K8s新手系列之Endponit

概述

官方文档：https://kubernetes.io/zh-cn/docs/reference/kubernetes-api/service-resources/endpoints-v1/

Endpoint简称ep

Endpoint是kubernetes中的一个资源对象，存储在etcd中，用来记录一个service对应的所有pod的访问地址，它是根据service配置文件中selector描述产生的。

一个Service由一组Pod组成，这些Pod通过Endpoints暴露出来，Endpoints是实现实际服务的端点集合。换句话说，service和pod之间的联系是通过endpoints实现的。

负载分发策略

对Service的访问被分发到了后端的Pod上去，目前kubernetes提供了两种负载分发策略：

• 如果不定义，默认使用kube-proxy的策略，比如随机、轮询
• 基于客户端地址的会话保持模式，即来自同一个客户端发起的所有请求都会转发到固定的一个Pod上，此模式可以使在spec中添加sessionAffinity:ClientIP选项

验证Endpoint

创建service和deploy

[root@master01 ~/ingress]# cat nginx-deploy.yaml
# service
apiVersion: v1
kind: Service
metadata:
  namespace: default
  name: nginx-svc-clusterip
spec:
  type: ClusterIP
  selector:
    app: nginx
  ports:
  - name: clusterip-nginx
    port: 80
    targetPort: 80
    protocol: TCP

---
# deploy
apiVersion: apps/v1
kind: Deployment
metadata:
  name: deployment-nginx
  namespace: default
spec:
  replicas: 3
  selector:
    matchLabels:
      app: nginx
  template:
    metadata:
      name: pod-nginx
      labels:
        app: nginx
    spec:
      containers:
      - name: container-nginx
        image: nginx:1.14.1
      restartPolicy: Always

查看创建成功对应的资源

# 查看pod
[root@master01 ~/ingress]# kubectl get po -o wide
NAME                                READY   STATUS             RESTARTS         AGE    IP                NODE     NOMINATED NODE   READINESS GATES
deployment-nginx-6d84458cd8-g8lkv   1/1     Running            0                35m    100.117.144.139   node01   <none>           <none>
deployment-nginx-6d84458cd8-j8m6c   1/1     Running            0                35m    100.95.185.234    node02   <none>           <none>
deployment-nginx-6d84458cd8-znr7t   1/1     Running            0                35m    100.117.144.140   node01   <none>           <none>

# 查看svc详情，注意Endpoints列表是和pod的IP保持一致的
[root@master01 ~/ingress]# kubectl describe svc nginx-svc-clusterip
Name:              nginx-svc-clusterip
Namespace:         default
Labels:            <none>
Annotations:       <none>
Selector:          app=nginx
Type:              ClusterIP
IP Family Policy:  SingleStack
IP Families:       IPv4
IP:                10.96.1.52
IPs:               10.96.1.52
Port:              clusterip-nginx  80/TCP
TargetPort:        80/TCP
Endpoints:         100.117.144.139:80,100.117.144.140:80,100.95.185.234:80
Session Affinity:  None
Events:            <none>

# 查看endpoints资源
[root@master01 ~/ingress]# kubectl get endpoints
NAME                  ENDPOINTS                                                 AGE
nginx-svc-clusterip   100.117.144.139:80,100.117.144.140:80,100.95.185.234:80   36m

当我们重启Pod之后看看Endpoints列表会发生什么

重启Pod

[root@master01 ~/ingress]# kubectl delete po deployment-nginx-6d84458cd8-g8lkv deployment-nginx-6d84458cd8-j8m6c deployment-nginx-6d84458cd8-znr7t
pod "deployment-nginx-6d84458cd8-g8lkv" deleted
pod "deployment-nginx-6d84458cd8-j8m6c" deleted
pod "deployment-nginx-6d84458cd8-znr7t" deleted

查看资源，发现对应的IP是会进行变化的

# 查看Pod
[root@master01 ~/ingress]# kubectl get po -o wide
NAME                                READY   STATUS             RESTARTS         AGE    IP                NODE     NOMINATED NODE   READINESS GATES
deployment-nginx-6d84458cd8-4z4cb   1/1     Running            0                7s     100.95.185.236    node02   <none>           <none>
deployment-nginx-6d84458cd8-ht2z9   1/1     Running            0                7s     100.117.144.141   node01   <none>           <none>
deployment-nginx-6d84458cd8-ns47j   1/1     Running            0                7s     100.95.185.238    node02   <none>           <none>
# 查看service
[root@master01 ~/ingress]# kubectl describe svc nginx-svc-clusterip
Name:              nginx-svc-clusterip
Namespace:         default
Labels:            <none>
Annotations:       <none>
Selector:          app=nginx
Type:              ClusterIP
IP Family Policy:  SingleStack
IP Families:       IPv4
IP:                10.96.1.52
IPs:               10.96.1.52
Port:              clusterip-nginx  80/TCP
TargetPort:        80/TCP
Endpoints:         100.117.144.141:80,100.95.185.236:80,100.95.185.238:80
Session Affinity:  None
Events:            <none>
#查看endpoints
[root@master01 ~/ingress]# kubectl get endpoints
NAME                  ENDPOINTS                                                AGE
kubernetes            10.0.0.30:6443                                           78m
nginx-svc-clusterip   100.117.144.141:80,100.95.185.236:80,100.95.185.238:80   39m

编写一个属于自己Endpoints

Endpoints资源是通过name和namespace两个字段与Service进行关联的，所以Endpoints的名称和Service的名称相同.

示例：

[root@master01 ~]# cat ep.yaml
# Endpoints资源
apiVersion: v1
kind: Endpoints
metadata:
    name: harbor-huangxin
subsets:
    - addresses:
      - ip: 10.0.0.250
      ports:
      - port: 80
        protocol: TCP
---
# Service资源
apiVersion: v1
kind: Service
metadata:
    name: harbor-huangxin
spec:
    type: ClusterIP
    ports:
        - port: 80
          targetPort: 80

查看一下资源

# 查看service
[root@master01 ~]# kubectl describe svc endpoint-huangsir
Name:              endpoint-huangsir
Namespace:         default
Labels:            <none>
Annotations:       <none>
Selector:          <none>
Type:              ClusterIP
IP Family Policy:  SingleStack
IP Families:       IPv4
IP:                10.96.1.127
IPs:               10.96.1.127
Port:              <unset>  80/TCP
TargetPort:        80/TCP
Endpoints:         10.0.0.250:80
Session Affinity:  None
Events:            <none>
# 查看endpoint
[root@master01 ~]# kubectl get ep endpoint-huangsir
NAME              ENDPOINTS       AGE
endpoint-huangsir   10.0.0.250:80   51s