f5-4kn4c 1/1 Running 0 3m25s 10.244.2.49 k8s-node02 <none> <none> app=nginx,pod-template-hash=56d78686f5 13 nginx-deployment-56d78686f5-khsnm 1/1 Running 0 100s 10.244.2.50 k8s-node02 <none> <none> app=nginx,pod-template-hash=56d78686f5 14 nginx-deployment-56d78686f5-t24qw 1/1 Running 0 2m44s 10.244.4.43 k8s-node01 <none> <none> app=nginx,pod-template-hash=56d78686f5
通过查询Deployment详情，知晓pod替换过程
1 [root@k8s-master controller]# kubectl describe deploy nginx-deployment 2 Name: nginx-deployment 3 Namespace: default 4 CreationTimestamp: Thu, 28 May 2020 00:04:09 +0800 5 Labels: app=nginx 6 Annotations: deployment.kubernetes.io/revision: 2 7 kubectl.kubernetes.io/last-applied-configuration: 8 {"apiVersion":"apps/v1","kind":"Deployment","metadata":{"annotations":{"kubernetes.io/change-cause":"kubectl apply --filename=nginx-deploy... 9 kubernetes.io/change-cause: 10 kubectl set image deployment/nginx-deployment nginx=registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.5 --record=true 11 Selector: app=nginx 12 Replicas: 3 desired | 3 updated | 3 total | 3 available | 0 unavailable 13 StrategyType: RollingUpdate 14 MinReadySeconds: 0 15 RollingUpdateStrategy: 25% max unavailable, 25% max surge 16 Pod Template: 17 Labels: app=nginx 18 Containers: 19 nginx: 20 Image: registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.5 21 Port: 80/TCP 22 Host Port: 0/TCP 23 Environment: <none> 24 Mounts: <none> 25 Volumes: <none> 26 Conditions: 27 Type Status Reason 28 ---- ------ ------ 29 Available True MinimumReplicasAvailable 30 Progressing True NewReplicaSetAvailable 31 OldReplicaSets: <none> 32 NewReplicaSet: nginx-deployment-56d78686f5 (3/3 replicas created) 33 Events: 34 Type Reason Age From Message 35 ---- ------ ---- ---- ------- 36 Normal ScalingReplicaSet 93s deployment-controller Scaled up replica set nginx-deployment-76b9d6bcf5 to 3 37 Normal ScalingReplicaSet 38s deployment-controller Scaled up replica set nginx-deployment-56d78686f5 to 1 38 Normal ScalingReplicaSet 37s deployment-controller Scaled down replica set nginx-deployment-76b9d6bcf5 to 2 39 Normal ScalingReplicaSet 37s deployment-controller Scaled up replica set nginx-deployment-56d78686f5 to 2 40 Normal ScalingReplicaSet 35s deployment-controller Scaled down replica set nginx-deployment-76b9d6bcf5 to 1 41 Normal ScalingReplicaSet 35s deployment-controller Scaled up replica set nginx-deployment-56d78686f5 to 3 42 Normal ScalingReplicaSet 34s deployment-controller Scaled down replica set nginx-deployment-76b9d6bcf5 to 0
多 Deployment 动态更新
当 Deployment 正在展开进行更新时，Deployment 会为每个更新创建一个新的 ReplicaSet 并开始向上扩展，之前的 ReplicaSet 会被添加到旧 ReplicaSets 队列并开始向下扩展。
例如，假设创建一个 Deployment 以创建 nginx:1.7.9 的 5 个副本，然后更新 Deployment 以创建 5 个 nginx:1.9.1 的副本，而此时只有 3 个nginx:1.7.9 的副本已创建。在这种情况下， Deployment 会立即开始杀死3个 nginx:1.7.9 Pods，并开始创建 nginx:1.9.1 Pods。它不等待 nginx:1.7.9 的 5 个副本完成后再更新为nginx:1.9.1。
回滚 Deployment
yaml文件方式
针对应用的每个镜像版本，都有对应deploy的yaml文件。不管是升级还是回滚都已轻松应对。如下
1 nginx-deployment-1.15.6.yaml 2 nginx-deployment-1.17.yaml 3 nginx-deployment-1.17.1.yaml 4 nginx-deployment-1.17.5.yaml
yaml文件中的信息，参考上文即可。
命令行方式
问题产生
假设在更新 Deployment 时犯了一个拼写错误，将镜像名称命名为了 nginx:1.1710 而不是 nginx:1.17.10
1 [root@k8s-master controller]# kubectl set image deployment/nginx-deployment nginx=registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.1710 --record 2 deployment.apps/nginx-deployment image updated
查看Deployment、ReplicaSet、Pod信息
1 [root@k8s-master controller]# kubectl get deploy -o wide --show-labels 2 NAME READY UP-TO-DATE AVAILABLE AGE CONTAINERS IMAGES SELECTOR LABELS 3 nginx-deployment 3/3 1 3 14m nginx registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.1710 app=nginx app=nginx 4 [root@k8s-master controller]# 5 [root@k8s-master controller]# kubectl get rs -o wide --show-labels 6 NAME DESIRED CURRENT READY AGE CONTAINERS IMAGES SELECTOR LABELS 7 nginx-deployment-55c7bdfb86 3 3 3 9m19s nginx registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17 app=nginx,pod-template-hash=55c7bdfb86 app=nginx,pod-template-hash=55c7bdfb86 8 nginx-deployment-56d78686f5 0 0 0 12m nginx registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.5 app=nginx,pod-template-hash=56d78686f5 app=nginx,pod-template-hash=56d78686f5 9 nginx-deployment-76b9d6bcf5 0 0 0 13m nginx registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.1 app=nginx,pod-template-hash=76b9d6bcf5 app=nginx,pod-template-hash=76b9d6bcf5 10 nginx-deployment-844d7bbb7f 1 1 0 64s nginx registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.1710 app=nginx,pod-template-hash=844d7bbb7f app=nginx,pod-template-hash=844d7bbb7f 11 [root@k8s-master controller]# 12 [root@k8s-master controller]# kubectl get pod -o wide --show-labels 13 NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES LABELS 14 nginx-deployment-55c7bdfb86-bwzk9 1/1 Running 0 10m 10.244.4.49 k8s-node01 <none> <none> app=nginx,pod-template-hash=55c7bdfb86 15 nginx-deployment-55c7bdfb86-cmvzg 1/1 Running 0 10m 10.244.2.55 k8s-node02 <none> <none> app=nginx,pod-template-hash=55c7bdfb86 16 nginx-deployment-55c7bdfb86-kjrrw 1/1 Running 0 10m 10.244.2.56 k8s-node02 <none> <none> app=nginx,pod-template-hash=55c7bdfb86 17 nginx-deployment-844d7bbb7f-pctwr 0/1 ImagePullBackOff 0 2m3s 10.244.4.51 k8s-node01 <none> <none> app=nginx,pod-template-hash=844d7bbb7f
需求：回滚到以前稳定的 Deployment 版本。
操作步骤如下：
检查 Deployment 修改历史
1 [root@k8s-master controller]# kubectl rollout history deployment/nginx-deployment 2 deployment.apps/nginx-deployment 3 REVISION CHANGE-CAUSE 4 1 kubectl apply --filename=nginx-deployment.yaml --record=true 5 2 kubectl set image deployment/nginx-deployment nginx=registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.5 --record=true 6 3 kubectl set image deployment/nginx-deployment nginx=registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17 --record=true 7 4 kubectl set image deployment/nginx-deployment nginx=registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.1710 --record=true
查看修改历史的详细信息，运行
1 [root@k8s-master controller]# kubectl rollout history deployment/nginx-deployment --revision=3 2 deployment.apps/nginx-deployment with revision #3 3 Pod Template: 4 Labels: app=nginx 5 pod-template-hash=55c7bdfb86 6 Annotations: kubernetes.io/change-cause: 7 kubectl set image deployment/nginx-deployment nginx=registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17 --record=true 8 Containers: 9 nginx: 10 Image: registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17 11 Port: 80/TCP 12 Host Port: 0/TCP 13 Environment: <none> 14 Mounts: <none> 15 Volumes: <none>
回滚到上一次修改（即版本 3）或指定版本
现在已决定撤消当前更新并回滚到以前的版本
1 # 回滚到上一版本 2 [root@k8s-master controller]# kubectl rollout undo deployment/nginx-deployment 3 deployment.apps/nginx-deployment rolled back 4 # 回滚到指定历史版本 5 [root@k8s-master controller]# kubectl rollout undo deployment/nginx-deployment --to-revision=2 6 deployment.apps/nginx-deployment rolled back
检查回滚是否成功、 Deployment 是否正在运行
1 [root@k8s-master controller]# kubectl get deploy -o wide --show-labels 2 NAME READY UP-TO-DATE AVAILABLE AGE CONTAINERS IMAGES SELECTOR LABELS 3 nginx-deployment 3/3 3 3 17h nginx registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.5 app=nginx app=nginx
获取 Deployment 描述信息
1 [root@k8s-master controller]# kubectl describe deployment 2 Name: nginx-deployment 3 Namespace: default 4 CreationTimestamp: Thu, 28 May 2020 00:04:09 +0800 5 Labels: app=nginx 6 Annotations: deployment.kubernetes.io/revision: 7 7 kubectl.kubernetes.io/last-applied-configuration: 8 {"apiVersion":"apps/v1","kind":"Deployment","metadata":{"annotations":{"kubernetes.io/change-cause":"kubectl apply --filename=nginx-deploy... 9 kubernetes.io/change-cause: 10 kubectl set image deployment/nginx-deployment nginx=registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.5 --record=true 11 Selector: app=nginx 12 Replicas: 3 desired | 3 updated | 3 total | 3 available | 0 unavailable 13 StrategyType: RollingUpdate 14 MinReadySeconds: 0 15 RollingUpdateStrategy: 25% max unavailable, 25% max surge 16 Pod Template: 17 Labels: app=nginx 18 Containers: 19 nginx: 20 Image: registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.5 21 Port: 80/TCP 22 Host Port: 0/TCP 23 Environment: <none> 24 Mounts: <none> 25 Volumes: <none> 26 Conditions: 27 Type Status Reason 28 ---- ------ ------ 29 Available True MinimumReplicasAvailable 30 Progressing True NewReplicaSetAvailable 31 OldReplicaSets: <none> 32 NewReplicaSet: nginx-deployment-56d78686f5 (3/3 replicas created) 33 Events: 34 Type Reason Age From Message 35 ---- ------ ---- ---- ------- 36 ……………… 37 Normal ScalingReplicaSet 107s deployment-controller Scaled up replica set nginx-deployment-56d78686f5 to 1 38 Normal ScalingReplicaSet 104s deployment-controller Scaled down replica set nginx-deployment-55c7bdfb86 to 2 39 Normal ScalingReplicaSet 104s deployment-controller Scaled up replica set nginx-deployment-56d78686f5 to 2 40 Normal ScalingReplicaSet 103s deployment-controller Scaled down replica set nginx-deployment-55c7bdfb86 to 1 41 Normal ScalingReplicaSet 103s deployment-controller Scaled up replica set nginx-deployment-56d78686f5 to 3 42 Normal ScalingReplicaSet 102s deployment-controller Scaled down replica set nginx-deployment-55c7bdfb86 to 0
扩容/缩容Deployment
操作过程如下
1 [root@k8s-master controller]# kubectl scale deployment/nginx-deployment --replicas=10 2 deployment.apps/nginx-deployment scaled 3 [root@k8s-master controller]# kubectl get deploy -o wide --show-labels 4 NAME READY UP-TO-DATE AVAILABLE AGE CONTAINERS IMAGES SELECTOR LABELS 5 nginx-deployment 10/10 10 10 17h nginx registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.5 app=nginx app=nginx 6 [root@k8s-master controller]# 7 [root@k8s-master controller]# kubectl get rs -o wide --show-labels 8 NAME DESIRED CURRENT READY AGE CONTAINERS IMAGES SELECTOR LABELS 9 nginx-deployment-55c7bdfb86 0 0 0 17h nginx registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17 app=nginx,pod-template-hash=55c7bdfb86 app=nginx,pod-template-hash=55c7bdfb86 10 nginx-deployment-56d78686f5 10 10 10 17h nginx registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.5 app=nginx,pod-template-hash=56d78686f5 app=nginx,pod-template-hash=56d78686f5 11 nginx-deployment-76b9d6bcf5 0 0 0 17h nginx registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.17.1 app=nginx,pod-template-hash=76b9d6bcf5 app=nginx,pod-template-hash=76b9d6bcf5 12 nginx-deployment-844d7bbb7f 0 0 0 17h nginx registry.cn-beijing.aliyuncs.com/google_registry/nginx:1.1710 app=nginx,pod-template-hash=844d7bbb7f app=nginx,pod-template-hash=844d7bbb7f 13 [root@k8s-master controller]# 14 [root@k8s-master controller]# kubectl get pod -o wide --show-labels 15 NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES LABELS 16 nginx-deployment-56d78686f5-4v5mj 1/1 Running 0 44s 10.244.2.64 k8s-node02 <none> <none> app=nginx,pod-template-hash=56d78686f5 17 nginx-deployment-56d78686f5-8m7mx 1/1 Running 0 44s 10.244.4.60 k8s-node01 <none> <none> app=nginx,pod-template-hash=56d78686f5 18 nginx-deployment-56d78686f5-c7wlb 1/1 Running 0 44s 10.244.4.59 k8s-node01 <none> <none> app=nginx,pod-template-hash=56d78686f5 19 nginx-deployment-56d78686f5-jg5lt 1/1 Running 0 44s 10.244.2.63 k8s-node02 <none> <none> app=nginx,pod-template-hash=56d78686f5 20 nginx-deployment-56d78686f5-jj58d 1/1 Running 0 11m 10.244.4.56 k8s-node01 <none> <none> app=nginx,pod-template-hash=56d78686f5 21 nginx-deployment-56d78686f5-k2kts 1/1 Running 0 11m 10.244.4.57 k8s-node01 <none> <none> app=nginx,pod-template-hash=56d78686f5 22 nginx-deployment-56d78686f5-qltkv 1/1 Running 0 44s 10.244.2.61 k8s-node02 <none> <none> app=nginx,pod-template-hash=56d78686f5 23 nginx-deployment-56d78686f5-r7vmm 1/1 Running 0 11m 10.244.2.60 k8s-node02 <none> <none> app=nginx,pod-template-hash=56d78686f5 24 nginx-deployment-56d78686f5-rxlpm 1/1 Running 0 44s 10.244.2.62 k8s-node02 <none> <none> app=nginx,pod-template-hash=56d78686f5 25 nginx-deployment-56d78686f5-vlzrf 1/1 Running 0 44s 10.244.4.58 k8s-node01 <none> <none> app=nginx,pod-template-hash=56d78686f5
清理策略Policy
可以在 Deployment 中设置 .spec.revisionHistoryLimit，以指定保留多少该 Deployment 的 ReplicaSets数量。其余的将在后台进行垃圾回收。默认情况下，是10。
注意：此字段设置为 0 将导致清理 Deployment 的所有历史记录，因此 Deployment 将无法通过命令行回滚。
相关阅读
1、Kubernetes K8S之kubectl命令详解及常用示例
完毕！
———END———
如果觉得不错就关注下呗 (-^O^-) ！
