K8S | Service服务发现

0130

服务发现与负载均衡。

一、背景

在微服务架构中，这里以开发环境「Dev」为基础来描述，在K8S集群中通常会开放：路由网关、注册中心、配置中心等相关服务，可以被集群外部访问；

对于测试「Tes」环境或者生产「Pro」环境，出于安全或者环境隔离性来考虑，在正常情况下只会开放网关服务，而「注册、配置」中心并不会对外暴露；

对于架构中的其它业务服务一般不会对外开放，在K8S集群内部服务间是可以正常通信的，对于「Dev」环境来说，研发会使用「注册、配置」中心，网关是系统的访问入口；

在K8S集群中，通过Service组件，可以快速简单的实现服务发现和负载均衡；

二、Service组件

1、简介

在K8S集群中是通过Pod组件来部署应用服务，Deployment组件实现Pod编排管理，Service组件实现应用的访问；

【Pod】自身的特点是临时的，使用过后直接抛弃的实体，这样在Pod创建和销毁的状态中，会导致IP地址发生变化，即无法使用固定的IP进行应用访问；

【Deployment】控制器通过管理ReplicaSet间接实现Pod管理，比如发布方式，更新和回滚策略，维持Pod副本数量，对应用进行快速的编排，但是并没有涉及应用的访问；

【Service】是将运行在一个或一组Pod上的网络应用程序公开为网络服务的方法，可以在不修改现有应用程序的情况下，使用服务发现机制访问到该应用；

基于Pod、Deployment、Service三个组件的协作，同一个应用的部署脚本可以在开发、测试、生产不同环境中复用；

2、基础语法

这里提供一个简单的【Service】语法做参考；

需要注意的是：在该脚本中没有指定服务类型即ServiceType，默认采用的是ClusterIP，通过集群的内部IP暴露服务，选择该值时服务只能够在集群内部访问；

三、内部服务发现

1、Pod创建

基于【Deployment】组件，创建「auto-serve」应用；

apiVersion: apps/v1
kind: Deployment
metadata:
  name: serve-deployment
  labels:
    app: auto-serve
spec:
  replicas: 1
  selector:
    matchLabels:
      app: auto-serve
  template:
    metadata:
      labels:
        app: auto-serve
    spec:
      containers:
        - name: auto-serve
          image: auto-serve:latest
          imagePullPolicy: Never
          ports:
            - containerPort: 8082
              name: auto-serve-port
apiVersion: apps/v1
kind: Deployment
metadata:
  name: serve-deployment
  labels:
    app: auto-serve
spec:
  replicas: 1
  selector:
    matchLabels:
      app: auto-serve
  template:
    metadata:
      labels:
        app: auto-serve
    spec:
      containers:
        - name: auto-serve
          image: auto-serve:latest
          imagePullPolicy: Never
          ports:
            - containerPort: 8082
              name: auto-serve-port
apiVersion: apps/v1
kind: Deployment
metadata:
  name: serve-deployment
  labels:
    app: auto-serve
spec:
  replicas: 1
  selector:
    matchLabels:
      app: auto-serve
  template:
    metadata:
      labels:
        app: auto-serve
    spec:
      containers:
        - name: auto-serve
          image: auto-serve:latest
          imagePullPolicy: Never
          ports:
            - containerPort: 8082
              name: auto-serve-port

执行创建命令

kubectl apply -f serve-deployment.yaml
kubectl apply -f serve-deployment.yaml
kubectl apply -f serve-deployment.yaml

2、Service创建

简单的脚本文件：app-service.yaml；

apiVersion: v1
kind: Service
metadata:
  name: app-service
spec:
  selector:
    app: auto-serve
  ports:
  - name: app-service-port
    protocol: TCP
    port: 8082
    targetPort: auto-serve-port
apiVersion: v1
kind: Service
metadata:
  name: app-service
spec:
  selector:
    app: auto-serve
  ports:
  - name: app-service-port
    protocol: TCP
    port: 8082
    targetPort: auto-serve-port
apiVersion: v1
kind: Service
metadata:
  name: app-service
spec:
  selector:
    app: auto-serve
  ports:
  - name: app-service-port
    protocol: TCP
    port: 8082
    targetPort: auto-serve-port

创建【Service】

kubectl apply -f app-service.yaml
kubectl apply -f app-service.yaml
kubectl apply -f app-service.yaml

查看【Service】，可以使用命令行或者界面；

kubectl describe svc app-service
kubectl describe svc app-service
kubectl describe svc app-service

删除【Service】

kubectl delete -f app-service.yaml
kubectl delete -f app-service.yaml
kubectl delete -f app-service.yaml

3、内部访问

在上面已经说明，当Type不指定时采用的是ClusterIP，只能在集群内部访问，集群外部的网络是无法访问的；

在【auto-client】服务中提供一段访问【auto-serve】接口的代码，并制作镜像【auto-client:3.3.3】，完成部署后查看日志打印；

@Component
public class HttpServiceJob {
    private static final Logger LOG = LoggerFactory.getLogger(HttpServiceJob.class.getName()) ;
    private static final String SERVER_NAME = "http://app-service:8082/serve";
    private static final String SERVER_IP = "http://10.103.252.94:8082/serve";
    /**
     * 每30秒执行一次
     */
    @Scheduled(fixedDelay = 30000)
    public void systemDate () {
        SimpleClientHttpRequestFactory factory = new SimpleClientHttpRequestFactory();
        factory.setReadTimeout(3000);
        factory.setConnectTimeout(6000);
        RestTemplate restTemplate = new RestTemplate(factory);
        try {
            Map<String, String> paramMap = new HashMap<>();
            String result = restTemplate.getForObject(SERVER_NAME, String.class, paramMap);
            LOG.info("service-name-resp::::" + result);
        } catch (Exception e) {
            e.printStackTrace();
        }
        try {
            Map<String, String> paramMap = new HashMap<>();
            String result = restTemplate.getForObject(SERVER_IP, String.class, paramMap);
            LOG.info("service-ip-resp::::" + result);
        } catch (Exception e) {
            e.printStackTrace();
        }
    }
}
@Component
public class HttpServiceJob {

    private static final Logger LOG = LoggerFactory.getLogger(HttpServiceJob.class.getName()) ;

    private static final String SERVER_NAME = "http://app-service:8082/serve";
    private static final String SERVER_IP = "http://10.103.252.94:8082/serve";

    /**
     * 每30秒执行一次
     */
    @Scheduled(fixedDelay = 30000)
    public void systemDate () {
        SimpleClientHttpRequestFactory factory = new SimpleClientHttpRequestFactory();
        factory.setReadTimeout(3000);
        factory.setConnectTimeout(6000);
        RestTemplate restTemplate = new RestTemplate(factory);

        try {
            Map<String, String> paramMap = new HashMap<>();
            String result = restTemplate.getForObject(SERVER_NAME, String.class, paramMap);
            LOG.info("service-name-resp::::" + result);
        } catch (Exception e) {
            e.printStackTrace();
        }

        try {
            Map<String, String> paramMap = new HashMap<>();
            String result = restTemplate.getForObject(SERVER_IP, String.class, paramMap);
            LOG.info("service-ip-resp::::" + result);
        } catch (Exception e) {
            e.printStackTrace();
        }
    }
}
@Component
public class HttpServiceJob {

    private static final Logger LOG = LoggerFactory.getLogger(HttpServiceJob.class.getName()) ;

    private static final String SERVER_NAME = "http://app-service:8082/serve";
    private static final String SERVER_IP = "http://10.103.252.94:8082/serve";

    /**
     * 每30秒执行一次
     */
    @Scheduled(fixedDelay = 30000)
    public void systemDate () {
        SimpleClientHttpRequestFactory factory = new SimpleClientHttpRequestFactory();
        factory.setReadTimeout(3000);
        factory.setConnectTimeout(6000);
        RestTemplate restTemplate = new RestTemplate(factory);

        try {
            Map<String, String> paramMap = new HashMap<>();
            String result = restTemplate.getForObject(SERVER_NAME, String.class, paramMap);
            LOG.info("service-name-resp::::" + result);
        } catch (Exception e) {
            e.printStackTrace();
        }

        try {
            Map<String, String> paramMap = new HashMap<>();
            String result = restTemplate.getForObject(SERVER_IP, String.class, paramMap);
            LOG.info("service-ip-resp::::" + result);
        } catch (Exception e) {
            e.printStackTrace();
        }
    }
}

在代码中通过服务名:端口和IP:端口都可以正常访问，在Pod中查看两个应用的日志，请求和响应都正常；

四、外部服务发现

1、NodePort类型

指定类型为NodePort的脚本：app-np-service.yaml；

apiVersion: v1
kind: Service
metadata:
  name: app-np-service
spec:
  type: NodePort
  selector:
    app: auto-serve
  ports:
    - protocol: TCP
      port: 8082
      targetPort: 8082
      nodePort: 30010
apiVersion: v1
kind: Service
metadata:
  name: app-np-service
spec:
  type: NodePort
  selector:
    app: auto-serve
  ports:
    - protocol: TCP
      port: 8082
      targetPort: 8082
      nodePort: 30010
apiVersion: v1
kind: Service
metadata:
  name: app-np-service
spec:
  type: NodePort
  selector:
    app: auto-serve
  ports:
    - protocol: TCP
      port: 8082
      targetPort: 8082
      nodePort: 30010

创建【Service】

kubectl apply -f app-np-service.yaml
kubectl apply -f app-np-service.yaml
kubectl apply -f app-np-service.yaml

使用NodePort类型，K8S控制平面会在指定的范围内分配端口，如果需要特定的端口号可以指定nodePort字段中的值，但是该类型需要自己设置负载均衡解决方案；

2、LoadBalancer类型

指定类型为LoadBalancer的脚本：app-lb-service.yaml；

apiVersion: v1
kind: Service
metadata:
  name: app-lb-service
spec:
  type: LoadBalancer
  selector:
    app: auto-serve
  ports:
    - protocol: TCP
      port: 8082
      targetPort: 8082
apiVersion: v1
kind: Service
metadata:
  name: app-lb-service
spec:
  type: LoadBalancer
  selector:
    app: auto-serve
  ports:
    - protocol: TCP
      port: 8082
      targetPort: 8082
apiVersion: v1
kind: Service
metadata:
  name: app-lb-service
spec:
  type: LoadBalancer
  selector:
    app: auto-serve
  ports:
    - protocol: TCP
      port: 8082
      targetPort: 8082

创建【Service】

kubectl apply -f app-lb-service.yaml
kubectl apply -f app-lb-service.yaml
kubectl apply -f app-lb-service.yaml

查看【Service】

在查看「app-lb-service」时，值得注意一下Endpoints的字段属性，这里就是Pod选择器选中的Pod；

kubectl get svc app-lb-service -o wide
NAME             TYPE           CLUSTER-IP      EXTERNAL-IP   PORT(S)          AGE     SELECTOR
app-lb-service   LoadBalancer   10.111.65.220   localhost     8082:30636/TCP   6m49s   app=auto-serve
kubectl describe svc app-lb-service
Name:                     app-lb-service
Namespace:                default
Labels:                   <none>
Annotations:              <none>
Selector:                 app=auto-serve
Type:                     LoadBalancer
IP Family Policy:         SingleStack
IP Families:              IPv4
IP:                       10.111.65.220
IPs:                      10.111.65.220
LoadBalancer Ingress:     localhost
Port:                     <unset>  8082/TCP
TargetPort:               8082/TCP
NodePort:                 <unset>  30636/TCP
Endpoints:                10.1.0.160:8082,10.1.0.161:8082,10.1.0.162:8082
Session Affinity:         None
External Traffic Policy:  Cluster
Events:                   <none>
kubectl get pods -o wide
NAME                               READY   STATUS    RESTARTS   AGE   IP           NODE          
serve-deployment-f6f6c5bbd-9qvgr   1/1     Running   0          39m   10.1.0.162   docker-desktop
serve-deployment-f6f6c5bbd-w7nj2   1/1     Running   0          39m   10.1.0.161   docker-desktop
serve-deployment-f6f6c5bbd-x7v4d   1/1     Running   0          39m   10.1.0.160   docker-desktop
kubectl get svc app-lb-service -o wide


NAME             TYPE           CLUSTER-IP      EXTERNAL-IP   PORT(S)          AGE     SELECTOR
app-lb-service   LoadBalancer   10.111.65.220   localhost     8082:30636/TCP   6m49s   app=auto-serve


kubectl describe svc app-lb-service


Name:                     app-lb-service
Namespace:                default
Labels:                   <none>
Annotations:              <none>
Selector:                 app=auto-serve
Type:                     LoadBalancer
IP Family Policy:         SingleStack
IP Families:              IPv4
IP:                       10.111.65.220
IPs:                      10.111.65.220
LoadBalancer Ingress:     localhost
Port:                     <unset>  8082/TCP
TargetPort:               8082/TCP
NodePort:                 <unset>  30636/TCP
Endpoints:                10.1.0.160:8082,10.1.0.161:8082,10.1.0.162:8082
Session Affinity:         None
External Traffic Policy:  Cluster
Events:                   <none>


kubectl get pods -o wide


NAME                               READY   STATUS    RESTARTS   AGE   IP           NODE          
serve-deployment-f6f6c5bbd-9qvgr   1/1     Running   0          39m   10.1.0.162   docker-desktop
serve-deployment-f6f6c5bbd-w7nj2   1/1     Running   0          39m   10.1.0.161   docker-desktop
serve-deployment-f6f6c5bbd-x7v4d   1/1     Running   0          39m   10.1.0.160   docker-desktop
kubectl get svc app-lb-service -o wide


NAME             TYPE           CLUSTER-IP      EXTERNAL-IP   PORT(S)          AGE     SELECTOR
app-lb-service   LoadBalancer   10.111.65.220   localhost     8082:30636/TCP   6m49s   app=auto-serve


kubectl describe svc app-lb-service


Name:                     app-lb-service
Namespace:                default
Labels:                   <none>
Annotations:              <none>
Selector:                 app=auto-serve
Type:                     LoadBalancer
IP Family Policy:         SingleStack
IP Families:              IPv4
IP:                       10.111.65.220
IPs:                      10.111.65.220
LoadBalancer Ingress:     localhost
Port:                     <unset>  8082/TCP
TargetPort:               8082/TCP
NodePort:                 <unset>  30636/TCP
Endpoints:                10.1.0.160:8082,10.1.0.161:8082,10.1.0.162:8082
Session Affinity:         None
External Traffic Policy:  Cluster
Events:                   <none>


kubectl get pods -o wide


NAME                               READY   STATUS    RESTARTS   AGE   IP           NODE          
serve-deployment-f6f6c5bbd-9qvgr   1/1     Running   0          39m   10.1.0.162   docker-desktop
serve-deployment-f6f6c5bbd-w7nj2   1/1     Running   0          39m   10.1.0.161   docker-desktop
serve-deployment-f6f6c5bbd-x7v4d   1/1     Running   0          39m   10.1.0.160   docker-desktop

五、参考源码

文档仓库：
https://gitee.com/cicadasmile/butte-java-note
脚本仓库：
https://gitee.com/cicadasmile/butte-auto-parent
文档仓库：
https://gitee.com/cicadasmile/butte-java-note

脚本仓库：
https://gitee.com/cicadasmile/butte-auto-parent
文档仓库：
https://gitee.com/cicadasmile/butte-java-note

脚本仓库：
https://gitee.com/cicadasmile/butte-auto-parent

© 版权声明
文章版权归作者所有，未经允许请勿转载，侵权请联系 admin@trc20.tw 删除。
THE END
博客文章
# 后端# Java# Kubernetes# K8s
喜欢就支持一下吧
点赞0

Warning: mysqli_query(): (HY000/3): Error writing file '/tmp/MYHATnsh' (Errcode: 28 - No space left on device) in /www/wwwroot/583.cn/wp-includes/class-wpdb.php on line 2345
admin的头像-五八三
相关推荐
评论 抢沙发
头像
欢迎您留下宝贵的见解！
提交
头像

昵称

图形验证码
取消
昵称

请填写用户信息：

  • 昵称(必填)
  • 邮箱(必填)
代码

请输入代码：

确认
图片
文章目录
文章目录
热门推荐