Appearance
SpringCloud微服务电商系统在Kubernetes集群中上线详细教程
该文章来自:https://jiangxl.blog.csdn.net/article/details/121740024
仅做个人备份学习使用。
1.微服务架构及理论概述
程序架构一共经历了3个阶段:单体架构、SOA架构、微服务架构。
微服务系统应用由原来的单体变成几十到几百个不同的工程,会产生例如包括服务间的依赖,服务如何拆封,内部接口规范,数据传递等等问题,尤其是服务拆分,需要团队熟悉业务流程,懂得取舍,要保证拆分的粒度服务既符合“高内聚,低耦合”的基本原则,还要兼顾业务的发展以及公司的愿景,要还要说服团队成员为之努力,并且积极投入,在多方中间取得平衡。
微服务从概念上讲就是将一个大的平台拆分成几个独立的小模块,这些模块有自己专门的数据库服务,即使这个单独模块出现了故障,也不会影响整个平台的使用,例如一个平台中有订单模块、新闻模块、咨询模块,即使新闻模块宕机了,订单模块、新闻模块依然工作正常,也不会造成整个网站的瘫痪。
对于微服务版本迭代而言,只需要针对特定功能所在的子项目上线即可,版本迭代更加灵活,而单体程序版本迭代,哪怕只是一个小小的功能也需要对整个系统进行更新。
目前主流的程序架构就是微服务系统,运用的技术栈更加丰富,例如kubernetes、CI/CD、Devops等等。
微服务项目在kubernetes环境部署主要有两种实现方式:无状态的部署、istio微服务网格部署。
可以把微服务理解为是一个大的框架,在这个框架里面内嵌各个子系统。
1.1.单体架构与微服务架构的区别
微服务的特点:
- 系统服务独立化
- 每个子项目系统都会独立开发、部署,有效的避免一个服务的bug影响整个系统的使用
- 技术栈灵活
- 每个子系统之间约定一个通信方式,例如MQ、API等,不再像单体服务那么的局限性
- 独立部署
- 每个微服务系统都独立部署,都有单独的war包,即使其中一个服务异常宕机,也不会对整个平台产生影响
- 扩展性强
- 每个微服务都可以部署多套,配合注册中心实现负载均衡能力
- 独立数据
- 每个微服务都有独立的存储组件,比如数据库、缓存等,互不影响
单体服务的特点
- 易于部署
- 单体服务只有一个程序包,只需要部署一个服务即可完成整个系统的建设
- 易于测试
- 测试用例也只需要设计一个即可
单体服务的不足
- 代码量级巨大,难以维护,如果系统中存在bug,修复起来牵扯功能太多,故障百出
- 构建、部署的成本巨大,服务都集成在一起,启动的时间也会非常缓慢
- 新人上手较难,如果整个服务也没有注释的情况下,新的开发人员上手会非常难
微服务与单体服务架构上的不同
单体服务:程序包只有一个,开发团队可能有一群人,用户之间访问程序包即可获取系统信息,后端对应一个数据库
微服务:程序包有多个,每个程序包都是不同的功能结合,且每个微服务程序都有自己的单独数据库存储数据,所有的微服务信息都会存储到注册中心,网关程序在最前面,当用户有请求进来后,首先发送给网关服务,然后网关根据注册中心的数据,将不同功能的请求转发给不同的微服务程序
1.2.微服务组件架构图
以电商平台为例,描述一个购买商品的微服务之间调用流程
用户首先访问到平台的静态首页,发送一个搜索请求从前端页面提交给微服务的网关服务,网关服务是对外的,也就是不管用户要操作什么请求,都需要由网关服务转发给具体的微服务,因为网关服务是入口,因此网关服务一般都是多台做成的负载均衡,搜索请求到达网关服务后,网关服务会根据请求的功能转发到搜索服务的程序上,搜索出来商品后,会调用商品服务程序,查看商品的详细情况,用户选择完商品后,商品服务再调用订单服务,完成订单的创建,再由订单服务调用支付服务,完成商品的购买。
所有的微服务接口信息都会注册到注册中心,当程序启动成功就会自动注册到注册中心,注册中心也会起到一个负载均衡的作用,根据网关的API请求,在注册中心上找到特定的微服务程序进行转发。
微服务程序还有自己的一套配置中心,保存着各个微服务的配置文件,所有的微服务程序可能都有自己独立的后端组件,比如MySQL、Redis、MQ、分布式存储。
微服务之间的通信可以采用接口API、MQ、RPAC等机制
微服务在kubernetes集群中部署时可以采用无状态服务的部署,对于数据库、Redis、MQ、存储还是建议部署在k8s集群之外。
注册中心与微服务之间的联系:发现微服务—>微服务注册到注册中心—>进行心跳监测,异常的节点会踢出
微服务程序的门户网站会配置Gateway网关服务的地址,由门户网站用户请求的微服务信息,转发给Gateway网关程序,所有的微服务都会注册在Eureka注册中心中,Gateway网关会将请求通过注册中心转发给对应的微服务程序。
1.3.微服务注册中心
微服务面对着很多问题:
- 如何记录一个微服务副本的接口地址,如何对多个微服务节点组成的集群做负载均衡
- 符合判断集群中某一个微服务是否可用
微服务面临的这些问题都可以通过注册中心去解决,在注册中心中可用记录某一个微服务的所有副本节点,当一个副本节点挂掉了,注册中心也会将其踢出集群,注册中心会将一个微服务的所有副本节点自动形成一个负载均衡
在微服务中配合注册中心地址之后,程序一旦启动,无需配置注册中心,且能够自动注册到注册中心
微服务接入注册中心流程:微服务程序---->注册---->注册中心服务器
注册中心调用微服务流程:微服务网关程序---->注册/查询微服务(调用微服务)---->注册中心---->微服务程序
当然也可以不适用注册中心,k8s自身的功能就有服务发现自动形成负载均衡,也可以用健康检查监控服务的运行状况
目前主流的注册中心:Eureka、Nacos
1.4.不同的部署环境对于程序配置文件如何管理
可以通过以下四种方式来管理不同环境使用不同的配置文件
- kubernetes configmap资源
- 可以根据线上、预发布、测试环境编写不同的configmap资源,然后将其挂载到对应的deployment资源上
- docker 容器启动脚本entrypoint.sh
- 根据环境的不同在entrypoint.sh声明环境变量,运行特定的配置文件
- java启动命令控制
- java程序可以同时存在多个程序配置文件,可以在配置文件中声明各自的变量,然后使用
java --spring.profiles.active=xxx xxx.jar
的方式来指定不同环境的配置文件
- java程序可以同时存在多个程序配置文件,可以在配置文件中声明各自的变量,然后使用
- 使用开源的统一配置中心程序,比如市面上主流的Apollo以及Disconf,这两款程序都有自己的图形化管理系统,配置可视化
2.微服务迁移至kubernetes平台流程
2.1.项目迁移到K8S平台流程
项目迁移到K8S平台大致流程分为以下几个阶段:
- 制作程序镜像
- 在K8s环境,所有的程序都是以Docker镜像来运行的,一个镜像包含了文件系统、程序的运行环境、程序本身,可以通过Dockerfile的形式制作镜像
- 使用资源控制器管理Pod资源
- 程序镜像运行需要依靠Pod资源,需要有Pod控制器去管理Pod资源
- 常见的Pod控制器:Deployment(无状态程序部署)、Statefulset(有状态程序部署)、Job/CronJob(批量处理)
- 将Pod资源使用服务发现对外进行暴露
- Services资源注意通过label标签与Pod进行关联,实现对一组Pod的自动负载均衡以及服务发现
- Services资源支持Cluster、Nodeport、LoadBalancer三种类型
- 对外发布应用程序
- 可以使用单独的nginx容器去反向代理Cordns解析的services名称将Pod应用对外发布
- 也可以使用ingress+services的方式将应用程序对外发布,使用ingress的前提是必须可以通过域名访问应用
- 架构图如下
- 后期的日志收集与监控
- 当应用在K8s程序运行后,对于程序的日志比如nginx、tomcat这些日志文件都不会做持久化,因此就需要考虑对日志进行收集,可以采用Filebeat+ELK方式对程序日志进行收集
- 使用Prometheus+Grafana对整个K8s平台进行全方面监控
2.2.传统部署与K8S部署的区别
传统方式部署项目方式:
首先由开发人员提交代码到Gitlab代码仓库,Jenkins触发更新从Gitlab上获取最新代码,通过maven将代码编译成war包或者jar包,再由Jenkins通过写好的自动化部署脚本或者Ansible自动化程序将war包部署到程序所在的云主机,云主机一定会是多台组成的负载均衡集群,将负载均衡VIP与域名进行绑定,用户通过访问域名由负载均衡转发至对应的后台主机。
传统架构运维环境一般由zabbix监控系统进行监控,日志采集会使用ELK平台。
Kubernetes平台部署项目方式:
首先由开发人员提交代码到Gitlab代码仓库,Jenkins触发更新从Gitlab上获取最新代码,通过写好流水线将程序打包成Docker镜像并推送至Harbor仓库,再由Jenkins调用K8s的Master Api将程序的镜像采用Deployment控制器部署到K8s集群,再由service资源暴露pod资源,最后由ingress或者nginx资源将程序发布到互联网。
3.微服务程序前期环境准备(配置、编译、制作镜像)
3.1.simple-microservice微服务项目介绍
本次使用的微服务为spring cloud的电商平台项目
项目名称:simple-microservice
simple-microservice微服务的各个程序:
- eureka-service(注册中心)
- gateway-service(网管服务)
- order-service(订单服务)
- product-service(商品服务)
- portal-service(门户网站)
- stock-service(库存服务)
用户请求首先到Protal前端页面,也就是程序的首页,在首页的各项功能操作都会由Gateway网关服务转发到各自的微服务程序上,比如请求一个订单服务,订单服务会事先注册到Eureka中,由Gateway将请求发送给Eureka,再由Eureka转发给具体的微服务程序,这些微服务程序都有自己单独的数据库服务。
微服务程序的门户网站会配置Gateway网关服务的地址,由门户网站用户请求的微服务信息,转发给Gateway网关程序,所有的微服务都会注册在Eureka注册中心中,Gateway网关会将请求通过注册中心转发给对应的微服务程序。
该微服务程序代码可以托管在gitlab代码仓库中,我们可以建立多个分支模拟从开发到上线的流程,理念:分别创建5个分支,在每个分支中增加对应的代码,最后合并到master分支,也可以直接就在master分支上新增配置文件,我是直接拉取dev1分支,只有程序代码的分支,然后新增一系列配置,最后上线。
拉取dev1分支,进行一系列配置的话,首先要把master分支里的protal程序代码同步到dev1分支里,因为dev1的protal程序代码有点bug,整个微服务运行完毕后,order订单服务会报错。
分别有五个分支:
- dev1:交付代码
- dev2:新增程序打包的Dockerfile文件
- dev3:新增K8s资源编排Yaml文件
- dev4:微服务链路监控
- dev5:新功能代码更新上线
- master:上线版本
3.2.simple微服务部署到K8S逻辑架构
simple-microservice微服务部署到k8s平台主要分为如下几个步骤:
- 修改程序配置文件,将数据库等各组件地址修改为当前环境
- 将各个微服务程序使用Maven编译成可部署的jar包
- 编写Dockerfile将程序做成docker镜像
对K8S环境的要求:
一套完整的K8S集群、有cordns可以解析Service地址、Harbor镜像仓库、Nginx容器将项目发布到互联网或者INgress资源将项目发布到互联网。
微服务部署到K8S环境之后架构是怎样的
首先用户访问网站的域名,请求达到前端页面,点击前端页面中的功能跳转到不同的微服务,比如商品服务,将商品服务的请求交个Gateway网关服务,由网关服务转发给商品的微服务程序Product,各个微服务之间的调用,都从Eureka注册中心读取各个微服务的信息完成调度,Product(商品)、Order(订单)、Stock(库存)都有单独的数据库,MySQL数据部署在K8S集群之外。
服务发布到互联网采用Nginx而不采用Ingress,Nginx可以配置更多的参数,将Nginx容器运行在每一个Node主机上,映射Node主机的80/443端口,通过在hosts文件里绑定node主机ip+程序域名的方式访问微服务。
3.3.准备微服务各程序的配置文件
在使用Maven之前,首先将程序的配置文件进行修改,编译成功之后再想修改配置文件内容,就需要重新编译了,将每个微服务的环境地址都修改成自己环境的地址。
拉取dev1分支展开全新配置
3.3.1.创建微服务各个程序的prod配置文件
eureka(注册中心)、gateway(网关)、order(订单)、portal(首页)、product(商品)、stock(库存)这几个微服务程序都需要配置,新建一个prod的配置文件,用于线上生产环境
order(订单)、product(商品)、stock(库存)这些微服务有两个包,分别是xxx-biz和xxx-api,只需要部署biz目录下的jar包即可,api的目录只是存在一些依赖
order(订单)、product(商品)、stock(库存)这些微服务都有自己单独的数据库
先将prod的配置文件创建出来,待eureka项目部署完成后,再进行进一步的修改,最后将程序编译
1.eureka-service配置文件
1.创建一个prod的配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1]\# cd eureka-service/src/main/resources/
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/eureka-service/src/main/resources]\# cp application-fat.yml application-prod.yml
2.程序使用prod的配置文件运行
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/eureka-service/src/main/resources]\# vim application.yml
server:
port: 8080 #eureka的端口号
spring:
application:
name: legendshop-basic-eureka
profiles:
active: prod #将这里修改为prod就会调用prod配置文件
3.配置eureka服务的配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/eureka-service/src/main/resources]\# vim application-prod.yml
eureka:
server:
renewal-percent-threshold: 0.9
enable-self-preservation: false
eviction-interval-timer-in-ms: 40000
instance:
hostname: 127.0.0.1
prefer-ip-address: true
client:
register-with-eureka: false
serviceUrl:
defaultZone: http://127.0.0.1:${server.port}/eureka/
fetch-registry: false
1.创建一个prod的配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1]\# cd eureka-service/src/main/resources/
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/eureka-service/src/main/resources]\# cp application-fat.yml application-prod.yml
2.程序使用prod的配置文件运行
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/eureka-service/src/main/resources]\# vim application.yml
server:
port: 8080 #eureka的端口号
spring:
application:
name: legendshop-basic-eureka
profiles:
active: prod #将这里修改为prod就会调用prod配置文件
3.配置eureka服务的配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/eureka-service/src/main/resources]\# vim application-prod.yml
eureka:
server:
renewal-percent-threshold: 0.9
enable-self-preservation: false
eviction-interval-timer-in-ms: 40000
instance:
hostname: 127.0.0.1
prefer-ip-address: true
client:
register-with-eureka: false
serviceUrl:
defaultZone: http://127.0.0.1:${server.port}/eureka/
fetch-registry: false
2.gateway-service配置文件
1.创建一个prod的配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1]\# cd gateway-service/src/main/resources/
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/gateway-service/src/main/resources]\# cp application-fat.yml application-prod.yml
2.程序使用prod的配置文件运行
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/gateway-service/src/main/resources]\# cat application.yml
server:
port: 8080
spring:
profiles:
active: prod
application:
name: @artifactId@
3.配置程序配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/gateway-service/src/main/resources]\# vim application-prod.yml
······
eureka:
instance:
prefer-ip-address: true
client:
register-with-eureka: true
fetch-registry: true
service-url:
defaultZone: http://eureka-0.eureka-service:8080/eureka/ #修改eureka地址
1.创建一个prod的配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1]\# cd gateway-service/src/main/resources/
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/gateway-service/src/main/resources]\# cp application-fat.yml application-prod.yml
2.程序使用prod的配置文件运行
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/gateway-service/src/main/resources]\# cat application.yml
server:
port: 8080
spring:
profiles:
active: prod
application:
name: @artifactId@
3.配置程序配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/gateway-service/src/main/resources]\# vim application-prod.yml
······
eureka:
instance:
prefer-ip-address: true
client:
register-with-eureka: true
fetch-registry: true
service-url:
defaultZone: http://eureka-0.eureka-service:8080/eureka/ #修改eureka地址
3.order-service配置文件
1.创建一个prod的配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1]\# cd order-service/order-service-biz/src/main/resources/
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/order-service/order-service-biz/src/main/resources]\# cp application-fat.yml application-prod.yml
2.程序使用prod的配置文件运行
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/order-service/order-service-biz/src/main/resources]\# vim application.yml
server:
port: 8080
spring:
profiles:
active: prod
application:
name: order-service
3.配置程序配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/order-service/order-service-biz/src/main/resources]\# vim application-prod.yml
spring:
datasource:
url: jdbc:mysql://192.168.20.11:3306/tb_order?characterEncoding=utf-8 #mysql数据库地址
username: root
password: 123456
driver-class-name: com.mysql.jdbc.Driver
eureka:
instance:
prefer-ip-address: true
client:
register-with-eureka: true
fetch-registry: true
service-url:
defaultZone: http://eureka-0.eureka-service:8080/eureka/ #eureka地址
1.创建一个prod的配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1]\# cd order-service/order-service-biz/src/main/resources/
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/order-service/order-service-biz/src/main/resources]\# cp application-fat.yml application-prod.yml
2.程序使用prod的配置文件运行
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/order-service/order-service-biz/src/main/resources]\# vim application.yml
server:
port: 8080
spring:
profiles:
active: prod
application:
name: order-service
3.配置程序配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/order-service/order-service-biz/src/main/resources]\# vim application-prod.yml
spring:
datasource:
url: jdbc:mysql://192.168.20.11:3306/tb_order?characterEncoding=utf-8 #mysql数据库地址
username: root
password: 123456
driver-class-name: com.mysql.jdbc.Driver
eureka:
instance:
prefer-ip-address: true
client:
register-with-eureka: true
fetch-registry: true
service-url:
defaultZone: http://eureka-0.eureka-service:8080/eureka/ #eureka地址
5.product-service配置文件
1.创建一个prod的配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1]\# cd product-service/product-service-biz/src/main/resources/
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/product-service/product-service-biz/src/main/resources]\# cp application-fat.yml application-prod.yml
2.程序使用prod的配置文件运行
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/product-service/product-service-biz/src/main/resources]\# vim application.yml
server:
port: 8080
spring:
profiles:
active: prod
application:
name: product-service
3.配置程序配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/product-service/product-service-biz/src/main/resources]\# vim application-prod.yml
spring:
datasource:
url: jdbc:mysql://192.168.20.11:3306/tb_product?characterEncoding=utf-8 #数据库地址
username: root
password: 123456
driver-class-name: com.mysql.jdbc.Driver
eureka:
instance:
prefer-ip-address: true
client:
register-with-eureka: true
fetch-registry: true
service-url:
defaultZone: http://eureka-0.eureka-service:8080/eureka/ #eureka地址
1.创建一个prod的配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1]\# cd product-service/product-service-biz/src/main/resources/
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/product-service/product-service-biz/src/main/resources]\# cp application-fat.yml application-prod.yml
2.程序使用prod的配置文件运行
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/product-service/product-service-biz/src/main/resources]\# vim application.yml
server:
port: 8080
spring:
profiles:
active: prod
application:
name: product-service
3.配置程序配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/product-service/product-service-biz/src/main/resources]\# vim application-prod.yml
spring:
datasource:
url: jdbc:mysql://192.168.20.11:3306/tb_product?characterEncoding=utf-8 #数据库地址
username: root
password: 123456
driver-class-name: com.mysql.jdbc.Driver
eureka:
instance:
prefer-ip-address: true
client:
register-with-eureka: true
fetch-registry: true
service-url:
defaultZone: http://eureka-0.eureka-service:8080/eureka/ #eureka地址
6.stock-service配置文件
1.创建一个prod的配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1]\# cd stock-service/stock-service-biz/src/main/resources/
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/stock-service/stock-service-biz/src/main/resources]\# cp application-fat.yml application-prod.yml
2.程序使用prod的配置文件运行
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/stock-service/stock-service-biz/src/main/resources]\# vim application.yml
server:
port: 8080
spring:
profiles:
active: prod
application:
name: stock-service
3.配置程序配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/stock-service/stock-service-biz/src/main/resources]\# vim application-prod.yml
spring:
datasource:
url: jdbc:mysql://192.168.20.11:3306/tb_stock?characterEncoding=utf-8 #数据库地址
username: root
password: 123456
driver-class-name: com.mysql.jdbc.Driver
eureka:
instance:
prefer-ip-address: true
client:
register-with-eureka: true
fetch-registry: true
service-url:
defaultZone: http://eureka-service.eureka:8080/eureka/ #eureka地址
1.创建一个prod的配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1]\# cd stock-service/stock-service-biz/src/main/resources/
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/stock-service/stock-service-biz/src/main/resources]\# cp application-fat.yml application-prod.yml
2.程序使用prod的配置文件运行
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/stock-service/stock-service-biz/src/main/resources]\# vim application.yml
server:
port: 8080
spring:
profiles:
active: prod
application:
name: stock-service
3.配置程序配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/stock-service/stock-service-biz/src/main/resources]\# vim application-prod.yml
spring:
datasource:
url: jdbc:mysql://192.168.20.11:3306/tb_stock?characterEncoding=utf-8 #数据库地址
username: root
password: 123456
driver-class-name: com.mysql.jdbc.Driver
eureka:
instance:
prefer-ip-address: true
client:
register-with-eureka: true
fetch-registry: true
service-url:
defaultZone: http://eureka-service.eureka:8080/eureka/ #eureka地址
4.portal-service配置文件
1.创建一个prod的配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1]\# cd portal-service/src/main/resources/
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/portal-service/src/main/resources]\# cp application-fat.yml application-prod.yml
2.程序使用prod的配置文件运行
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/portal-service/src/main/resources]\# vim application.yml
server:
port: 8080
undertow:
io-threads: 16
worker-threads: 256
buffer-size: 1024
direct-buffers: true
spring:
application:
name: portal-service
profiles:
active: prod
3.配置程序配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/portal-service/src/main/resources]\# vim application-prod.yml
eureka:
instance:
prefer-ip-address: true
client:
service-url:
defaultZone: http://eureka-0.eureka-service:8080/eureka/ #指定eureka地址
register-with-eureka: true
fetch-registry: true
······
1.创建一个prod的配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1]\# cd portal-service/src/main/resources/
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/portal-service/src/main/resources]\# cp application-fat.yml application-prod.yml
2.程序使用prod的配置文件运行
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/portal-service/src/main/resources]\# vim application.yml
server:
port: 8080
undertow:
io-threads: 16
worker-threads: 256
buffer-size: 1024
direct-buffers: true
spring:
application:
name: portal-service
profiles:
active: prod
3.配置程序配置文件
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/portal-service/src/main/resources]\# vim application-prod.yml
eureka:
instance:
prefer-ip-address: true
client:
service-url:
defaultZone: http://eureka-0.eureka-service:8080/eureka/ #指定eureka地址
register-with-eureka: true
fetch-registry: true
······
3.3.2.调整每个微服务的pom文件增加prod环境
由于我们新加了一个prod线上环境的配置文件,因此也需要在pom.xml中配置,配置完成后可以编译一下,看看是否可以做成可部署的程序包
1.安装maven
[root@binary-k8s-master1 ~]\# yum -y install java-1.8.0-openjdk maven
2.调整pom.xml增加prod环境配置
[root@binary-k8s-master1 simple-microservice-dev1]\# vim pom.xml
······ #65行左右
<profile>
<id>prod</id>
<properties>
<profileActive>prod</profileActive>
</properties>
</profile>
</profiles>
······
2.编译程序
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1]\# mvn clean package -D maven.test.skip=true -P prod
3.查看编译成功的war包
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1]\# find . -name "*.jar"
./basic-common/basic-common-core/target/basic-common-core.jar
./eureka-service/target/eureka-service.jar
./gateway-service/target/gateway-service.jar
./order-service/order-service-api/target/order-service-api.jar
./order-service/order-service-biz/target/order-service-biz.jar
./portal-service/target/portal-service.jar
./product-service/product-service-api/target/product-service-api.jar
./product-service/product-service-biz/target/product-service-biz.jar
./stock-service/stock-service-api/target/stock-service-api.jar
./stock-service/stock-service-biz/target/stock-service-biz.jar
1.安装maven
[root@binary-k8s-master1 ~]\# yum -y install java-1.8.0-openjdk maven
2.调整pom.xml增加prod环境配置
[root@binary-k8s-master1 simple-microservice-dev1]\# vim pom.xml
······ #65行左右
<profile>
<id>prod</id>
<properties>
<profileActive>prod</profileActive>
</properties>
</profile>
</profiles>
······
2.编译程序
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1]\# mvn clean package -D maven.test.skip=true -P prod
3.查看编译成功的war包
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1]\# find . -name "*.jar"
./basic-common/basic-common-core/target/basic-common-core.jar
./eureka-service/target/eureka-service.jar
./gateway-service/target/gateway-service.jar
./order-service/order-service-api/target/order-service-api.jar
./order-service/order-service-biz/target/order-service-biz.jar
./portal-service/target/portal-service.jar
./product-service/product-service-api/target/product-service-api.jar
./product-service/product-service-biz/target/product-service-biz.jar
./stock-service/stock-service-api/target/stock-service-api.jar
./stock-service/stock-service-biz/target/stock-service-biz.jar
3.4.制作微服务程序所使用的的底层系统镜像
3.4.1.为程序环境构建底层系统镜像
由于是SpringCloud微服务系统,SpringCloud属于JAVA语言开发的项目,因此要有JAVA启动环境。
1.拉取centos7.5作为底层容器
[root@binary-k8s-master1 ~]\# docker pull centos:7.5.1804
2.启动容器
[root@binary-k8s-master1 ~]\# docker run -itd centos:7.5.1804
3.进入系统容器进行环境配置
[root@binary-k8s-master1 ~]\# docker exec -it 462237f9a0a bash
#安装java环境
[root@2fcfb51f04e0 /]\# yum -y install java-1.8.0-openjdk net-tools
[root@2fcfb51f04e0 /]\# java -version
openjdk version "1.8.0_302"
OpenJDK Runtime Environment (build 1.8.0_302-b08)
OpenJDK 64-Bit Server VM (build 25.302-b08, mixed mode)
[root@2fcfb51f04e0 /]\# mkdir /data/simple-microservice -p
#调整系统时区
[root@2fcfb51f04e0 /]\# yum -y intall tzdata
[root@2fcfb51f04e0 /]\# ln -sf /usr/share/zoneinfo/Asia/Shanghai /etc/localtime
4.将容器提交为镜像
[root@binary-k8s-master1 ~]\# docker commit 2fcfb51f04e0 centos-java:v1
5.配置Harbor镜像仓库
[root@binary-k8s-master1 ~]\# vim /etc/docker/daemon.json
{
"registry-mirrors": ["https://9wn5tbfh.mirror.aliyuncs.com"],
"insecure-registries": ["192.168.20.10"]
}
[root@binary-k8s-master1 ~]\# systemctl restart docker
6.提交镜像到Harbor仓库
[root@binary-k8s-master1 ~]\# docker login -u admin -p Harbor12345 192.168.20.10
[root@binary-k8s-master1 ~]\# docker tag centos-java:v1 192.168.20.11/base/centos-java:v1
[root@binary-k8s-master1 ~]\# docker push 192.168.20.11/base/centos-java:v1
1.拉取centos7.5作为底层容器
[root@binary-k8s-master1 ~]\# docker pull centos:7.5.1804
2.启动容器
[root@binary-k8s-master1 ~]\# docker run -itd centos:7.5.1804
3.进入系统容器进行环境配置
[root@binary-k8s-master1 ~]\# docker exec -it 462237f9a0a bash
#安装java环境
[root@2fcfb51f04e0 /]\# yum -y install java-1.8.0-openjdk net-tools
[root@2fcfb51f04e0 /]\# java -version
openjdk version "1.8.0_302"
OpenJDK Runtime Environment (build 1.8.0_302-b08)
OpenJDK 64-Bit Server VM (build 25.302-b08, mixed mode)
[root@2fcfb51f04e0 /]\# mkdir /data/simple-microservice -p
#调整系统时区
[root@2fcfb51f04e0 /]\# yum -y intall tzdata
[root@2fcfb51f04e0 /]\# ln -sf /usr/share/zoneinfo/Asia/Shanghai /etc/localtime
4.将容器提交为镜像
[root@binary-k8s-master1 ~]\# docker commit 2fcfb51f04e0 centos-java:v1
5.配置Harbor镜像仓库
[root@binary-k8s-master1 ~]\# vim /etc/docker/daemon.json
{
"registry-mirrors": ["https://9wn5tbfh.mirror.aliyuncs.com"],
"insecure-registries": ["192.168.20.10"]
}
[root@binary-k8s-master1 ~]\# systemctl restart docker
6.提交镜像到Harbor仓库
[root@binary-k8s-master1 ~]\# docker login -u admin -p Harbor12345 192.168.20.10
[root@binary-k8s-master1 ~]\# docker tag centos-java:v1 192.168.20.11/base/centos-java:v1
[root@binary-k8s-master1 ~]\# docker push 192.168.20.11/base/centos-java:v1
3.4.2.为每个微服务程序编写Dockerfile
设计概念:
采用3.3.1中制作的底层镜像
第一步:在镜像里面创建/data/simple-microservice/微服务名称
这样的目录用于存放程序的jar包
第二步:将程序的jar包拷贝至镜像中
第三步:生成一个启动脚本,由于服务名称都不一样也就是每个程序的路径都不同,想要做到一个底层镜像多个程序复用,那么就需要在启动脚本里手动指定程序的路径,在脚本中指定程序的部署路径以及日志路径
第四步:暴露8080端口
将Dockerfile放到程序编译的目录也就是与src在同一目录
1.eureka-service服务
在4.2.2.会写
在4.2.2.会写
2.gateway-service服务
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/gateway-service]\# vim Dockerfile
FROM 192.168.20.11/base/centos-java:v1
RUN mkdir /data/simple-microservice/gateway-service -p
COPY ./target/gateway-service.jar /data/simple-microservice/gateway-service/
RUN echo -e "#!/bin/bash \njava -jar /data/simple-microservice/gateway-service/gateway-service.jar > /data/simple-microservice/gateway-service/gateway-service.log & \ntail -f /data/simple-microservice/gateway-service/gateway-service.log" > /data/entrypoint.sh && chmod a+x /data/entrypoint.sh
EXPOSE 8080
ENTRYPOINT /data/entrypoint.sh
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/gateway-service]\# vim Dockerfile
FROM 192.168.20.11/base/centos-java:v1
RUN mkdir /data/simple-microservice/gateway-service -p
COPY ./target/gateway-service.jar /data/simple-microservice/gateway-service/
RUN echo -e "#!/bin/bash \njava -jar /data/simple-microservice/gateway-service/gateway-service.jar > /data/simple-microservice/gateway-service/gateway-service.log & \ntail -f /data/simple-microservice/gateway-service/gateway-service.log" > /data/entrypoint.sh && chmod a+x /data/entrypoint.sh
EXPOSE 8080
ENTRYPOINT /data/entrypoint.sh
3.order-service服务
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/order-service/order-service-biz]\# vim Dockerfile
FROM 192.168.20.11/base/centos-java:v1
RUN mkdir /data/simple-microservice/order-service -p
COPY ./target/order-service-biz.jar /data/simple-microservice/order-service/
RUN echo -e "#!/bin/bash \njava -jar /data/simple-microservice/order-service/order-service-biz.jar > /data/simple-microservice/order-service/order-service-biz.log & \ntail -f /data/simple-microservice/order-service/order-service-biz.log" > /data/entrypoint.sh && chmod a+x /data/entrypoint.sh
EXPOSE 8080
ENTRYPOINT /data/entrypoint.sh
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/order-service/order-service-biz]\# vim Dockerfile
FROM 192.168.20.11/base/centos-java:v1
RUN mkdir /data/simple-microservice/order-service -p
COPY ./target/order-service-biz.jar /data/simple-microservice/order-service/
RUN echo -e "#!/bin/bash \njava -jar /data/simple-microservice/order-service/order-service-biz.jar > /data/simple-microservice/order-service/order-service-biz.log & \ntail -f /data/simple-microservice/order-service/order-service-biz.log" > /data/entrypoint.sh && chmod a+x /data/entrypoint.sh
EXPOSE 8080
ENTRYPOINT /data/entrypoint.sh
4.portal-service服务
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/portal-service]\# vim Dockerfile
FROM 192.168.20.11/base/centos-java:v1
RUN mkdir /data/simple-microservice/portal-service -p
COPY ./target/portal-service.jar /data/simple-microservice/portal-service/
RUN echo -e "#!/bin/bash \njava -jar /data/simple-microservice/portal-service/portal-service.jar > /data/simple-microservice/portal-service/portal-service.log & \ntail -f /data/simple-microservice/portal-service/portal-service.log" > /data/entrypoint.sh && chmod a+x /data/entrypoint.sh
EXPOSE 8080
ENTRYPOINT /data/entrypoint.sh
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/portal-service]\# vim Dockerfile
FROM 192.168.20.11/base/centos-java:v1
RUN mkdir /data/simple-microservice/portal-service -p
COPY ./target/portal-service.jar /data/simple-microservice/portal-service/
RUN echo -e "#!/bin/bash \njava -jar /data/simple-microservice/portal-service/portal-service.jar > /data/simple-microservice/portal-service/portal-service.log & \ntail -f /data/simple-microservice/portal-service/portal-service.log" > /data/entrypoint.sh && chmod a+x /data/entrypoint.sh
EXPOSE 8080
ENTRYPOINT /data/entrypoint.sh
5.product-service服务
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/product-service/product-service-biz]\# vim Dockerfile
FROM 192.168.20.11/base/centos-java:v1
RUN mkdir /data/simple-microservice/product-service -p
COPY ./target/product-service-biz.jar /data/simple-microservice/product-service/
RUN echo -e "#!/bin/bash \njava -jar /data/simple-microservice/product-service/product-service-biz.jar > /data/simple-microservice/product-service/product-service-biz.log & \ntail -f /data/simple-microservice/product-service/product-service-biz.log" > /data/entrypoint.sh && chmod a+x /data/entrypoint.sh
EXPOSE 8080
ENTRYPOINT /data/entrypoint.sh
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/product-service/product-service-biz]\# vim Dockerfile
FROM 192.168.20.11/base/centos-java:v1
RUN mkdir /data/simple-microservice/product-service -p
COPY ./target/product-service-biz.jar /data/simple-microservice/product-service/
RUN echo -e "#!/bin/bash \njava -jar /data/simple-microservice/product-service/product-service-biz.jar > /data/simple-microservice/product-service/product-service-biz.log & \ntail -f /data/simple-microservice/product-service/product-service-biz.log" > /data/entrypoint.sh && chmod a+x /data/entrypoint.sh
EXPOSE 8080
ENTRYPOINT /data/entrypoint.sh
6.stock-service服务
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/stock-service/stock-service-biz]\# cat Dockerfile
FROM 192.168.20.11/base/centos-java:v1
RUN mkdir /data/simple-microservice/stock-service -p
COPY ./target/stock-service-biz.jar /data/simple-microservice/stock-service/
RUN echo -e "#!/bin/bash \njava -jar /data/simple-microservice/stock-service/stock-service-biz.jar > /data/simple-microservice/stock-service/stock-service-biz.log & \ntail -f /data/simple-microservice/stock-service/stock-service-biz.log" > /data/entrypoint.sh && chmod a+x /data/entrypoint.sh
EXPOSE 8080
ENTRYPOINT /data/entrypoint.sh
[root@binary-k8s-master1 ~/springcloud/simple-microservice-dev1/stock-service/stock-service-biz]\# cat Dockerfile
FROM 192.168.20.11/base/centos-java:v1
RUN mkdir /data/simple-microservice/stock-service -p
COPY ./target/stock-service-biz.jar /data/simple-microservice/stock-service/
RUN echo -e "#!/bin/bash \njava -jar /data/simple-microservice/stock-service/stock-service-biz.jar > /data/simple-microservice/stock-service/stock-service-biz.log & \ntail -f /data/simple-microservice/stock-service/stock-service-biz.log" > /data/entrypoint.sh && chmod a+x /data/entrypoint.sh
EXPOSE 8080
ENTRYPOINT /data/entrypoint.sh
3.5.调整Protal微服务中连接Gateway网关的地址
Protal是我们电商平台的门户网站,也就是前端首页,在Protal微服务中会填写好其他微服务程序的地址,Gateway网关程序的地址就是写死在Protal项目的js里,在我们定义好Gateway网关服务的域名后,就要将域名写入到对应的js文件里,否则当项目启动之后是无法调通的。
需要修改下列这四个文件
portal-service/src/main/resources/static/js/orderList.js
portal-service/src/main/resources/static/js/productList.js
打开这四个文件,将里面的gateway.ctnrs.com域名修改为gateway.jiangxl.com域名
vim修改命令:%s/gateway.ctnrs.com/gateway.jiangxl.com/g
打开这四个文件,将里面的gateway.ctnrs.com域名修改为gateway.jiangxl.com域名
vim修改命令:%s/gateway.ctnrs.com/gateway.jiangxl.com/g
3.6.部署Ingress用于将微服务程序发布到互联网
1.下载ingress部署的yaml文件
[root@binary-k8s-master1 ~]\# wget https://raw.githubusercontent.com/kubernetes/ingress-nginx/nginx-0.30.0/deploy/static/mandatory.yaml
[root@binary-k8s-master1 ~]\# wget https://raw.githubusercontent.com/kubernetes/ingress-nginx/nginx-0.30.0/deploy/static/provider/baremetal/service-nodeport.yaml
2.调整ingress nodeport端口
[root@binary-k8s-master1 ~/springcloud/ingress]\# vim service-nodeport.yaml
······
ports:
- name: http
port: 80
targetPort: 80
nodePort: 80
protocol: TCP
- name: https
port: 443
targetPort: 443
nodePort: 443
protocol: TCP
······
3.部署ingress
[root@binary-k8s-master1 ~/springcloud/ingress]\# kubectl apply -f mandatory.yaml
4.查看ingress部署的结果
[root@binary-k8s-master1 ~]\# kubectl get all -n ingress-nginx
NAME READY STATUS RESTARTS AGE
pod/nginx-ingress-controller-766867958b-n7dwj 1/1 Running 0 49m
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
service/ingress-nginx NodePort 10.0.0.107 <none> 80:80/TCP,443:443/TCP 49m
NAME READY UP-TO-DATE AVAILABLE AGE
deployment.apps/nginx-ingress-controller 1/1 1 1 49m
NAME DESIRED CURRENT READY AGE
replicaset.apps/nginx-ingress-controller-766867958b 1 1 1 49m
1.下载ingress部署的yaml文件
[root@binary-k8s-master1 ~]\# wget https://raw.githubusercontent.com/kubernetes/ingress-nginx/nginx-0.30.0/deploy/static/mandatory.yaml
[root@binary-k8s-master1 ~]\# wget https://raw.githubusercontent.com/kubernetes/ingress-nginx/nginx-0.30.0/deploy/static/provider/baremetal/service-nodeport.yaml
2.调整ingress nodeport端口
[root@binary-k8s-master1 ~/springcloud/ingress]\# vim service-nodeport.yaml
······
ports:
- name: http
port: 80
targetPort: 80
nodePort: 80
protocol: TCP
- name: https
port: 443
targetPort: 443
nodePort: 443
protocol: TCP
······
3.部署ingress
[root@binary-k8s-master1 ~/springcloud/ingress]\# kubectl apply -f mandatory.yaml
4.查看ingress部署的结果
[root@binary-k8s-master1 ~]\# kubectl get all -n ingress-nginx
NAME READY STATUS RESTARTS AGE
pod/nginx-ingress-controller-766867958b-n7dwj 1/1 Running 0 49m
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
service/ingress-nginx NodePort 10.0.0.107 <none> 80:80/TCP,443:443/TCP 49m
NAME READY UP-TO-DATE AVAILABLE AGE
deployment.apps/nginx-ingress-controller 1/1 1 1 49m
NAME DESIRED CURRENT READY AGE
replicaset.apps/nginx-ingress-controller-766867958b 1 1 1 49m
4.在kubernetes集群部署simple微服务项目
simple微服务项目在K8S集群中部署,只有eureka服务采用statefulset控制器部署,其余所有微服务均采用deployment控制器部署,所有的微服务通过ingress发布在互联网。
simple微服务电商项目在K8S集群部署流程
1)部署simple电商微服务项目的数据库环境
2)在K8S中部署Eureka注册中心
3)修改每个微服务的配置文件,指定eureka注册中心集群的地址
4)使用maven将每个微服务构建成可部署的jar包
5)部署Gateway关服务
6)部署protal、order、product、stock服务
4.1.部署simple微服务程序的mysql数据库环境
1.安装MySQL数据库
[root@harbor-mysql ~]\# rpm -ivh http://repo.mysql.com/yum/mysql-5.6-community/el/7/x86_64/mysql-community-release-el7-5.noarch.rpm
[root@harbor-mysql ~]\# yum install mysql-community-server
2.启动MySQL
[root@harbor-mysql ~]\# systemctl start mysqld
[root@harbor-mysql ~]\# systemctl enable mysqld
3.设置数据库密码
[root@harbor-mysql ~]\# mysqladmin -u root password '123456'
4.创建order(订单)、product(商品)、stock(库存)微服务的数据库
[root@harbor-mysql ~]\# mysql -uroot -p123456
mysql> create database tb_order;
Query OK, 1 row affected (0.00 sec)
mysql> create database tb_product;
Query OK, 1 row affected (0.01 sec)
mysql> create database tb_stock;
Query OK, 1 row affected (0.00 sec)
5.导入数据库数据
#将数据库文件传到数据库服务器
[root@binary-k8s-master1 simple-microservice-dev1]\# scp -r db root@192.168.20.11:/root
#导入order数据库的数据
mysql> use tb_order
mysql> source /root/db/order.sql
#导入product数据库的数据
mysql> use tb_product
mysql> source /root/db/product.sql
#导入stock数据库的数据
mysql> use tb_stock;
mysql> source /root/db/stock.sql
6.允许数据库远程访问
mysql> grant all privileges on *.* to 'root'@'%' identified by '123456' with grant option;
Query OK, 0 rows affected (0.00 sec)
1.安装MySQL数据库
[root@harbor-mysql ~]\# rpm -ivh http://repo.mysql.com/yum/mysql-5.6-community/el/7/x86_64/mysql-community-release-el7-5.noarch.rpm
[root@harbor-mysql ~]\# yum install mysql-community-server
2.启动MySQL
[root@harbor-mysql ~]\# systemctl start mysqld
[root@harbor-mysql ~]\# systemctl enable mysqld
3.设置数据库密码
[root@harbor-mysql ~]\# mysqladmin -u root password '123456'
4.创建order(订单)、product(商品)、stock(库存)微服务的数据库
[root@harbor-mysql ~]\# mysql -uroot -p123456
mysql> create database tb_order;
Query OK, 1 row affected (0.00 sec)
mysql> create database tb_product;
Query OK, 1 row affected (0.01 sec)
mysql> create database tb_stock;
Query OK, 1 row affected (0.00 sec)
5.导入数据库数据
#将数据库文件传到数据库服务器
[root@binary-k8s-master1 simple-microservice-dev1]\# scp -r db root@192.168.20.11:/root
#导入order数据库的数据
mysql> use tb_order
mysql> source /root/db/order.sql
#导入product数据库的数据
mysql> use tb_product
mysql> source /root/db/product.sql
#导入stock数据库的数据
mysql> use tb_stock;
mysql> source /root/db/stock.sql
6.允许数据库远程访问
mysql> grant all privileges on *.* to 'root'@'%' identified by '123456' with grant option;
Query OK, 0 rows affected (0.00 sec)
4.2.部署eureka微服务注册中心
Eureka在整个微服务架构中起到至关重要的的作用,必须保证Eureka服务高可用。
Eureka服务采用statefulset有状态应用部署方式再K8S集群中进行部署。
Eureka本身就可以利用自身的服务发现来实现集群的高可用,只需要在Eureka配置文件中指定各个Eureka节点的地址就可以实现。
将程序部署在K8S之后,如何能够保证每个Eureka Pod的地址唯一不变,就需要用到Statefulset控制器采用有状态的应用部署,使用Statefulset控制器之后,每个Pod的名称都是唯一不变的,并且每个Pod都有一个DNS服务发现地址,通过这个DNS地址就可以请求到Pod中的服务,主要是coredns组件来实现的,因此集群还需要安装有coredns服务,我们可以将每一个Eureka Pod的DNS服务发现地址写入到Eureka的配置文件中,每个Eureka节点相互注册,最终形成高可用集群。
Statefulset+Headless部署的程序DNS名格式:<statefulsetName-index>.<service-name>.<namespace-name>.svc.cluster.local
后面的.svc.cluster.local可以去掉,K8S默认就是.svc.cluster.local。
<statefulsetName-index>
就是Pod的名称,-index是每一个Pod名称后面的数字编号,Statefulset控制器部署好的Pod,会在每一个Pod名称后面加上一个数字编号,从0开始,一次累加,保证所有的Pod都有一个独立的访问入口
<service-name>
是程序的service资源的名称
<namespace-name>
是程序所在的namespace名称
我们的Eureka规定的节点数为3个,也就是最终statefulset会启动3个Pod副本组成Eureka集群。
4.2.1.修改Eureka配置文件增加各节点地址
Eureka集群节点每个Pod副本服务发现地址如下,需要将其配置在eureka中
http://eureka-0.eureka-service.simple-ms`
`http://eureka-0.eureka-service.simple-ms`
`http://eureka-0.eureka-service.simple-ms
http://eureka-0.eureka-service.simple-ms`
`http://eureka-0.eureka-service.simple-ms`
`http://eureka-0.eureka-service.simple-ms
修改Eureka的配置文件,采用Dns服务发现地址的形式将每个节点的地址都写在配置文件中,当Eureka启动之后,每个节点都会相互发现最终形成高可用集群。
1.修改Eureka配置文件中的节点地址
[root@binary-k8s-master1 eureka-service]\# vim src/main/resources/application-prod.yml
eureka:
server:
renewal-percent-threshold: 0.9
enable-self-preservation: false
eviction-interval-timer-in-simple-ms: 40000
instance:
hostname: 127.0.0.1
prefer-ip-address: false
client:
register-with-eureka: true
serviceUrl:
defaultZone: http://eureka-0.eureka.simple-ms:${server.port}/eureka/,http://eureka-1.eureka.simple-ms:${server.port}/eureka/,http://eureka-2.eureka.simple-ms:${server.port}/eureka/
fetch-registry: true
2.使用maven将最新的Eureka编译
[root@binary-k8s-master1 simple-microservice-dev1]\# mvn clean package -D maven.test.skip=true -P prod
注意:只能在项目最外层的目录中执行mvn clean package
1.修改Eureka配置文件中的节点地址
[root@binary-k8s-master1 eureka-service]\# vim src/main/resources/application-prod.yml
eureka:
server:
renewal-percent-threshold: 0.9
enable-self-preservation: false
eviction-interval-timer-in-simple-ms: 40000
instance:
hostname: 127.0.0.1
prefer-ip-address: false
client:
register-with-eureka: true
serviceUrl:
defaultZone: http://eureka-0.eureka.simple-ms:${server.port}/eureka/,http://eureka-1.eureka.simple-ms:${server.port}/eureka/,http://eureka-2.eureka.simple-ms:${server.port}/eureka/
fetch-registry: true
2.使用maven将最新的Eureka编译
[root@binary-k8s-master1 simple-microservice-dev1]\# mvn clean package -D maven.test.skip=true -P prod
注意:只能在项目最外层的目录中执行mvn clean package
4.2.2.调整Eureka的Dockerfile
由于我们在Eureka配置文件中配置的是以dns名称来互相发现对方,并不是以IP地址形式发现,因此我们需要配置Dockerfile,在启动程序时将配置文件中的hostname字段值修改为当前Pod的dns名称。
因为我们的Eureka是多Pod组成的高可用集群,因此不能将dns名称写死在配置文件中,不然每个Pod都是同样的地址,集群是组件不成功的,所以就需要在Dockerfile中定义程序启动命令时,将dns名称地址设置成变量,由Pod资源将变量值传入到启动命令中,程序启动时就会用Pod的dns名称地址替换掉配置文件中的IP地址。
1.修改DOckerfile文件内容
[root@binary-k8s-master1 eureka-service]\# vim Dockerfile
FROM 192.168.20.11/base/centos-java:v1
RUN mkdir /data/simple-microservice/eureka-service -p
COPY ./target/eureka-service.jar /data/simple-microservice/eureka-service/
RUN echo -e "#!/bin/bash \njava -jar -Deureka.instance.hostname=${MY_POD_NAME}.eureka.simple-ms /data/simple-microservice/eureka-service/eureka-service.jar > /data/simple-microservice/eureka-service/eureka-service.log & \ntail -f /data/simple-microservice/eureka-service/eureka-service.log" > /data/entrypoint.sh && chmod a+x /data/entrypoint.sh
EXPOSE 8080
ENTRYPOINT /data/entrypoint.sh
1.修改DOckerfile文件内容
[root@binary-k8s-master1 eureka-service]\# vim Dockerfile
FROM 192.168.20.11/base/centos-java:v1
RUN mkdir /data/simple-microservice/eureka-service -p
COPY ./target/eureka-service.jar /data/simple-microservice/eureka-service/
RUN echo -e "#!/bin/bash \njava -jar -Deureka.instance.hostname=${MY_POD_NAME}.eureka.simple-ms /data/simple-microservice/eureka-service/eureka-service.jar > /data/simple-microservice/eureka-service/eureka-service.log & \ntail -f /data/simple-microservice/eureka-service/eureka-service.log" > /data/entrypoint.sh && chmod a+x /data/entrypoint.sh
EXPOSE 8080
ENTRYPOINT /data/entrypoint.sh
启动命令单独拉出来看一下,-Deureka.instance.hostname
这个参数表示修改配置文件中hostname参数的值,${MY_POD_NAME}
这个变量使从deployment资源文件中定义的env变量,将Pod的名称赋给MY_POD_NAME变量,最后引入到启动命令中,最后在配置文件中hostname字段的值应该是eureka-0.eureka-service.simple-ms
java -jar -Deureka.instance.hostname=${MY_POD_NAME}.eureka-service.simple-ms /data/simple-microservice/eureka-service/eureka-service.jar
java -jar -Deureka.instance.hostname=${MY_POD_NAME}.eureka-service.simple-ms /data/simple-microservice/eureka-service/eureka-service.jar
4.2.3.构建Eureka服务的Docker镜像
[root@binary-k8s-master1 eureka-service]\# docker build -t 192.168.20.11/simple-microservice/eureka-service:v1 .
[root@binary-k8s-master1 eureka-service]\# docker images | grep eureka
192.168.20.11/simp-microservice/eureka-service v1 1dc64fc9ba09 About a minute ago 800MB
[root@binary-k8s-master1 eureka-service]\# docker push 192.168.20.11/simple-microservice/eureka-service:v1
[root@binary-k8s-master1 eureka-service]\# docker build -t 192.168.20.11/simple-microservice/eureka-service:v1 .
[root@binary-k8s-master1 eureka-service]\# docker images | grep eureka
192.168.20.11/simp-microservice/eureka-service v1 1dc64fc9ba09 About a minute ago 800MB
[root@binary-k8s-master1 eureka-service]\# docker push 192.168.20.11/simple-microservice/eureka-service:v1
4.2.4.编写Eureka部署YAML资源文件
1)创建命名空间
1.创建命名空间
# kubectl create ns simple-ms
namespace/simple-ms created
2.创建资源编排YAML文件存放路径
[root@binary-k8s-master1 ~]\# mkdir springcloud/simple-microservice-dev1/k8s/
1.创建命名空间
# kubectl create ns simple-ms
namespace/simple-ms created
2.创建资源编排YAML文件存放路径
[root@binary-k8s-master1 ~]\# mkdir springcloud/simple-microservice-dev1/k8s/
2)登录Harbor
可以不在每一个Node节点使用docker login登陆Harbor仓库,可以将Harbor仓库的认证信息做成secret,通过使用secret的方式连接Harbor。
[root@binary-k8s-master1 ~]\# kubectl create secret docker-registry registry-pull-secret --docker-server=192.168.20.11 --docker-username=admin --docker-password=Harbor12345 --docker-email=admin@ctnrs.com -n simple-ms
[root@binary-k8s-master1 ~]\# kubectl create secret docker-registry registry-pull-secret --docker-server=192.168.20.11 --docker-username=admin --docker-password=Harbor12345 --docker-email=admin@ctnrs.com -n simple-ms
3)编写资源YAML文件
eureka要做成集群式的服务,所以要采用statefulset控制器部署,并采用headless service保证statefulset创建的pod名称唯一性,最后通过ingress将eureka服务发布在互联网环境。
[root@binary-k8s-master1 k8s]\# vim eureka-deploy.yaml
---
#ingress资源
apiVersion: extensions/v1beta1
kind: Ingress #类型为ingress
metadata: #定义元数据信息:资源名称、资源所在的命名空间
name: eureka
namespace: simple-ms
spec:
rules: #定义ingress规则
- host: eureka.jiangxl.com #定义eureka使用的域名
http: #采用http类型的访问方式
paths: #定义程序的url路径
- path: /
backend: #针对定义的程序路径关于对应的后端service资源
serviceName: eureka
servicePort: 8080
---
#headless service资源
apiVersion: v1
kind: Service
metadata:
name: eureka
namespace: simple-ms
spec:
clusterIP: None #ClusterIP资源设置为Node就代表headless类型的service
ports: #定义暴露的eureka端口号
- port: 8080
name: eureka
selector: #标签选择器,关联后端pod
project: simple
app: eureka
---
#statefulset资源
apiVersion: apps/v1
kind: StatefulSet
metadata:
name: eureka
namespace: simple-ms
spec:
replicas: 3 #副本数
selector: #标签选择器
matchLabels:
project: simple
app: eureka
serviceName: "eureka" #service的名称,也就是将来的pod名称
template:
metadata:
labels:
project: simple
app: eureka
spec:
imagePullSecrets: #harbor镜像仓库的凭据信息
- name: registry-pull-secret
containers:
- name: eureka
image: 192.168.20.11/simple-microservice/eureka-service:v1
ports:
- protocol: TCP
containerPort: 8080
env: #增加一个环境变量,变量名为MY_POD_NAME,值为pod的名称,这个变量信息会传入dockerfile中
- name: MY_POD_NAME
valueFrom:
fieldRef:
fieldPath: metadata.name
resources: #资源配额
requests: #最小资源限制,cpu为0.5,内存为256M
cpu: 0.5
memory: 256Mi
limits: #最大资源限制,cpu为1,内存为1G
cpu: 1
memory: 1Gi
readinessProbe: #就绪性健康检查
tcpSocket: #以tcp 8080端口作为健康检查的机制
port: 8080
initialDelaySeconds: 60 #容器启动60秒后进行检测
periodSeconds: 10 #探测的频率
livenessProbe: #存活性健康检查
tcpSocket: #以tcp 8080端口作为健康检查的机制
port: 8080
initialDelaySeconds: 60 #容器启动60秒后进行检查
periodSeconds: 10 #探测的频率为10秒一次
[root@binary-k8s-master1 k8s]\# vim eureka-deploy.yaml
---
#ingress资源
apiVersion: extensions/v1beta1
kind: Ingress #类型为ingress
metadata: #定义元数据信息:资源名称、资源所在的命名空间
name: eureka
namespace: simple-ms
spec:
rules: #定义ingress规则
- host: eureka.jiangxl.com #定义eureka使用的域名
http: #采用http类型的访问方式
paths: #定义程序的url路径
- path: /
backend: #针对定义的程序路径关于对应的后端service资源
serviceName: eureka
servicePort: 8080
---
#headless service资源
apiVersion: v1
kind: Service
metadata:
name: eureka
namespace: simple-ms
spec:
clusterIP: None #ClusterIP资源设置为Node就代表headless类型的service
ports: #定义暴露的eureka端口号
- port: 8080
name: eureka
selector: #标签选择器,关联后端pod
project: simple
app: eureka
---
#statefulset资源
apiVersion: apps/v1
kind: StatefulSet
metadata:
name: eureka
namespace: simple-ms
spec:
replicas: 3 #副本数
selector: #标签选择器
matchLabels:
project: simple
app: eureka
serviceName: "eureka" #service的名称,也就是将来的pod名称
template:
metadata:
labels:
project: simple
app: eureka
spec:
imagePullSecrets: #harbor镜像仓库的凭据信息
- name: registry-pull-secret
containers:
- name: eureka
image: 192.168.20.11/simple-microservice/eureka-service:v1
ports:
- protocol: TCP
containerPort: 8080
env: #增加一个环境变量,变量名为MY_POD_NAME,值为pod的名称,这个变量信息会传入dockerfile中
- name: MY_POD_NAME
valueFrom:
fieldRef:
fieldPath: metadata.name
resources: #资源配额
requests: #最小资源限制,cpu为0.5,内存为256M
cpu: 0.5
memory: 256Mi
limits: #最大资源限制,cpu为1,内存为1G
cpu: 1
memory: 1Gi
readinessProbe: #就绪性健康检查
tcpSocket: #以tcp 8080端口作为健康检查的机制
port: 8080
initialDelaySeconds: 60 #容器启动60秒后进行检测
periodSeconds: 10 #探测的频率
livenessProbe: #存活性健康检查
tcpSocket: #以tcp 8080端口作为健康检查的机制
port: 8080
initialDelaySeconds: 60 #容器启动60秒后进行检查
periodSeconds: 10 #探测的频率为10秒一次
4.2.5.在K8S集群中部署Eureka注册中心
1.部署eureka服务
[root@binary-k8s-master1 k8s]\# kubectl apply -f eureka-deploy.yaml
2.查看部署的资源状态
[root@binary-k8s-master1 k8s]\# kubectl get all -n simple-ms
NAME READY STATUS RESTARTS AGE
pod/eureka-0 1/1 Running 1 12h
pod/eureka-1 1/1 Running 0 12h
pod/eureka-2 1/1 Running 0 12h
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
service/eureka ClusterIP None <none> 8080/TCP 13h
NAME READY AGE
statefulset.apps/eureka 3/3 13h
3.查看eureka的ingress状态
[root@binary-k8s-master1 k8s]\# kubectl get ingress -n simple-ms
NAME CLASS HOSTS ADDRESS PORTS AGE
eureka <none> eureka.jiangxl.com 10.0.0.107 80 13h
1.部署eureka服务
[root@binary-k8s-master1 k8s]\# kubectl apply -f eureka-deploy.yaml
2.查看部署的资源状态
[root@binary-k8s-master1 k8s]\# kubectl get all -n simple-ms
NAME READY STATUS RESTARTS AGE
pod/eureka-0 1/1 Running 1 12h
pod/eureka-1 1/1 Running 0 12h
pod/eureka-2 1/1 Running 0 12h
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
service/eureka ClusterIP None <none> 8080/TCP 13h
NAME READY AGE
statefulset.apps/eureka 3/3 13h
3.查看eureka的ingress状态
[root@binary-k8s-master1 k8s]\# kubectl get ingress -n simple-ms
NAME CLASS HOSTS ADDRESS PORTS AGE
eureka <none> eureka.jiangxl.com 10.0.0.107 80 13h
首先在本地hosts文件中绑定eureka域名解析,然后在浏览器中输入eureka.jiangxl.com访问eureka服务。
4.3.配置各个微服务程序连接Eureka集群的地址
在程序的配置文件中填写eureka集群所有节点的服务发现地址
1.gateway-service配置文件
# vim gateway-service/src/main/resources/application-prod.yml
······
eureka:
instance:
prefer-ip-address: true
client:
register-with-eureka: true
fetch-registry: true
service-url:
defaultZone: http://eureka-0.eureka.simple-ms:8080/eureka,http://eureka-1.eureka.simple-ms:8080/eureka,http://eureka-2.eureka.simple-ms:8080/eureka
# vim gateway-service/src/main/resources/application-prod.yml
······
eureka:
instance:
prefer-ip-address: true
client:
register-with-eureka: true
fetch-registry: true
service-url:
defaultZone: http://eureka-0.eureka.simple-ms:8080/eureka,http://eureka-1.eureka.simple-ms:8080/eureka,http://eureka-2.eureka.simple-ms:8080/eureka
2.order-service配置文件
# vim order-service/order-service-biz/src/main/resources/application-prod.yml
spring:
datasource:
url: jdbc:mysql://192.168.20.11:3306/tb_order?characterEncoding=utf-8
username: root
password: 123456
driver-class-name: com.mysql.jdbc.Driver
eureka:
instance:
prefer-ip-address: true
client:
register-with-eureka: true
fetch-registry: true
service-url:
defaultZone: http://eureka-0.eureka.simple-ms:8080/eureka,http://eureka-1.eureka.simple-ms:8080/eureka,http://eureka-2.eureka.simple-ms:8080/eureka
# vim order-service/order-service-biz/src/main/resources/application-prod.yml
spring:
datasource:
url: jdbc:mysql://192.168.20.11:3306/tb_order?characterEncoding=utf-8
username: root
password: 123456
driver-class-name: com.mysql.jdbc.Driver
eureka:
instance:
prefer-ip-address: true
client:
register-with-eureka: true
fetch-registry: true
service-url:
defaultZone: http://eureka-0.eureka.simple-ms:8080/eureka,http://eureka-1.eureka.simple-ms:8080/eureka,http://eureka-2.eureka.simple-ms:8080/eureka
3.product-service配置文件
# vim product-service/product-service-biz/src/main/resources/application-prod.yml
spring:
datasource:
url: jdbc:mysql://192.168.20.11:3306/tb_product?characterEncoding=utf-8
username: root
password: 123456
driver-class-name: com.mysql.jdbc.Driver
eureka:
instance:
prefer-ip-address: true
client:
register-with-eureka: true
fetch-registry: true
service-url:
defaultZone: http://eureka-0.eureka.simple-ms:8080/eureka,http://eureka-1.eureka.simple-ms:8080/eureka,http://eureka-2.eureka.simple-ms:8080/eureka
# vim product-service/product-service-biz/src/main/resources/application-prod.yml
spring:
datasource:
url: jdbc:mysql://192.168.20.11:3306/tb_product?characterEncoding=utf-8
username: root
password: 123456
driver-class-name: com.mysql.jdbc.Driver
eureka:
instance:
prefer-ip-address: true
client:
register-with-eureka: true
fetch-registry: true
service-url:
defaultZone: http://eureka-0.eureka.simple-ms:8080/eureka,http://eureka-1.eureka.simple-ms:8080/eureka,http://eureka-2.eureka.simple-ms:8080/eureka
4.stock-service配置文件
# vim stock-service/stock-service-biz/src/main/resources/application-prod.yml
spring:
datasource:
url: jdbc:mysql://192.168.20.11:3306/tb_stock?characterEncoding=utf-8
username: root
password: 123456
driver-class-name: com.mysql.jdbc.Driver
eureka:
instance:
prefer-ip-address: true
client:
register-with-eureka: true
fetch-registry: true
service-url:
defaultZone: http://eureka-0.eureka.simple-ms:8080/eureka,http://eureka-1.eureka.simple-ms:8080/eureka,http://eureka-2.eureka.simple-ms:8080/eureka
# vim stock-service/stock-service-biz/src/main/resources/application-prod.yml
spring:
datasource:
url: jdbc:mysql://192.168.20.11:3306/tb_stock?characterEncoding=utf-8
username: root
password: 123456
driver-class-name: com.mysql.jdbc.Driver
eureka:
instance:
prefer-ip-address: true
client:
register-with-eureka: true
fetch-registry: true
service-url:
defaultZone: http://eureka-0.eureka.simple-ms:8080/eureka,http://eureka-1.eureka.simple-ms:8080/eureka,http://eureka-2.eureka.simple-ms:8080/eureka
5.portal-service配置文件
# vim portal-service/src/main/resources/application-prod.yml
eureka:
instance:
prefer-ip-address: true
client:
service-url:
defaultZone: http://eureka-0.eureka.simple-ms:8080/eureka,http://eureka-1.eureka.simple-ms:8080/eureka,http://eureka-2.eureka.simple-ms:8080/eureka
register-with-eureka: true
fetch-registry: true
······
# vim portal-service/src/main/resources/application-prod.yml
eureka:
instance:
prefer-ip-address: true
client:
service-url:
defaultZone: http://eureka-0.eureka.simple-ms:8080/eureka,http://eureka-1.eureka.simple-ms:8080/eureka,http://eureka-2.eureka.simple-ms:8080/eureka
register-with-eureka: true
fetch-registry: true
······
4.4.将各个程序代码编译成可部署的程序
[root@binary-k8s-master1 simple-microservice-dev1]\# mvn clean package -D maven.test.skip=true -P prod
[root@binary-k8s-master1 simple-microservice-dev1]\# mvn clean package -D maven.test.skip=true -P prod
4.3.中部署Gateway网关服务
4.3.1.将Gateway程序打包成docker镜像
[root@binary-k8s-master1 gateway-service]\# docker build -t 192.168.20.11/simple-microservice/gateway-service:v1 .
[root@binary-k8s-master1 gateway-service]\# docker build -t 192.168.20.11/simple-microservice/gateway-service:v1 .
4.3.2.在K8S集群中部署Gateway网关服务
所有程序的资源YAML文件都存放在/root/springcloud/simple-microservice-dev1/k8s路径。
微服务程序的话就不需要有状态应用部署了,直接采用deployment无状态应用部署的控制器,字段配置和eureka服务的statefulset配置参数几乎是一样的,微服务的各个程序也都采用有deployment控制器去部署,每个程序的yaml文件也只有名字、使用的镜像不同,其余所有参数均一致。
gateway是网关服务,需要配置一个ingress和service资源,将其发布在集群之外。
字段配置和statefulset的一致,不再做注释信息
# vim gateway-deploy.yaml
---
#ingress资源YAML
apiVersion: extensions/v1beta1
kind: Ingress
metadata:
name: gateway
namespace: simple-ms
spec:
rules:
- host: gateway.jiangxl.com
http:
paths:
- path: /
backend:
serviceName: gateway
servicePort: 8080
---
#service资源YAML
apiVersion: v1
kind: Service
metadata:
name: gateway
namespace: simple-ms
spec:
ports:
- port: 8080
name: gateway
selector:
project: simple
app: gateway
---
#deployment资源YAML
apiVersion: apps/v1
kind: Deployment
metadata:
name: gateway
namespace: simple-ms
spec:
replicas: 1 #副本数设置1个,后续再加
selector:
matchLabels:
project: simple
app: gateway
template:
metadata:
labels:
project: simple
app: gateway
spec:
imagePullSecrets:
- name: registry-pull-secret
containers:
- name: gateway
image: 192.168.20.11/simple-microservice/gateway-service:v1
imagePullPolicy: Always
ports:
- protocol: TCP
containerPort: 8080
resources:
requests:
cpu: 0.5
memory: 256Mi
limits:
cpu: 1
memory: 1Gi
readinessProbe:
tcpSocket:
port: 8080
initialDelaySeconds: 60
periodSeconds: 10
livenessProbe:
tcpSocket:
port: 8080
initialDelaySeconds: 60
periodSeconds: 10
# vim gateway-deploy.yaml
---
#ingress资源YAML
apiVersion: extensions/v1beta1
kind: Ingress
metadata:
name: gateway
namespace: simple-ms
spec:
rules:
- host: gateway.jiangxl.com
http:
paths:
- path: /
backend:
serviceName: gateway
servicePort: 8080
---
#service资源YAML
apiVersion: v1
kind: Service
metadata:
name: gateway
namespace: simple-ms
spec:
ports:
- port: 8080
name: gateway
selector:
project: simple
app: gateway
---
#deployment资源YAML
apiVersion: apps/v1
kind: Deployment
metadata:
name: gateway
namespace: simple-ms
spec:
replicas: 1 #副本数设置1个,后续再加
selector:
matchLabels:
project: simple
app: gateway
template:
metadata:
labels:
project: simple
app: gateway
spec:
imagePullSecrets:
- name: registry-pull-secret
containers:
- name: gateway
image: 192.168.20.11/simple-microservice/gateway-service:v1
imagePullPolicy: Always
ports:
- protocol: TCP
containerPort: 8080
resources:
requests:
cpu: 0.5
memory: 256Mi
limits:
cpu: 1
memory: 1Gi
readinessProbe:
tcpSocket:
port: 8080
initialDelaySeconds: 60
periodSeconds: 10
livenessProbe:
tcpSocket:
port: 8080
initialDelaySeconds: 60
periodSeconds: 10
4.3.3.在K8S集群中部署Gateway网关服务
1.部署网关服务
[root@binary-k8s-master1 k8s]\# kubectl apply -f gateway-deploy.yaml
2.查看gateway pod资源的状态
[root@binary-k8s-master1 k8s]\# kubectl get pod -n simple-ms
NAME READY STATUS RESTARTS AGE
eureka-0 1/1 Running 1 2d12h
eureka-1 1/1 Running 0 2d12h
eureka-2 1/1 Running 0 2d12h
gateway-7c687f6c45-mcwpf 1/1 Running 0 2m2s
3.查看deployment和svc的资源状态
[root@binary-k8s-master1 k8s]\# kubectl get deploy,svc -n simple-ms
NAME READY UP-TO-DATE AVAILABLE AGE
deployment.apps/gateway 1/1 1 1 17h
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
service/eureka ClusterIP None <none> 8080/TCP 2d13h
service/gateway ClusterIP 10.0.0.188 <none> 8080/TCP 17h
4.查看gateway ingress资源状态
[root@binary-k8s-master1 k8s]\# kubectl get ingress -n simple-ms
NAME CLASS HOSTS ADDRESS PORTS AGE
eureka <none> eureka.jiangxl.com 10.0.0.107 80 2d13h
gateway <none> gateway.jiangxl.com 10.0.0.107 80 17h
1.部署网关服务
[root@binary-k8s-master1 k8s]\# kubectl apply -f gateway-deploy.yaml
2.查看gateway pod资源的状态
[root@binary-k8s-master1 k8s]\# kubectl get pod -n simple-ms
NAME READY STATUS RESTARTS AGE
eureka-0 1/1 Running 1 2d12h
eureka-1 1/1 Running 0 2d12h
eureka-2 1/1 Running 0 2d12h
gateway-7c687f6c45-mcwpf 1/1 Running 0 2m2s
3.查看deployment和svc的资源状态
[root@binary-k8s-master1 k8s]\# kubectl get deploy,svc -n simple-ms
NAME READY UP-TO-DATE AVAILABLE AGE
deployment.apps/gateway 1/1 1 1 17h
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
service/eureka ClusterIP None <none> 8080/TCP 2d13h
service/gateway ClusterIP 10.0.0.188 <none> 8080/TCP 17h
4.查看gateway ingress资源状态
[root@binary-k8s-master1 k8s]\# kubectl get ingress -n simple-ms
NAME CLASS HOSTS ADDRESS PORTS AGE
eureka <none> eureka.jiangxl.com 10.0.0.107 80 2d13h
gateway <none> gateway.jiangxl.com 10.0.0.107 80 17h
4.4.编写快速部署微服务到K8S集群的脚本
由于微服务程序众多,并且每个程序在K8S集群部署的流程也基本一样,因此我们编写一个部署脚本,只要准备好程序的jar包以及资源YAML文件就可以将其一键部署到K8S集群,当前也可以通过Helm来实现。
1.脚本内容
实现逻辑:
1.将harbor仓库的地址、所有微服务的名称、k8s yaml文件的路径、微服务代码所在的路径都定义成环境变量。
2.如果只想部署其中一个微服务的话,通过位置传参的方式来实现,也就是再定义一个微服务名称的变量,变量值为KaTeX parse error: Expected '}', got 'EOF' at end of input: {1:-{service_list}},将要部署的微服务名称通过位置传参的方式,将值传入到service_list变量名里,覆盖原service_list变量值,实现单个微服务的部署,如果没有使用位置参数,那么默认部署全部的微服务。
3.进入微服务代码所在的路径。然后通过for循环的方式,遍历service_list变量中的各个微服务,然后本次循环的微服务代码路径,判断是否有包含biz的目录,如果有biz目录则切入到biz目录,然后通过变量的形式定义镜像名称,构建镜像、推送至镜像仓库。
4.将最新打包好的docker进在yaml文件中通过sed命令替换,最后在k8s集群中部署这个微服务。
${1:-${service_list}}
:这个的意思就是如果$1
位置参数没有传入值,那么service_list变量的值就是$service_list
变量的值,相当于一个判断。
本脚本可以自动化完成微服务到K8S集群的一个部署,包括程序的编译、打包成docker镜像、推荐镜像到Harbor仓库,最后将程序部署到K8S集群
#!/bin/bash
registry_addr=192.168.20.11
service_list="eureka-service gateway-service order-service product-service stock-service portal-service"
service_list=${1:-${service_list}}
current_dir=/root/springcloud/simple-microservice-dev1/k8s
service_dir=$(dirname $current_dir)
cd $service_dir
#mvn clean package -Dmaven.test.skip=true
for service in $service_list; do
cd $service_dir/$service
biz_count=`ls | grep biz | wc -l`
if [ $biz_count -eq 1 ];then
cd ${service}-biz
fi
service_name=`echo $service | awk -F- '{print $1}'`
image_name=${registry_addr}/simple-microservice/${service}:v1
yaml_name=${service_name}-deploy.yaml
docker build -t ${image_name} .
docker push ${image_name}
sed -i -r "s#(image: )(.*)#\1$image_name#" ${current_dir}/${yaml_name}
kubectl apply -f ${current_dir}/${yaml_name}
done
#!/bin/bash
registry_addr=192.168.20.11
service_list="eureka-service gateway-service order-service product-service stock-service portal-service"
service_list=${1:-${service_list}}
current_dir=/root/springcloud/simple-microservice-dev1/k8s
service_dir=$(dirname $current_dir)
cd $service_dir
#mvn clean package -Dmaven.test.skip=true
for service in $service_list; do
cd $service_dir/$service
biz_count=`ls | grep biz | wc -l`
if [ $biz_count -eq 1 ];then
cd ${service}-biz
fi
service_name=`echo $service | awk -F- '{print $1}'`
image_name=${registry_addr}/simple-microservice/${service}:v1
yaml_name=${service_name}-deploy.yaml
docker build -t ${image_name} .
docker push ${image_name}
sed -i -r "s#(image: )(.*)#\1$image_name#" ${current_dir}/${yaml_name}
kubectl apply -f ${current_dir}/${yaml_name}
done
2.脚本使用方式
执行本脚本前提是必须先将程序部署的YAML文件准备好,放在/root/springcloud/simple-microservice-dev1/k8s路径
脚本的用法:
sh deploy_k8s.sh 微服务名称
脚本的用法:
sh deploy_k8s.sh 微服务名称
4.5.使用脚本部署Order订单服务
4.5.1.编写资源部署YAML文件
# vim order-deploy.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: order
namespace: simple-ms
spec:
replicas: 1
selector:
matchLabels:
project: simple
app: order
template:
metadata:
labels:
project: simple
app: order
spec:
imagePullSecrets:
- name: registry-pull-secret
containers:
- name: order
image: 192.168.20.11/simple-microservice/order-service:v1
imagePullPolicy: Always
ports:
- protocol: TCP
containerPort: 8080
resources:
requests:
cpu: 0.5
memory: 256Mi
limits:
cpu: 1
memory: 1Gi
readinessProbe:
tcpSocket:
port: 8080
initialDelaySeconds: 60
periodSeconds: 10
livenessProbe:
tcpSocket:
port: 8080
initialDelaySeconds: 60
periodSeconds: 10
# vim order-deploy.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: order
namespace: simple-ms
spec:
replicas: 1
selector:
matchLabels:
project: simple
app: order
template:
metadata:
labels:
project: simple
app: order
spec:
imagePullSecrets:
- name: registry-pull-secret
containers:
- name: order
image: 192.168.20.11/simple-microservice/order-service:v1
imagePullPolicy: Always
ports:
- protocol: TCP
containerPort: 8080
resources:
requests:
cpu: 0.5
memory: 256Mi
limits:
cpu: 1
memory: 1Gi
readinessProbe:
tcpSocket:
port: 8080
initialDelaySeconds: 60
periodSeconds: 10
livenessProbe:
tcpSocket:
port: 8080
initialDelaySeconds: 60
periodSeconds: 10
4.5.2在K8S集群中部署Order订单服务
1.使用脚本将Order服务部署到K8S集群
[root@binary-k8s-master1 k8s]\# sh deploy_k8s.sh order-service
执行完脚本,order程序的docker镜像已经制作好并且推送到了harbor仓库,已经使用sed命令将镜像的名称在yaml中替换成了最新的镜像,部署到了K8S集群
[root@binary-k8s-master1 k8s]\# sh deploy_k8s.sh order-service
执行完脚本,order程序的docker镜像已经制作好并且推送到了harbor仓库,已经使用sed命令将镜像的名称在yaml中替换成了最新的镜像,部署到了K8S集群
2.查看程序资源的状态
1.查看pod的状态
[root@binary-k8s-master1 k8s]\# kubectl get all -n simple-ms
NAME READY STATUS RESTARTS AGE
pod/eureka-0 1/1 Running 1 2d15h
pod/eureka-1 1/1 Running 0 2d15h
pod/eureka-2 1/1 Running 0 2d15h
pod/gateway-8665d5d56-tx2m8 1/1 Running 0 39m
pod/order-6694f4c474-cp5d6 1/1 Running 0 35m
2.查看deployment资源的状态
[root@binary-k8s-master1 k8s]\# kubectl get deploy -n simple-ms
NAME READY UP-TO-DATE AVAILABLE AGE
gateway 1/1 1 1 21h
order 1/1 1 1 36m
1.查看pod的状态
[root@binary-k8s-master1 k8s]\# kubectl get all -n simple-ms
NAME READY STATUS RESTARTS AGE
pod/eureka-0 1/1 Running 1 2d15h
pod/eureka-1 1/1 Running 0 2d15h
pod/eureka-2 1/1 Running 0 2d15h
pod/gateway-8665d5d56-tx2m8 1/1 Running 0 39m
pod/order-6694f4c474-cp5d6 1/1 Running 0 35m
2.查看deployment资源的状态
[root@binary-k8s-master1 k8s]\# kubectl get deploy -n simple-ms
NAME READY UP-TO-DATE AVAILABLE AGE
gateway 1/1 1 1 21h
order 1/1 1 1 36m
4.6.使用脚本部署Product商品服务
4.6.1.编写资源部署YAML文件
# vim product-deploy.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: product
namespace: simple-ms
spec:
replicas: 1
selector:
matchLabels:
project: simple
app: product
template:
metadata:
labels:
project: simple
app: product
spec:
imagePullSecrets:
- name: registry-pull-secret
containers:
- name: product
image: 192.168.20.11/simple-microservice/product-service:v1
imagePullPolicy: Always
ports:
- protocol: TCP
containerPort: 8080
resources:
requests:
cpu: 0.5
memory: 256Mi
limits:
cpu: 1
memory: 1Gi
readinessProbe:
tcpSocket:
port: 8080
initialDelaySeconds: 60
periodSeconds: 10
livenessProbe:
tcpSocket:
port: 8080
initialDelaySeconds: 60
periodSeconds: 10
# vim product-deploy.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: product
namespace: simple-ms
spec:
replicas: 1
selector:
matchLabels:
project: simple
app: product
template:
metadata:
labels:
project: simple
app: product
spec:
imagePullSecrets:
- name: registry-pull-secret
containers:
- name: product
image: 192.168.20.11/simple-microservice/product-service:v1
imagePullPolicy: Always
ports:
- protocol: TCP
containerPort: 8080
resources:
requests:
cpu: 0.5
memory: 256Mi
limits:
cpu: 1
memory: 1Gi
readinessProbe:
tcpSocket:
port: 8080
initialDelaySeconds: 60
periodSeconds: 10
livenessProbe:
tcpSocket:
port: 8080
initialDelaySeconds: 60
periodSeconds: 10
4.6.2.在K8S集群中熟不熟Product商品服务
1.使用脚本在K8S集群中部署product服务
[root@binary-k8s-master1 k8s]\# sh deploy_k8s.sh product-service
[root@binary-k8s-master1 k8s]\# sh deploy_k8s.sh product-service
2.查看程序资源的状态
1.查看pod的状态
[root@binary-k8s-master1 k8s]\# kubectl get pod -n simple-ms
NAME READY STATUS RESTARTS AGE
eureka-1 1/1 Running 3 2d16h
eureka-2 1/1 Running 2 2d16h
gateway-8665d5d56-tx2m8 1/1 Running 2 74m
order-6694f4c474-cp5d6 1/1 Running 4 69m
product-77c8bb6847-vbxnr 1/1 Running 2 21m
2.查看deployment资源的状态
[root@binary-k8s-master1 k8s]\# kubectl get deploy -n simple-ms
NAME READY UP-TO-DATE AVAILABLE AGE
gateway 1/1 1 1 21h
order 1/1 1 1 72m
product 1/1 1 1 24m
1.查看pod的状态
[root@binary-k8s-master1 k8s]\# kubectl get pod -n simple-ms
NAME READY STATUS RESTARTS AGE
eureka-1 1/1 Running 3 2d16h
eureka-2 1/1 Running 2 2d16h
gateway-8665d5d56-tx2m8 1/1 Running 2 74m
order-6694f4c474-cp5d6 1/1 Running 4 69m
product-77c8bb6847-vbxnr 1/1 Running 2 21m
2.查看deployment资源的状态
[root@binary-k8s-master1 k8s]\# kubectl get deploy -n simple-ms
NAME READY UP-TO-DATE AVAILABLE AGE
gateway 1/1 1 1 21h
order 1/1 1 1 72m
product 1/1 1 1 24m
4.7.使用脚本部署Stock库存服务
4.7.1.编写资源部署YAML文件
# vim product-deploy.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: product
namespace: simple-ms
spec:
replicas: 1
selector:
matchLabels:
project: simple
app: product
template:
metadata:
labels:
project: simple
app: product
spec:
imagePullSecrets:
- name: registry-pull-secret
containers:
- name: product
image: 192.168.20.11/simple-microservice/product-service:v1
imagePullPolicy: Always
ports:
- protocol: TCP
containerPort: 8080
resources:
requests:
cpu: 0.5
memory: 256Mi
limits:
cpu: 1
memory: 1Gi
readinessProbe:
tcpSocket:
port: 8080
initialDelaySeconds: 60
periodSeconds: 10
livenessProbe:
tcpSocket:
port: 8080
initialDelaySeconds: 60
periodSeconds: 10
# vim product-deploy.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: product
namespace: simple-ms
spec:
replicas: 1
selector:
matchLabels:
project: simple
app: product
template:
metadata:
labels:
project: simple
app: product
spec:
imagePullSecrets:
- name: registry-pull-secret
containers:
- name: product
image: 192.168.20.11/simple-microservice/product-service:v1
imagePullPolicy: Always
ports:
- protocol: TCP
containerPort: 8080
resources:
requests:
cpu: 0.5
memory: 256Mi
limits:
cpu: 1
memory: 1Gi
readinessProbe:
tcpSocket:
port: 8080
initialDelaySeconds: 60
periodSeconds: 10
livenessProbe:
tcpSocket:
port: 8080
initialDelaySeconds: 60
periodSeconds: 10
4.7.2.在K8S集群中部署Stock库存服务
1.使用脚本在K8S集群中部署stock服务
[root@binary-k8s-master1 k8s]\# sh deploy_k8s.sh stock-service
[root@binary-k8s-master1 k8s]\# sh deploy_k8s.sh stock-service
2.查看程序资源的状态
1.查看pod的状态
[root@binary-k8s-master1 ~]\# kubectl get pod -n simple-ms
NAME READY STATUS RESTARTS AGE
eureka-0 1/1 Running 5 2d16h
eureka-1 1/1 Running 3 2d16h
eureka-2 1/1 Running 2 2d16h
gateway-8665d5d56-tx2m8 1/1 Running 2 112m
order-6694f4c474-cp5d6 1/1 Running 4 107m
product-77c8bb6847-vbxnr 1/1 Running 2 59m
stock-6cf98bb445-mxf7b 1/1 Running 0 30m
2.查看deployment资源的状态
[root@binary-k8s-master1 ~]\# kubectl get deploy -n simple-ms
NAME READY UP-TO-DATE AVAILABLE AGE
gateway 1/1 1 1 22h
order 1/1 1 1 108m
product 1/1 1 1 60m
stock 2/2 2 2 31m
1.查看pod的状态
[root@binary-k8s-master1 ~]\# kubectl get pod -n simple-ms
NAME READY STATUS RESTARTS AGE
eureka-0 1/1 Running 5 2d16h
eureka-1 1/1 Running 3 2d16h
eureka-2 1/1 Running 2 2d16h
gateway-8665d5d56-tx2m8 1/1 Running 2 112m
order-6694f4c474-cp5d6 1/1 Running 4 107m
product-77c8bb6847-vbxnr 1/1 Running 2 59m
stock-6cf98bb445-mxf7b 1/1 Running 0 30m
2.查看deployment资源的状态
[root@binary-k8s-master1 ~]\# kubectl get deploy -n simple-ms
NAME READY UP-TO-DATE AVAILABLE AGE
gateway 1/1 1 1 22h
order 1/1 1 1 108m
product 1/1 1 1 60m
stock 2/2 2 2 31m
4.8.使用脚本部署Portal前端首页
4.8.1.编写资源部署YAML文件
# vim portal-deploy.yaml
apiVersion: extensions/v1beta1
kind: Ingress
metadata:
name: portal
namespace: simple-ms
spec:
rules:
- host: portal.jiangxl.com
http:
paths:
- path: /
backend:
serviceName: portal
servicePort: 8080
---
apiVersion: v1
kind: Service
metadata:
name: portal
namespace: simple-ms
spec:
ports:
- port: 8080
name: portal
selector:
project: simple
app: portal
---
apiVersion: apps/v1
kind: Deployment
metadata:
name: portal
namespace: simple-ms
spec:
replicas: 1
selector:
matchLabels:
project: simple
app: portal
template:
metadata:
labels:
project: simple
app: portal
spec:
imagePullSecrets:
- name: registry-pull-secret
containers:
- name: portal
image: 192.168.20.11/simple-microservice/portal-service:v1
imagePullPolicy: Always
ports:
- protocol: TCP
containerPort: 8080
resources:
requests:
cpu: 0.5
memory: 256Mi
limits:
cpu: 1
memory: 1Gi
readinessProbe:
tcpSocket:
port: 8080
initialDelaySeconds: 60
periodSeconds: 10
livenessProbe:
tcpSocket:
port: 8080
initialDelaySeconds: 60
periodSeconds: 10
# vim portal-deploy.yaml
apiVersion: extensions/v1beta1
kind: Ingress
metadata:
name: portal
namespace: simple-ms
spec:
rules:
- host: portal.jiangxl.com
http:
paths:
- path: /
backend:
serviceName: portal
servicePort: 8080
---
apiVersion: v1
kind: Service
metadata:
name: portal
namespace: simple-ms
spec:
ports:
- port: 8080
name: portal
selector:
project: simple
app: portal
---
apiVersion: apps/v1
kind: Deployment
metadata:
name: portal
namespace: simple-ms
spec:
replicas: 1
selector:
matchLabels:
project: simple
app: portal
template:
metadata:
labels:
project: simple
app: portal
spec:
imagePullSecrets:
- name: registry-pull-secret
containers:
- name: portal
image: 192.168.20.11/simple-microservice/portal-service:v1
imagePullPolicy: Always
ports:
- protocol: TCP
containerPort: 8080
resources:
requests:
cpu: 0.5
memory: 256Mi
limits:
cpu: 1
memory: 1Gi
readinessProbe:
tcpSocket:
port: 8080
initialDelaySeconds: 60
periodSeconds: 10
livenessProbe:
tcpSocket:
port: 8080
initialDelaySeconds: 60
periodSeconds: 10
4.8.2.在K8S集群中熟不熟Portal存服务
1.使用脚本在K8S集群中部署portal服务
[root@binary-k8s-master1 k8s]\# sh deploy_k8s.sh portal-service
[root@binary-k8s-master1 k8s]\# sh deploy_k8s.sh portal-service
2.查看程序资源的状态
到此为止,所有的simple电商平台的微服务程序以及全部部署完毕。
1.查看所有微服务的程序
[root@binary-k8s-master1 k8s]\# kubectl get all -n simple-ms
NAME READY STATUS RESTARTS AGE
pod/eureka-0 1/1 Running 5 2d16h
pod/eureka-1 1/1 Running 3 2d16h
pod/eureka-2 1/1 Running 2 2d16h
pod/gateway-8665d5d56-tx2m8 1/1 Running 2 117m
pod/order-6694f4c474-cp5d6 1/1 Running 4 112m
pod/portal-d5d55b784-7kntp 1/1 Running 0 2m53s
pod/product-77c8bb6847-vbxnr 1/1 Running 2 64m
pod/stock-6cf98bb445-mxf7b 1/1 Running 0 35m
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
service/eureka ClusterIP None <none> 8080/TCP 2d18h
service/gateway ClusterIP 10.0.0.188 <none> 8080/TCP 22h
service/portal ClusterIP 10.0.0.190 <none> 8080/TCP 2m53s
NAME READY UP-TO-DATE AVAILABLE AGE
deployment.apps/gateway 1/1 1 1 22h
deployment.apps/order 1/1 1 1 112m
deployment.apps/portal 1/1 1 1 2m53s
deployment.apps/product 1/1 1 1 65m
deployment.apps/stock 1/1 1 1 35m
NAME DESIRED CURRENT READY AGE
replicaset.apps/gateway-8665d5d56 1 1 1 117m
replicaset.apps/order-6694f4c474 1 1 1 112m
replicaset.apps/portal-d5d55b784 1 1 1 2m53s
replicaset.apps/product-77c8bb6847 1 1 1 64m
replicaset.apps/stock-6cf98bb445 1 1 1 35m
NAME READY AGE
statefulset.apps/eureka 3/3 2d18h
2.查看所有微服务的ingress资源
[root@binary-k8s-master1 k8s]\# kubectl get ingress -n simple-ms
NAME CLASS HOSTS ADDRESS PORTS AGE
eureka <none> eureka.jiangxl.com 10.0.0.107 80 2d18h
gateway <none> gateway.jiangxl.com 10.0.0.107 80 22h
portal <none> portal.jiangxl.com 10.0.0.107 80 3m16s
1.查看所有微服务的程序
[root@binary-k8s-master1 k8s]\# kubectl get all -n simple-ms
NAME READY STATUS RESTARTS AGE
pod/eureka-0 1/1 Running 5 2d16h
pod/eureka-1 1/1 Running 3 2d16h
pod/eureka-2 1/1 Running 2 2d16h
pod/gateway-8665d5d56-tx2m8 1/1 Running 2 117m
pod/order-6694f4c474-cp5d6 1/1 Running 4 112m
pod/portal-d5d55b784-7kntp 1/1 Running 0 2m53s
pod/product-77c8bb6847-vbxnr 1/1 Running 2 64m
pod/stock-6cf98bb445-mxf7b 1/1 Running 0 35m
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
service/eureka ClusterIP None <none> 8080/TCP 2d18h
service/gateway ClusterIP 10.0.0.188 <none> 8080/TCP 22h
service/portal ClusterIP 10.0.0.190 <none> 8080/TCP 2m53s
NAME READY UP-TO-DATE AVAILABLE AGE
deployment.apps/gateway 1/1 1 1 22h
deployment.apps/order 1/1 1 1 112m
deployment.apps/portal 1/1 1 1 2m53s
deployment.apps/product 1/1 1 1 65m
deployment.apps/stock 1/1 1 1 35m
NAME DESIRED CURRENT READY AGE
replicaset.apps/gateway-8665d5d56 1 1 1 117m
replicaset.apps/order-6694f4c474 1 1 1 112m
replicaset.apps/portal-d5d55b784 1 1 1 2m53s
replicaset.apps/product-77c8bb6847 1 1 1 64m
replicaset.apps/stock-6cf98bb445 1 1 1 35m
NAME READY AGE
statefulset.apps/eureka 3/3 2d18h
2.查看所有微服务的ingress资源
[root@binary-k8s-master1 k8s]\# kubectl get ingress -n simple-ms
NAME CLASS HOSTS ADDRESS PORTS AGE
eureka <none> eureka.jiangxl.com 10.0.0.107 80 2d18h
gateway <none> gateway.jiangxl.com 10.0.0.107 80 22h
portal <none> portal.jiangxl.com 10.0.0.107 80 3m16s
4.9.查看Eureka注册中心中各个微服务的信息
每个微服务程序均在Eureka注册中心中注册完成。
4.10.使用simple微服务电商平台
到4.8步骤之后,我们的simple微服务电商平台已经全部部署在K8S集群了,可以在本机hosts里绑定simple微服务的域名,操作电商平台。
这个微服务平台的背景图其实是京东的一张图而已,是不可以点的。
只能在商品服务中查询能购买的商品,然后进行购买,最后在订单服务中查询。
搜索商品时会调用product商品微服务,下单后,查询订单服务时会调用order订单服务。
点击查询商品服务,我们来下单一个美女观察一下效果,下单后会提示下单成功。
点击查询订单服务就可以看到我吗下的订单了
5.基于K8S集群的微服务的扩容与升级
5.1.微服务的扩容
微服务在物理机环境运行时,网站并发量高的情况下,扩容只能去购买云服务器,然后搭建一套一模一样的服务,手动组成负载均衡集群,增加一台其实还好,如果要增加几十台,太费人工成本了。
当微服务迁移到K8S集群之后,扩容其实变得非常简单,只需要增加程序Pod的副本数就可以完成程序的扩容。
命令如下,以扩展order商品微服务为例。
将order商品服务的副本数调整为3个,原本是1个,现在扩容到3个
[root@binary-k8s-master1 k8s]\# kubectl scale deploy order --replicas=3 -n simple-ms
deployment.apps/order scaled
可以看到order订单服务目前已经有3个pod副本了
[root@binary-k8s-master1 ~]\# kubectl get pod -n simple-ms
NAME READY STATUS RESTARTS AGE
eureka-0 1/1 Running 1 6m15s
eureka-1 1/1 Running 0 7m59s
eureka-2 1/1 Running 1 9m40s
gateway-848874fc9d-pwp9w 1/1 Running 0 48m
order-6694f4c474-cp5d6 1/1 Running 4 3h29m
order-6694f4c474-kh8hm 1/1 Running 0 96s
order-6694f4c474-rg7kw 1/1 Running 0 96s
portal-6fdcd765bd-mtq68 1/1 Running 1 24m
product-784d9f77b5-rxsf2 1/1 Running 0 64m
stock-6cf98bb445-l8gxb 1/1 Running 0 133m
缩容也很简单,将order的副本数缩减成1个
[root@binary-k8s-master1 k8s]\# kubectl scale deploy order --replicas=1 -n simple-ms
deployment.apps/order scaled
成功缩减到1个副本
[root@binary-k8s-master1 ~]\# kubectl get pod -n simple-ms
NAME READY STATUS RESTARTS AGE
eureka-0 1/1 Running 1 9m24s
eureka-1 1/1 Running 0 11m
eureka-2 1/1 Running 1 12m
gateway-848874fc9d-pwp9w 1/1 Running 0 51m
order-6694f4c474-rg7kw 1/1 Running 0 4m45s
portal-6fdcd765bd-mtq68 1/1 Running 1 27m
product-784d9f77b5-rxsf2 1/1 Running 0 67m
stock-6cf98bb445-l8gxb 1/1 Running 0 136m
将order商品服务的副本数调整为3个,原本是1个,现在扩容到3个
[root@binary-k8s-master1 k8s]\# kubectl scale deploy order --replicas=3 -n simple-ms
deployment.apps/order scaled
可以看到order订单服务目前已经有3个pod副本了
[root@binary-k8s-master1 ~]\# kubectl get pod -n simple-ms
NAME READY STATUS RESTARTS AGE
eureka-0 1/1 Running 1 6m15s
eureka-1 1/1 Running 0 7m59s
eureka-2 1/1 Running 1 9m40s
gateway-848874fc9d-pwp9w 1/1 Running 0 48m
order-6694f4c474-cp5d6 1/1 Running 4 3h29m
order-6694f4c474-kh8hm 1/1 Running 0 96s
order-6694f4c474-rg7kw 1/1 Running 0 96s
portal-6fdcd765bd-mtq68 1/1 Running 1 24m
product-784d9f77b5-rxsf2 1/1 Running 0 64m
stock-6cf98bb445-l8gxb 1/1 Running 0 133m
缩容也很简单,将order的副本数缩减成1个
[root@binary-k8s-master1 k8s]\# kubectl scale deploy order --replicas=1 -n simple-ms
deployment.apps/order scaled
成功缩减到1个副本
[root@binary-k8s-master1 ~]\# kubectl get pod -n simple-ms
NAME READY STATUS RESTARTS AGE
eureka-0 1/1 Running 1 9m24s
eureka-1 1/1 Running 0 11m
eureka-2 1/1 Running 1 12m
gateway-848874fc9d-pwp9w 1/1 Running 0 51m
order-6694f4c474-rg7kw 1/1 Running 0 4m45s
portal-6fdcd765bd-mtq68 1/1 Running 1 27m
product-784d9f77b5-rxsf2 1/1 Running 0 67m
stock-6cf98bb445-l8gxb 1/1 Running 0 136m
5.2.微服务的升级
微服务在物理机环境更新时只需要拉取最新的代码,编译成jar包,在服务器上部署就可以了,但是微服务迁移到K8S集群之后,更新过程就变得比较复杂了,首先更新的代码,然后编译成jar包,再通过Dockerfile将jar包做成docker镜像,然后在k8s集群中部署,当然这些流程可以通过Jenkins实现自动化上线。
5.2.1.微调Portal前端首页的代码
将平台的这些文件修改成jiangxl.com的内容完成一次portal前端首页的升级
首先找到记录这些文字的文件
# vim templates/index.ftl
# vim templates/orderList.ftl
# vim templates/productList.ftl
进入文件后在末行模式输入:
%s/容器学院 www.ctnrs.com/jiangxl.com/g
%s/www.ctnrs.com/jiangxl.com/g
然后将所有的容器学院的文字删除
# vim templates/index.ftl
# vim templates/orderList.ftl
# vim templates/productList.ftl
进入文件后在末行模式输入:
%s/容器学院 www.ctnrs.com/jiangxl.com/g
%s/www.ctnrs.com/jiangxl.com/g
然后将所有的容器学院的文字删除
5.2.2.编译最新代码
[root@binary-k8s-master1 simple-microservice-dev1]\# mvn clean package -D maven.test.skip=true -P prod
[root@binary-k8s-master1 simple-microservice-dev1]\# mvn clean package -D maven.test.skip=true -P prod
5.2.3.通过脚本将Protal最新代码打包成镜像
直接使用前面部署程序的脚本即可
[root@binary-k8s-master1 k8s]\# sh deploy_k8s.sh portal-service
[root@binary-k8s-master1 k8s]\# sh deploy_k8s.sh portal-service
5.2.4.升级K8S集群中的Protal微服务
[root@binary-k8s-master1 k8s]\# kubectl -n simple-ms set image deployment portal portal=192.168.20.11/simple-microservice/portal-service:v1 --record
[root@binary-k8s-master1 k8s]\# kubectl -n simple-ms set image deployment portal portal=192.168.20.11/simple-microservice/portal-service:v1 --record
程序更新完成后刷新页面就可以看到我们更新的页面内容