k8s部署eureka集群

发表于   |   更新于 | 分类于 | | 阅读次数: | 总览

eureka 不阐述介绍，这里直接开始在kubernetes下部署eureka集群

1、配置文件的增加

eureka 只一个有状态的服务，部署有状态服务我们可以使用StatefulSet

1.1、增加dockerfile

1
2
3
4
5
6
7
8
9
10
11
12
13
$ cat Dockerfile
FROM docker.io/xxlaila/centos7.6-jdk1.8:latest
MAINTAINER xxlaila "cq_xxlaila@163.com"
# Install dependent plugin

ADD target/kxl-eureka.jar /opt/webapps/kxl-eureka.jar
ADD application.yaml /opt/webapps/application.yaml

WORKDIR /opt/webapps

EXPOSE 8080

ENTRYPOINT ["java", "-jar", "-Dspring.profiles.active=dev", "kxl-eureka.jar"]

1.2、修改配置文件

在做eureka集群的时候，application.yaml的配置文件很重要，配置文件做不好，将会直接影响到eureka的启动，还有集群的模式

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
$ cat application.yaml
spring:
  application:
    name: kxl-eureka
server:
  port: 8761
eureka:
  instance:
#    prefer-ip-address: true
    hostname: ${EUREKA_HOST_NAME:peer1} #服务主机名
    appname: ${spring.application.name}  #服务名称，默认为 unknow 这里直接取 spring.application.name 了
    # server 从最后一次收到心跳到移除废弃服务的超时时间（秒）
    lease-expiration-duration-in-seconds: 90
    # client 给 server 发送心跳的间隔时间（秒），比 lease-expiration-duration-in-seconds 小
    lease-renewal-interval-in-seconds: 30

  client:
    serviceUrl:
      defaultZone: ${EUREKA_URL_LIST:http://peer1:8761/eureka/} # 指定服务中心 eureka server的地址
    # client 是否从eureka上拉取注册信息， server模式可关掉
    fetch-registry: ${BOOL_FETCH:true}   # 是否拉取 eureka server 的注册信息。 默认为true
    # client 是否注册到eureka上， server模式可关掉
    register-with-eureka: ${BOOL_REGISTER:true}  # 是否把服务中心本身当做eureka client 注册。默认为true
    # client 间隔多久去拉去服务信息(秒)
    registry-fetch-interval-seconds: 30
  server:
    # 自我保护机制，应对网络闪断情况，大面积丢失过多的client，不删除服务
    enable-self-preservation: ${SELF_PRESERVATION:true}     # 是否开启自我保护。 默认为 true.
    # 每分钟心跳数 实际/期望，如果小于阈值(threshold)，则触发自我保护机制
    renewal-percent-threshold: 0.85
    # 扫描失效服务的间隔时间（毫秒）
    eviction-interval-timer-in-ms: 60000

  application:
    name: ${EUREKA_APPLICATION_NAME:eureka-server}

1.3、创建 eureka docker镜像

1
2
$ docker build -t xxlaila/kxl-eureka:v1 .
$ docker push xxlaila/kxl-eureka:v1

2、在k8s创建eureka集群

2.1、创建eureka集群

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
$ cat kxl-eureka.yaml
---
apiVersion: v1
kind: Service
metadata:
  name: eureka
  namespace: kube-dev
  labels:
    app: eureka
spec:
  ports:
  - port: 8761
    name: eureka
  clusterIP: None
  selector:
    app: eureka
---
apiVersion: apps/v1beta1
kind: StatefulSet
metadata:
  name: eureka
  namespace: kube-dev
spec:
  serviceName: "eureka"
  replicas: 3
  selector:
    matchLabels:
      app: eureka
  template:
    metadata:
      labels:
        app: eureka
    spec:
      containers:
      - name: eureka
        image:  docker.io/xxlaila/kxl-eureka:v1
        ports:
        - containerPort: 8761
        env:
        - name: MY_POD_NAME
          valueFrom:
            fieldRef:
              fieldPath: metadata.name
          # Due to camelcase issues with "defaultZone" and "preferIpAddress", _JAVA_OPTIONS is used here
        - name: eureka_client_serviceUrl_defaultZone
          #value: http://eureka-0.eureka.<namespace>.svc.cluster.local:8761/eureka/,http://eureka-1.eureka.<namespace>.svc.cluster.local:8761/eureka/
          value: http://eureka-0.eureka:8761/eureka/,http://eureka-1.eureka:8761/eureka/,http://eureka-2.eureka:8761/eureka/
        - name: EUREKA_CLIENT_REGISTERWITHEUREKA
          value: "true"
        - name: EUREKA_CLIENT_FETCHREGISTRY
          value: "true"
          # In the docker image, this is set to localhost. Otherwise, we could leave this empty.
          # The hostnames must match with the the eureka serviceUrls, otherwise the replicas are reported as unavailable in the eureka dashboard
        - name: EUREKA_INSTANCE_HOSTNAME
          #value: "$(MY_POD_NAME).eureka.<namespace.svc.cluster.local"
          value: "$(MY_POD_NAME).eureka"
          #value: eureka
          # For the other (stateless) services, this should probably be set to true, since their pods have no DNS-resolvable  hostnames
       #- name: EUREKA_INSTANCE_PREFERIPADDRESS
       #  value: "false"
  # No need to start the pods in order. We just need the stable network identity
  podManagementPolicy: "Parallel"

在做这个的时候其实遇到了很多坑，也是参考一些文章才完成的,参考文献

2.2、创建eureka Ingress

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
$ cat eureka-ingress.yaml
apiVersion: extensions/v1beta1
kind: Ingress
metadata:
  name: eureka-ingress
  namespace: kube-dev
  annotations:
    kubernetes.io/ingress.class: "nginx"
    nginx.ingress.kubernetes.io/secure-backends: "true"
spec:
  rules:
  - host: eureka.xxlaila.io
    http:
      paths:
      - backend:
          serviceName: eureka
          servicePort: 8761

2.3、执行创建

1
2
3
4
5
6
7
8
9
10
11
12
13
14
$ kubectl apply -f .
$ kubectl get pods -n kube-dev
NAME       READY   STATUS    RESTARTS   AGE
eureka-0   1/1     Running   0          21m
eureka-1   1/1     Running   0          21m
eureka-2   1/1     Running   0          21m
$ kubectl get pods,svc,rs -n kube-dev
NAME           READY   STATUS    RESTARTS   AGE
pod/eureka-0   1/1     Running   0          21m
pod/eureka-1   1/1     Running   0          21m
pod/eureka-2   1/1     Running   0          21m

NAME             TYPE        CLUSTER-IP   EXTERNAL-IP   PORT(S)    AGE
service/eureka   ClusterIP   None         <none>        8761/TCP   21m

2.4、访问验证

img
通过域名访问：
img

3、eureka 环境

    根据2.4小节可以看到，enviroonment为test，我们在dockerfile指定的为dev，所以这里就有点差池，但是查看了一下资料，这个要么就写多个application.yaml的配置文件，要么就打多个包，这样就比较麻烦，而且考虑到公司微服务的特殊性，既要满足于公司的微服务架构，有要考虑的模版的通用性，还需要考虑运维维护的便捷性。下面一起来看看基于公司的定制化来改变这个局限性。

3.1、公司的系统环境变量

1
2
3
4
5
6
7
8
CONFIG_API_SERVER=http://api.conf.xxlaila.io
RUN_CLUSTER=default
RUN_MODE=AUTO
RUN_ENV=demo
CONFIG_API_SERVER：公司配置中心api的地址，app拉取配置中心的配置
RUN_CLUSTER：默认集群，做灰度发布使用
RUN_MODE：
RUN_ENV：当前系统所运行的环境

3.2、修改eureka的配置文件

增加配置eureka:environment: ${RUN_ENV}

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
$ cat application.yaml
spring:
  application:
    name: kxl-eureka
server:
  port: 8761
eureka:
  environment: ${RUN_ENV}
  instance:
#    prefer-ip-address: true
    hostname: ${EUREKA_HOST_NAME:peer1}
    appname: ${spring.application.name}
    # server 从最后一次收到心跳到移除废弃服务的超时时间（秒）
    lease-expiration-duration-in-seconds: 90
    # client 给 server 发送心跳的间隔时间（秒），比 lease-expiration-duration-in-seconds 小
    lease-renewal-interval-in-seconds: 30

  client:
    serviceUrl:
      defaultZone: ${EUREKA_URL_LIST:http://peer1:8761/eureka/}
    # client 是否从eureka上拉取注册信息， server模式可关掉
    fetch-registry: ${BOOL_FETCH:true}
    # client 是否注册到eureka上， server模式可关掉
    register-with-eureka: ${BOOL_REGISTER:true}
    # client 间隔多久去拉去服务信息(秒)
    registry-fetch-interval-seconds: 30
  server:
    # 自我保护机制，应对网络闪断情况，大面积丢失过多的client，不删除服务
    enable-self-preservation: ${SELF_PRESERVATION:true}
    # 每分钟心跳数 实际/期望，如果小于阈值(threshold)，则触发自我保护机制
    renewal-percent-threshold: 0.85
    # 扫描失效服务的间隔时间（毫秒）
    eviction-interval-timer-in-ms: 60000

  application:
    name: ${EUREKA_APPLICATION_NAME:eureka-server}

3.3、修改eureka部署文件

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
$ cat kxl-eureka.yaml 
---
apiVersion: v1
kind: Service
metadata:
  name: eureka
  namespace: kube-dev
  labels:
    app: eureka
spec:
  ports:
  - port: 8761
    name: eureka
  clusterIP: None
  selector:
    app: eureka
---
apiVersion: apps/v1beta1
kind: StatefulSet
metadata:
  name: eureka
  namespace: kube-dev
spec:
  serviceName: "eureka"
  replicas: 3
  selector:
    matchLabels:
      app: eureka
  template:
    metadata:
      labels:
        app: eureka
    spec:
      containers:
      - name: eureka
        image:  docker.io/xxlaila/kxl-eureka:v2
        ports:
        - containerPort: 8761
        env:
        - name: MY_POD_NAME
          valueFrom:
            fieldRef:
              fieldPath: metadata.name
          # Due to camelcase issues with "defaultZone" and "preferIpAddress", _JAVA_OPTIONS is used here
        - name: eureka_client_serviceUrl_defaultZone
          #value: http://eureka-0.eureka.<namespace>.svc.cluster.local:8761/eureka/,http://eureka-1.eureka.<namespace>.svc.cluster.local:8761/eureka/
          value: http://eureka-0.eureka:8761/eureka/,http://eureka-1.eureka:8761/eureka/,http://eureka-2.eureka:8761/eureka/
        - name: EUREKA_CLIENT_REGISTERWITHEUREKA
          value: "true"
        - name: EUREKA_CLIENT_FETCHREGISTRY
          value: "true"
          # In the docker image, this is set to localhost. Otherwise, we could leave this empty.
          # The hostnames must match with the the eureka serviceUrls, otherwise the replicas are reported as unavailable in the eureka dashboard
        - name: EUREKA_INSTANCE_HOSTNAME
          #value: "$(MY_POD_NAME).eureka.<namespace.svc.cluster.local"
          value: "$(MY_POD_NAME).eureka"
        - name: RUN_ENV
          value: test
        - name: CONFIG_API_SERVER
          value: http://api.conf.xxlaila.io
        - name: RUN_CLUSTER
          value: default
        - name: RUN_MODE
          value: AUTO
          #value: eureka
          # For the other (stateless) services, this should probably be set to true, since their pods have no DNS-resolvable  hostnames
       #- name: EUREKA_INSTANCE_PREFERIPADDRESS
       #  value: "false"
  # No need to start the pods in order. We just need the stable network identity
  podManagementPolicy: "Parallel"

3.4、重建pod

pod重建以后我们经过访问可以看到

img
img

后续持续优化

坚持原创技术分享，您的支持将鼓励我继续创作！
0%