docker-swarm容器固定到node节点启动

Posted 2020-12-25

               Docker-compose的编排

首先这里需要创建一个目录mkdir /docker-compose然后在里边创建一个docker-compose.yml的文件。

具体包括：

Images、hostname、ports、networks、volumes、deploy这些内容

 

其中包括nginx+tomcat1+tomcat2三个服务

 

version: "3"               这是最新的版本

services:

  nginx:

    image: nginx5:latest

    hostname: nginx

    ports:

      - "8021:80"

    networks:

      - "mynet"

    volumes:

      - "/srv/jetsen/nginx:/nginx"

      - "nginx://usr/local/nginx-1.6.2"

    deploy:

      placement:

        constraints:

          - engin.labels.function == nodeone

 

  tomcat1:

    image: tomcat15:latest

    hostname: tomcat1

    ports:

      - "8088:8080"

    networks:

      - "mynet"

   volumes:

      - "/srv/jetsen/tomcat1:/tomcat1"

      - "tomcat1jdk:/usr/local/jdk1.7.0_79"

      - "tomcat1tomcat://usr/local/apache-tomcat-7.0.61/"

    deploy:

      placement:

        constraints:

          - engine.labels.function == nodeone

 

  tomcat2:

    image: tomcat15:latest

    hostname: tomcat2

    ports:

      - "8089:8080"

    networks:

      - "mynet"

    volumes:

      - "/srv/jetsen/tomcat2:/tomcat2"

      - "tomcat2jdk:/usr/local/jdk1.7.0_79"

      - "tomcat2tomcat://usr/local/apache-tomcat-7.0.61/"

    deploy:

      placement:

        constraints:

          - engine.labels.function == nodetwo

volumes:

  nginx:

  tomcat1tomcat:

  tomcat1jdk:

  tomcat2tomcat:

  tomcat2jdk:

networks:

  mynet:

external: true

 

另外需要修改docker的主配置文件

 

--label=function=nodeone： 一个标签【别名】

--label=host=node1   ：  主机名

--registry-mirror=http://e0d76b42.m.daocloud.io：国内的镜像

 

注：另外node2的服务器节点也需要修改

 

完成之后通过docker stack deploy -c docker-compose.yml  【mynet：名字】

启动这个docker-compose编排文件

 

查看docker service  list  

 

查看运行的节点

 

 

                          Nginx+tomcat1+tomcat2反向代理docker

 

实现上面的三个容器运行必须要有三个dockerfile写出来的启动镜像

 

Nginx：

 

这里使用的是nginx1.6.2版本

ENTRYPOINT /usr/local/nginx/sbin/nginx && tail -F /usr/local/nginx/logs/access.log

FROM docker.io/centos:7.2.1511

MAINTAINER  LXF

RUN  yum -y install gcc gcc-c++  kernel-headers    dracut  make auotmake autoconf libtool pcre pcre-devel zlib zlib-devel openssl openssl-devel

ADD ./nginx-1.6.2  /usr/local/nginx-1.6.2

RUN  cd /usr/local/nginx-1.6.2 && ./configure --prefix=/usr/local/nginx --with-http_stub_status_module && make &&make install

EXPOSE 80

ENTRYPOINT /usr/local/nginx/sbin/nginx && tail -F /usr/local/nginx/logs/access.log

 

 

 

Tomcat1的镜像：

 

Tomcat使用的是7版本

FROM docker.io/centos:7.2.1511

MAINTAINER LXF

RUN yum -y install gcc gcc- c++ make   glibc  kernel-headers tar  glibc-common

ADD ./apache-tomcat-7.0.61 /usr/local/apache-tomcat-7.0.61

ADD ./jdk1.7.0_79  /usr/local/jdk1.7.0_79

ENV  JAVA_HOME=/usr/local/jdk1.7.0_79

ENV  PATH=$JAVA_HOME/bin:$PATH

FROM docker.io/centos:7.2.1511

MAINTAINER LXF

RUN yum -y install gcc gcc- c++ make   glibc  kernel-headers tar  glibc-common

ADD ./apache-tomcat-7.0.61 /usr/local/apache-tomcat-7.0.61

ADD ./jdk1.7.0_79  /usr/local/jdk1.7.0_79

ENV  JAVA_HOME=/usr/local/jdk1.7.0_79

ENV  PATH=$JAVA_HOME/bin:$PATH

ENV  CATALINA_HOME=/usr/local/apache-tomcat-7.0.61

ENV  PATH=$JAVA_HOME/bin:$CATALINA_HOME/bin:$PATH

ADD ./server.xml   /usr/local/apache-tomcat-7.0.61/conf/server.xml

RUN mkdir -p /web/webapp1

ADD index.jsp /web/webapp1/index.jsp

RUN chmod +x   /usr/local/apache-tomcat-7.0.61/bin/*

EXPOSE 8080

ENTRYPOINT  /usr/local/apache-tomcat-7.0.61/bin/startup.sh  && tail -F /usr/local/apache-tomcat-7.0.61/logs/catalina.out

~                          

 

Tomcat2的镜像和tomcat1的没有区别就是在index.Jsp文件有区别

 

完成之后开始生产成镜像   通过docker  build  -t  镜像   /路径

 

 

aa是nginx的镜像

 

完成之后需要去编写docker-compose.yml文件

内容如下

version: "3"

services:

  tomcat1:

    image: tomcat:1

    hostname: tomcat1

    expose:

      - 8080

    networks:

      - "mynet"

    volumes:

      - "/srv/jetsen/tomcat1:/tomcat1"

      - "tomcat1jdk:/usr/local/jdk1.7.0_79"

      - "tomcat1tomcat://usr/local/apache-tomcat-7.0.61/"

    deploy:

      placement:

        constraints:

          - engine.labels.function == nodeone

 

  tomcat2:

    image: tomcat:2

    hostname: tomcat2

    expose:

      - 8080

    networks:

      - "mynet"

    volumes:

      - "/srv/jetsen/tomcat2:/tomcat2"

      - "tomcat2jdk:/usr/local/jdk1.7.0_79"

      - "tomcat2tomcat://usr/local/apache-tomcat-7.0.61/"

    deploy:

      placement:

        constraints:

          - engine.labels.function == nodetwo

  nginx:

    image: aa:aa

    hostname: nginx

    ports:

      - "8021:80"

    networks:

      - "mynet"

    volumes:

      - "/srv/jetsen/nginx:/nginx"

      - "nginx1:/usr/local/nginx-1.6.2"

      - "nginx2://usr/local/nginx"

    deploy:

      placement:

        constraints:

          - engine.labels.function == nodeone

 

volumes:

  nginx:

  nginx1:

  nginx2:

  tomcat1tomcat:

  tomcat1jdk:

  tomcat2tomcat:

  tomcat2jdk:

networks:

  mynet:

    external: true

 

 

注：卷组一定要在最后标注，另外加红颜色的路径要创建才行       加绿色的表示申明的作用使用的卷组和网络

完成之后需要启动这个docker-compose文件

       

自己定义名字

当然删除的命令是

看一下起来的容器

 

前面说过tomcat2在woeker节点上工作

                  

启动三个容器之后进行反向代理的配置

需要在nginx的主配置文件配置upstream，可以去卷组里面配置

 

                             手动添加

 

 

添加在卷组当中会自动和容器的数据同步，防止容器挂掉数据丢失

 

完成之后需要测试一下

 

反向代理测试：

 

 

 

 

 

 

 

注：如果有业务需要需要将docker-compose文件单独分开那么需要编写单独的文件，比如讲上面的分为三分，每一份都需要申明网络和卷组。当然使用同一个网络mynet的话，启动的容器之间还是可以互相同信得