使用docker swarm 搭建CentOS7 集群

canyanol

已于 2023-01-31 16:06:59 修改

阅读量367

点赞数

分类专栏：技能文章标签： docker 运维容器

于 2023-01-13 15:58:02 首次发布

本文链接：https://blog.csdn.net/canyanol/article/details/128670214

版权

技能专栏收录该内容

6 篇文章 0 订阅

订阅专栏

使用Virtual Box桥接网络，部分需要登录才能访问外网的需要配置双网卡

桥接网卡配置：

vi /etc/sysconfig/network-scripts/ifcfg-enp0s8

BOOTPROTO=static # 自动获取ip
ONBOOT="yes"

IPADDR=192.168.2.111
GATEWAY=192.168.2.1
NETMASK=255.255.255.0

网络常用命令

service network restart --重启网卡
ip addr --查看ip
hostnamectl set-hostname node1 --设置机器名称
systemctl stop firewalld //临时关闭关闭防火墙
systemctl disable firewalld //禁止开机启动关闭防火墙

安装docker

yum -y update --更新
yum install -y yum-utils device-mapper-persistent-data lvm2 --安装需要的软件包

设置源

//官方镜像

yum-config-manager --add-repo https://download.docker.com/linux/centos/docker-ce.repo

//阿里镜像

yum-config-manager --add-repo http://mirrors.aliyun.com/docker-ce/linux/centos/docker-ce.repo

yum list docker-ce --showduplicates | sort -r --查看版本

yum install -y docker-ce --这里默认安装的是最新稳定版,你也可以指定版本号安装

systemctl start docker --启动

systemctl enable docker --加入自启

docker login --username=canyanol@qq.com registry.cn-hangzhou.aliyuncs.com 登录远程仓库

docker swarm 集群

node1:管理节点
node2：计算节点
node3:计算节点

node1 执行

docker swarm init --advertise-addr 192.168.2.151

node2 ,node3加入

docker swarm join --token XXXXXX

docker service

-- docker service update命令  https://www.yiibai.com/docker/service_update.html

docker service create --replicas 2 --network jdt --name queue -p 80:80 registry.cn-hangzhou.aliyuncs.com/my-work/work:jdt-queue-v3-115 --创建一个服务
docker service ls：查看所有的service基本信息
docker service ps [SERVICE_ID]：查看具体某个service的信息
docker service rm queue --删除服务

docker service update --replicas=3 queue  更新服务
docker service update --image registry.cn-hangzhou.aliyuncs.com/my-work/work:jdt-queue-v3-114 queue  更新镜像版本
 docker service scale queue=3  缩放单个服务

docker service ps queue 查看服务

docker node ls 查看计算节点
docker node update --availability drain node2 更新节点   上线/下线

docker service update \
--image nginx:2.0 \
--update-parallelism 2 \
--update-delay 20s my-nginx

# 滚动更新 将镜像更新为2.0 每次更新两个副本 每20s更新一次

使用图形化工具portainer 管理docker

docker run -d -p 9000:9000 --name=portainer   -v "/var/run/docker.sock:/var/run/docker.sock"   
-v /host/data:/data   portainer/portainer

timescaledb:latest Stack

version: '3.6'
services:
  timescaledb14:
    image: timescale/timescaledb:latest-pg14
    environment:
      - POSTGRES_PASSWORD=123456
    volumes:
      - /data/postgresql:/var/lib/postgresql/data
    deploy:
      replicas: 1
      restart_policy:
        condition: any
      resources:
        limits:
          cpus: "0.2"
          memory: 512M
      update_config:
        parallelism: 1 # 每次更新1个副本
        delay: 5s # 每次更新间隔 
        monitor: 10s # 单次更新多长时间后没有结束则判定更新失败
        max_failure_ratio: 0.1 # 更新时能容忍的最大失败率
        order: start-first # 更新顺序为新任务启动优先
      placement:
          constraints: [node.role == manager] #运行在 管理角色机器上 该hostname为指定容器在哪个主机启动 - node.hostname == node1
    ports:
      - 5432:5432
    networks:
      - mynet

networks:
  mynet:
    external: true

queue 的部署指令


version: '3.6'
services:
  webserver: # webserver 服务
    image: nginx:1.23.3
    ports:
      - 80:80
      - 443:443
    networks:
      - mynet
    volumes:
       - /data/nginx/conf.d:/etc/nginx/conf.d
       - /data/nginx/nginx.conf:/etc/nginx/nginx.conf
    deploy:
      replicas: 1
      placement:
          constraints: [node.role == manager] #运行在 管理角色机器上 该hostname为指定容器在哪个主机启动 - node.hostname == node1  
      restart_policy:
        condition: any
      update_config:
        parallelism: 1 # 每次更新1个副本
        delay: 5s # 每次更新间隔 
        monitor: 10s # 单次更新多长时间后没有结束则判定更新失败
        max_failure_ratio: 0.1 # 更新时能容忍的最大失败率
        order: start-first # 更新顺序为新任务启动优先
      
  webapi: # webapi 服务
    image: registry.cn-hangzhou.aliyuncs.com/canyanol/queue:3.120
    ports:
      - 8111:80
    networks:
      - mynet
#    depends_on:
#      - redis6
#      - mysql5 
    deploy:
      replicas: 2
      restart_policy:
        condition: any
      resources:
        limits:
          cpus: "0.5"
          memory: 512M
      update_config:
        parallelism: 1 # 每次更新1个副本
        delay: 3s # 每次更新间隔 
        monitor: 20s # 单次更新多长时间后没有结束则判定更新失败
        max_failure_ratio: 0.1 # 更新时能容忍的最大失败率
        order: start-first # 更新顺序为新任务启动优先
      placement:
          constraints: [node.role == worker] #运行在 管理角色机器上 该hostname为指定容器在哪个主机启动 - node.hostname == node1
  
  redis6:
    image: redis:6.0.6
    restart: always
    ports:
      - 6379:6379
    networks:
      - mynet
    privileged: true
    command: redis-server /etc/redis/redis.conf --appendonly yes
    volumes:
      - /data/redis/data:/data
      - /data/redis/conf/redis.conf:/etc/redis/redis.conf
    deploy:
      placement:
        constraints: [node.role == manager]

  mysql5:
    restart: always
    image: mysql:5.7.22
    container_name: mysql
    ports:
      - 3306:3306
    environment:
      TZ: Asia/Shanghai
      MYSQL_ROOT_PASSWORD: 123456
    deploy:
      replicas: 1
      restart_policy:
        condition: any
      placement:
          constraints: [node.role == manager]
    command:
      --character-set-server=utf8mb4
      --collation-server=utf8mb4_general_ci
      --explicit_defaults_for_timestamp=true
      --lower_case_table_names=1
      --max_allowed_packet=128M
      --sql-mode="STRICT_TRANS_TABLES,NO_AUTO_CREATE_USER,NO_ENGINE_SUBSTITUTION,NO_ZERO_DATE,NO_ZERO_IN_DATE,ERROR_FOR_DIVISION_BY_ZERO"
    volumes:
      - /data/mysql/conf:/etc/mysql
      - /data/mysql/logs:/var/log/mysql
      - /data/mysql/data:/var/lib/mysql
    networks:
      - mynet

networks:
  mynet:
    external: true

queue_service.yaml

traefik+ jaeger+mysql5+redis+helloworld+webapi

traefik 实现反向代理，负载均衡，限流，熔断，健康检查。jaegertracing实现分布式追踪


version: '3.6'
services:
  traefik: # 反向代理服务
    image: traefik:2.9.6
    ports:
      - "80:80"
      - "8080:8080" # traefik dashboard
    command:
      - --api.insecure=true # set to 'false' on production
      - --api.dashboard=true # see https://docs.traefik.io/v2.0/operations/dashboard/#secure-mode for how to secure the dashboard
      - --api.debug=true # enable additional endpoints for debugging and profiling
      - --log.level=DEBUG # debug while we get it working, for more levels/info see https://docs.traefik.io/observability/logs/
      - --providers.docker=true
      - --providers.docker.swarmMode=true
      - --providers.docker.exposedbydefault=false
      - --providers.docker.network=proxy
      - --entryPoints.web.address=:80
      # 日志追踪 jaeger
      - --tracing.jaeger.collector.endpoint=http://jaegertracing:14268/api/traces?format=jaeger.thrift
        #- --entryPoints.elderServer.address=:7122
    volumes:
      - /var/run/docker.sock:/var/run/docker.sock
    networks:
      - mynet
    deploy:
      labels:
        - "traefik.enable=true"
        - "traefik.http.routers.api.rule=Host(`traefik.xxx.com`)"
        - "traefik.http.routers.api.service=api@internal" # Let the dashboard access the traefik api
      placement:
          constraints: [node.role == manager] #运行在 管理角色机器上 该hostname为指定容器在哪个主机启动 - node.hostname == node1

  webapi: # webapi 服务
    image: registry.cn-hangzhou.aliyuncs.com/canyanol/queue:3.120
    networks:
      - mynet
    deploy:
      labels:
        # 对外暴露容器服务
        - "traefik.enable=true"
        # 对外访问的路由地址，路由规则请参考官网
        # https://docs.traefik.io/routing/routers/ Host(`www.a.com`)
        - "traefik.http.routers.webapi.rule=Host(`b.a.com`)"
        
        # 对外暴露的入口点
        - "traefik.http.routers.webapi.entrypoints=web"
        
        # 容器内的入口点，treafik无法获知你的服务的访问入口点
        # 所以你必须以此告诉Traefik
        # Traefik同时会在此对横向拓展的容器建立负载均衡
        # 更多见https://docs.traefik.io/routing/services/
        # 负载均衡 https://doc.traefik.io/traefik/routing/services/#load-balancing
        
        - "traefik.http.services.webapi.loadbalancer.server.port=80"
        # 启用粘性会话后，将在初始响应上设置标头，让客户端知道哪个服务器处理第一个响应。 在后续请求中，为了保持会话与同一服务器处于活动状态，客户端应发送具有设置值的 cookie
        #- "traefik.http.services.webapi.loadbalancer.sticky.cookie={}"
        
        #- "traefik.http.services.webapi.loadbalancer.sticky.cookie.name=loadbalancer"
        #- "traefik.http.services.webapi.loadbalancer.sticky.cookie.secure=true"
        #- "traefik.http.services.webapi.loadbalancer.sticky.cookie.httpOnly=true"
        
        # 健康检查 https://doc.traefik.io/traefik/routing/services/#health-check
        - "traefik.http.services.webapi.loadbalancer.healthCheck.path=/api/services/app/Session/GetCurrentLoginInformations"
        - "traefik.http.services.webapi.loadbalancer.healthCheck.interval=5s"
        - "traefik.http.services.webapi.loadbalancer.healthCheck.timeout=3s"
        
        # 限流 https://doc.traefik.io/traefik/middlewares/http/ratelimit/
        #average 是给定源允许的最大速率，默认情况下以每秒请求数为单位
        - "traefik.http.middlewares.webapi-ratelimit.ratelimit.average=100"
        # burst 是在同一任意短时间内允许通过的最大请求数。
        - "traefik.http.middlewares.webapi-ratelimit.ratelimit.burst=50"

        # 熔断 断路器可防止系统将请求堆叠到不正常的服务，从而导致级联故障  https://doc.traefik.io/traefik/middlewares/http/circuitbreaker/
        # 当 30% 的请求返回 5XX 状态代码或网络错误比率达到 10% 时，触发断路器
        - "traefik.http.middlewares.webapi-latency-check.circuitbreaker.expression=ResponseCodeRatio(500, 600, 0, 600) > 0.30 || NetworkErrorRatio() > 0.10"
        
        # 添加到中间件
        - "traefik.http.routers.webapi.middlewares=webapi-ratelimit,webapi-latency-check"
      replicas: 2 # 2个实例
      restart_policy: # 重启策略
        condition: any
      resources: # 资源配置
        limits:
          cpus: "0.5"
          memory: 512M
      update_config: # 更新配置
        parallelism: 1 # 每次更新1个副本
        delay: 3s # 每次更新间隔 
        monitor: 20s # 单次更新多长时间后没有结束则判定更新失败
        max_failure_ratio: 0.1 # 更新时能容忍的最大失败率
        order: start-first # 更新顺序为新任务启动优先
      placement:
          constraints: [node.role == worker] #运行在 管理角色机器上 该hostname为指定容器在哪个主机启动 - node.hostname == node1
  
  helloworld:
    image: tutum/hello-world:latest
    networks:
     # 与traefik同一个网络
     - mynet
    deploy:
      labels:
        # 对外暴露容器服务
        - "traefik.enable=true"
        # 对外访问的路由地址，路由规则请参考官网
        # https://docs.traefik.io/routing/routers/
        - "traefik.http.routers.helloworld.rule=Host(`a.a.com`)"
        # 对外暴露的入口点
        - "traefik.http.routers.helloworld.entrypoints=web"
        # 容器内的入口点，treafik无法获知你的服务的访问入口点
        # 所以你必须以此告诉Traefik
        # Traefik同时会在此对横向拓展的容器建立负载均衡
        # 更多见https://docs.traefik.io/routing/services/
        - "traefik.http.services.helloworld.loadbalancer.server.port=80"
  
  redis6: # redis 服务
    image: redis:6.0.6
    restart: always
    ports:
      - 6379:6379
    networks:
      - mynet
    command: redis-server /etc/redis/redis.conf --appendonly yes
    volumes:
      - /data/redis/data:/data
      - /data/redis/conf/redis.conf:/etc/redis/redis.conf
    deploy:
      placement:
        constraints: [node.role == manager]

  mysql5: # mysql 数据库服务
    image: mysql:5.7.22
    ports:
      - 3306:3306
    networks:
      - mynet
    environment:
      TZ: Asia/Shanghai
      MYSQL_ROOT_PASSWORD: 123456
    deploy:
      replicas: 1
      restart_policy:
        condition: any
      placement:
          constraints: [node.role == manager]
    volumes:
      - /data/mysql/conf/my.cnf:/etc/mysql/mysql.conf.d/mysqld.cnf
      - /data/mysql/logs:/var/log/mysql
      - /data/mysql/data:/var/lib/mysql


  jaegertracing: # 分布式追踪 https://zhuanlan.zhihu.com/p/524695029
    image: jaegertracing/all-in-one:1.33
    ports:
      - 14268:14268 # 接收日志
      - 14269:14269
      - 16686:16686 # web ui
      - 14250:14250
      - 9411:9411
    networks:
      - mynet
    environment:
      COLLECTOR_ZIPKIN_HOST_PORT: 9411
    deploy:
      replicas: 1
      restart_policy:
        condition: any
      placement:
          constraints: [node.role == manager]
      
networks:
  mynet:
    external: true