什么是Pod

• Pod是K8S最小调度单位。
• Pod可以由一个或者多个容器组合而成。
• Pod容器共享Volume
• 很少会直接创建一个Pod,在大多数情况下，会通过RC,RS,Deployment,DaemonSet,Job等控制器完成对一组Pod副本的创建，调度以及生命周期的自动控制任务。

通过Deployment部署一个nginx

• 新建并且编辑nignx-deployment.yaml

apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx-deployment
spec:
  selector:
    matchLabels:
      app: nginx
  replicas: 3
  template: 
    metadata:
      labels:
        app: nginx
    spec:
      containers:
      - name: nginx
        image: nginx:latest
        ports:
        - containerPort: 80

• 创建Deployment

kubectl create -f nignx-deployment.yaml
# 执行一下命令可以查看结果
kubectl get deployments
kubectl get rs
kubectl get pod -o wide
# [root@k8s-master k8s-yaml]# kubectl get pods
# NAME                               READY   STATUS    RESTARTS   AGE
# nginx-deployment-59c9f8dff-kbdg6   1/1     Running   0          47s
# nginx-deployment-59c9f8dff-lwnj2   1/1     Running   0          47s
# nginx-deployment-59c9f8dff-whjt8   1/1     Running   0          47s

创建Service

apiVersion: v1
kind: Service
metadata:
  name: nginx
spec:
  ports:
  - port: 8888
    targetPort: 80
  selector:
    app:nginx

• NodePort类型：在node上打开一个端口，将该端口得流量导入到kube-proxy，然后kube-proxy进一步到对应得pod

apiVersion: v1
kind: Service
metadata: 
  name: nodeportservice
spec:
  type: NodePort
  selector:
    app: nginx
  ports:
  - port: 8889
    targetPort: 80

ConfigMap

• 生产为容器内的环境变量
• 设置容器启动命令的启动参数
• 以Volume的形式挂载为容器内部的文件或者目录

创建

• 通过YAML配置文件方式创建

# configmap.yaml
apiVersion: v1
kind: ConfigMap
metadata:
  name: cm-var
data: 
  level: info
  env: development

[root@k8s-master k8s-yaml]# kubectl create -f configmap.yaml 
configmap/cm-var created
[root@k8s-master k8s-yaml]# kubectl get cm
NAME     DATA   AGE
cm-var   2      20s
kubectl describe configmap

Pod中使用ConfigMap

# pod-configmap.yaml
apiVersion: v1
kind: Pod
metadata:
  name: cm-test-pod
spec:
  containers:
  - name: cn-test
    image: busybox
    command: ["/bin/sh", "-c", "env | grep | APP"]
    env:                            
    - name: APPLOGLEVEL      # 定义环境变量的名称
      valueFrom:
        configMapKeyRef:
          name: cm-var       # 环境变量的值取自以上configmap.yaml定义的cm-var
          key: level         # key值为level
    - name: APPENV
      valueFrom:
        configMapKeyRef:
          name: cm-var
          key: env

• k8s在1.6的版本开始，引入了一个新的字段envFrom

# pod-envFrom-configmap.yaml
apiVersion: v1
kind: Pod
metadata:
  name: cm-test-pod
spec:
  containers:
  - name: cn-test
    image: busybox
    command: ["/bin/sh", "-c", "env"]
    envFrom:
    - configMapRef:
      name: cm-var

限制条件

• ConfigMap必须在Pod之前创建
• Config受NameSpace的限制，必须同一个NameSpace下
• 支持被Api Server管理的Pod才能使用【比如静态pod无法使用】

容器内获取Pod的信息

# pod-dapi.yaml
apiVersion: v1
kind: Pod
metadata:
  name: cm-test-pod
spec:
  containers:
  - name: cn-test
    image: busybox
    command: ["/bin/sh", "-c", "env"]
    env:                            
    - name: podName      # 定义环境变量的名称
      valueFrom:
        fieldRef:
          podName: metadata.name
    - name: podId
      valueFrom:
        fieldRef:
          podId: status.podIP

Pod升级和回滚

• Deployment的升级

Nignx由1.7.9升级到1.9.1

# deployment-nginx.yaml
apiVersion: v1
kind: Deplopment
metadata:
  name: deployment-nginx
spec:
  selector:
    matchLabels:
      app: nginx
  replicas: 3
  template:
    metadata:
      labels:
        app: nginx
    spec:
      containers:
      - name: nginx
        image: nginx:1.7.9
        ports:
        - containersPort: 80

# 创建
# kubeclt create -f deployment-nginx.yaml
# 升级
# kebectl set image deplopment/deplopment-nginx nginx=nginx:1.9.1

• Deplopment的回滚

# 查看回滚命令
kubect rollout history deplopment/deployment-nginx
# 回滚到上一版本
kubectl rollout undo deployment/deployment-nginx

Pod的扩容

• 手动扩容

通过执行kubectl scale命令或者通过restful api对一个deployment/rc/rs进行pod的副本数量的设置。

• 以上面deployment-nginx.yaml为例[yaml中设置启动3个]

kubectl scale deplopment deployment-nginx --replicase = 5
# 如果设置的replicase 小于初始化的，会“杀掉”一些运行中的Pod.

• 自动扩容

需要用户根据某个性能指标或者自定义业务指标，并且指定pod的副本数量范围，系统将在这个范围内根据性能指标的变化进行调整。

• HPA的资源对象处于api的“autoscaling”中，目前有v1和v2两个版本，v1只支持基于CPU使用率的自动扩缩容。v2版本则支持基于任意指标的扩容配置，包括基于资源使用率，pod指标

• v1版本需要预先安装Heapster组件或者Metrics Server 用于采集pod的CPU使用率

apiVserion: autoscaling/v1
kind: HorizontalPodAutoscaler
metadata:
  name: v1-auto-nginx
spec:
  scaleTargetRef: # 目标作用对象，可以是deplopment,rs,rc
    apiVsersion: apps/v1
    kind: Deplopment
    name: deplopment-nginx
  minReplicas: 1
  maxReplicas: 10
  targetCPUUtilizationPercentage: 50 # 期望每个pod的cpu使用率都是50%

• v2版本

apiVersion: autoscaling/v2beta2
kind: HorizontalPodAutoscaler
metadata:
  name: v2-auto-nginx
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: deployment
    name: deplopment-nginx
  minReplicas: 1
  maxReplicas: 10
  metrics:
  - type: Resource
    resource:
      name: cpu
      target: 
        type: Utilization
        averageUtilization: 50