0
mazhilin · 2020年01月06日

Linux Centos7 环境搭建Docker部署Zookeeper分布式集群服务实战

OpenJDKJavaPythonLinuxGolang

Zookeeper完全分布式集群服务

准备好3台服务器:

  • [x]A-> centos-helios:192.168.19.1
  • [x]B-> centos-hestia:192.168.19.2
  • [x]C-> centos-hebe:192.168.19.3
⚠️ [注意事项]：
1.一定要确定3台服务器之间可以相互通信
2.如果是采用的阿里云等服务器的话,一定要配置主机名与IP映射，否则会连接拒绝等问题

配置Zookeeper安装目录

  • 在3台宿主机配置zookeeper安装目录：/docker/zookeeper
  • 并且在文件夹创建 data 和logs 目录：mkdir -p /docker/zookeeper/data && mkdir -p /docker/zookeeper/logs
mkdir -p /docker/zookeeper/data
mkdir -p /docker/zookeeper/logs
  • 授权文件目录：chmod -R 777 /docker/zookeeper/

安装和部署Zookeeper

拉取zookeeper镜像：docker pull zookeeper:3.5.6

[root@centos-helios ~]# docker pull zookeeper:3.5.6
3.5.6: Pulling from library/zookeeper
8ec398bc0356: Pull complete 
8e866b095956: Pull complete 
e18af1784b1f: Pull complete 
9cec9f51936c: Pull complete 
475dcbe0e2e9: Pull complete 
222ebb7346b3: Pull complete 
f0552f5b3f6c: Pull complete 
cdfcdbc43b13: Pull complete 
Digest: sha256:6b6b5f7fb6a47d2b311df5af1718af5a425a679dbb844d77913fa68d1a8bf0fd
Status: Downloaded newer image for zookeeper:3.5.6
docker.io/library/zookeeper:3.5.6
[root@centos-helios ~]#
⚠️ [注意事项]：
1.由于Zookeeper是一个常规的基础服务组件，我这直接拉取Docker hub镜像，有兴趣的可以自行构建镜像部署
2.这里只对centos-helios有举例，另外2台需要执行相同操作，以下不再赘述

临时安装Zookeeper服务：zookeeper-tmp-server

docker run -itd -p 2181:2181 --restart always --name=zookeeper-tmp-server  --privileged=true \
-v /docker/develop/zookeeper/data:/data \
-v /docker/develop/zookeeper/datalog:/datalog \
-v /docker/develop/zookeeper/logs:/logs \
-e "ZOO_STANDALONE_ENABLED=true" \
-e "TZ=Asia/Shanghai" \
-e "ZOO_4LW_COMMANDS_WHITELIST=*" \
-e "ZOO_AUTOPURGE_SNAPRETAINCOUNT=5" \
-e "ZOO_AUTOPURGE_PURGEINTERVAL=24" \
-e "ZOO_MAX_CLIENT_CNXNS=64" \
-e "ZOO_TICK_TIME=2000" \
-e "ZOO_INIT_LIMIT=100" \
-e "ZOO_SYNC_LIMIT=5" \
zookeeper:3.5.6

复制zookeeper-tmp-server的conf到/docker/zookeeper：
1.查看容器列表：docker ps --format "table {{.ID}}\t{{.Names}}\t{{.Ports}}"

[root@centos-helios zookeeper]# docker ps --format "table {{.ID}}\t{{.Names}}\t{{.Ports}}"
CONTAINER ID        NAMES                  PORTS
71ad055739b6        zookeeper-tmp-server   2888/tcp, 3888/tcp, 0.0.0.0:2181->2181/tcp, 8080/tcp
[root@centos-helios zookeeper]#

2.执行Docker cp 命令：docker cp Zookeeper容器ID:/conf /docker/zookeeper/

[root@centos-helios zookeeper]# docker cp 71ad055739b6:/conf /docker/zookeeper/
[root@centos-helios zookeeper]#

centos-helios：docker cp 71ad055739b6:/conf /docker/zookeeper/
centos-hebe：docker cp bf9df9966efb:/conf /docker/zookeeper/
centos-hestia：docker cp 4b4f8e813c80:/conf /docker/zookeeper/

编写zookeeper的zoo.cfg配置文件：
1.服务器集群配置：

  • [x]centos-helios:192.168.19.1:2888:3888
  • [x]centos-hestia:192.168.19.2:2888:3888
  • [x]centos-hebe:192.168.19.3:2888:3888

2.对应的配置文件配置如下：

  • [x]centos-helios:/docker/zookeeper/conf/zoo.cfg
tickTime=2000
initLimit=100
syncLimit=5
dataDir=/data
dataLogDir=/datalog
clientPort=2181
quorumListenOnAllIPs=true
preAllocSize=64M
globalOutstandingLimit=100000
snapCount=50000
maxClientCnxns=64
minSessionTimeout=4000
maxSessionTimeout=40000
4lw.commands.whitelist=*
admin.enableServer=true
server.1=192.168.19.1:2888:3888
server.2=192.168.19.2:2888:3888
server.3=192.168.19.3:2888:3888
tcpKeepAlive=true
ipReachableTimeout=0
cnxTimeout=3
electionAlg=3
autopurge.snapRetainCount=5
autopurge.purgeInterval=24
#忽略ACL验证
skipACL=yes
forceSync=yes
  • [x]centos-hestia:/docker/zookeeper/conf/zoo.cfg
tickTime=2000
initLimit=100
syncLimit=5
dataDir=/data
dataLogDir=/datalog
clientPort=2181
quorumListenOnAllIPs=true
preAllocSize=64M
globalOutstandingLimit=100000
snapCount=50000
maxClientCnxns=64
minSessionTimeout=4000
maxSessionTimeout=40000
4lw.commands.whitelist=*
admin.enableServer=true
server.1=192.168.19.1:2888:3888
server.2=192.168.19.2:2888:3888
server.3=192.168.19.3:2888:3888
tcpKeepAlive=true
ipReachableTimeout=0
cnxTimeout=3
electionAlg=3
autopurge.snapRetainCount=5
autopurge.purgeInterval=24
#忽略ACL验证
skipACL=yes
forceSync=yes
  • [x]centos-hebe:/docker/zookeeper/conf/zoo.cfg
tickTime=2000
initLimit=100
syncLimit=5
dataDir=/data
dataLogDir=/datalog
clientPort=2181
quorumListenOnAllIPs=true
preAllocSize=64M
globalOutstandingLimit=100000
snapCount=50000
maxClientCnxns=64
minSessionTimeout=4000
maxSessionTimeout=40000
4lw.commands.whitelist=*
admin.enableServer=true
server.1=192.168.19.1:2888:3888
server.2=192.168.19.2:2888:3888
server.3=192.168.19.3:2888:3888
tcpKeepAlive=true
ipReachableTimeout=0
cnxTimeout=3
electionAlg=3
autopurge.snapRetainCount=5
autopurge.purgeInterval=24
#忽略ACL验证
skipACL=yes
forceSync=yes
⚠️ [注意事项]：如果是采用阿里云等云服务器部署,一般设置为：
server.1=阿里云[外网IP OR 内网IP]:2888:3888
server.2=阿里云[外网IP OR 内网IP]:2888:3888
server.3=阿里云[外网IP OR 内网IP]:2888:3888

编写正式安装Dcoekr命令：

  • [x]centos-helios：
docker run -itd -p 2181:2181 -p 2888:2888 -p 3888:3888 --restart always --name=zookeeper-server --hostname zookeeper-server --privileged=true \
-v /docker/zookeeper/conf/zoo.cfg:/conf/zoo.cfg \
-v /docker/zookeeper/data:/data \
-v /docker/zookeeper/datalog:/datalog \
-v /docker/zookeeper/logs:/logs \
-e "TZ=Asia/Shanghai" \
-e "ZOO_MY_ID=1" \
-e "JAVA_OPTS=-server -Xms512m -Xmx512m -Xmn256m -Duser.home=/opt -XX:MetaspaceSize=128m -XX:MaxMetaspaceSize=320m -XX:+AlwaysPreTouch -XX:-UseBiasedLocking" \
zookeeper:3.5.6
  • [x]centos-hestia:
docker run -itd -p 2181:2181 -p 2888:2888 -p 3888:3888 --restart always --name=zookeeper-server --hostname zookeeper-server --privileged=true \
-v /docker/zookeeper/conf/zoo.cfg:/conf/zoo.cfg \
-v /docker/zookeeper/data:/data \
-v /docker/zookeeper/datalog:/datalog \
-v /docker/zookeeper/logs:/logs \
-e "TZ=Asia/Shanghai" \
-e "ZOO_MY_ID=2" \
-e "JAVA_OPTS=-server -Xms512m -Xmx512m -Xmn256m -Duser.home=/opt -XX:MetaspaceSize=128m -XX:MaxMetaspaceSize=320m -XX:+AlwaysPreTouch -XX:-UseBiasedLocking" \
zookeeper:3.5.6
  • [x]centos-hebe：
docker run -itd -p 2181:2181 -p 2888:2888 -p 3888:3888 --restart always --name=zookeeper-server --hostname zookeeper-server --privileged=true \
-v /docker/zookeeper/conf/zoo.cfg:/conf/zoo.cfg \
-v /docker/zookeeper/data:/data \
-v /docker/zookeeper/datalog:/datalog \
-v /docker/zookeeper/logs:/logs \
-e "TZ=Asia/Shanghai" \
-e "ZOO_MY_ID=3" \
-e "JAVA_OPTS=-server -Xms512m -Xmx512m -Xmn256m -Duser.home=/opt -XX:MetaspaceSize=128m -XX:MaxMetaspaceSize=320m -XX:+AlwaysPreTouch -XX:-UseBiasedLocking" \
zookeeper:3.5.6

后续服务配置

1.开放端口[2181 2888 3888]

  • [1].启动防火墙:systemctl start firewalld.service
  • [2].查看防火墙状态：systemctl status firewalld.service
  • [3].开放端口:firewall-cmd --zone=public --add-port=端口号/tcp --permanent
firewall-cmd --zone=public --add-port=2181/tcp --permanent
firewall-cmd --zone=public --add-port=2888/tcp --permanent
firewall-cmd --zone=public --add-port=3888/tcp --permanent
  • [4].加载初始化防火墙生效：firewall-cmd --reload
  • [5].停用防火墙：systemctl stop firewalld.service
⚠️ [注意事项]：
1.如果是Centos7以及以上系统，防火墙默认是firewall,对于Centos7版本请自行进行配置
2.依据实测，对于2888和3888等端口必须开放，否则在使用Zookeeper服务的时候会报连接[KeeperErrorCode = ConnectionLoss]错误

2.配置启动以及测试zookeeper服务:

  • [x]启动容器Zookeeper服务顺序:A->B->C-重启A
  • [x]进入容器输入：bin/zkServer.sh status

centos-helios:

root@zookeeper-server:/apache-zookeeper-3.5.6-bin# bin/zkServer.sh status
ZooKeeper JMX enabled by default
Using config: /conf/zoo.cfg
Client port found: 2181. Client address: localhost.
Mode: follower
root@zookeeper-server:/apache-zookeeper-3.5.6-bin#

centos-hestia:

root@zookeeper-server:/apache-zookeeper-3.5.6-bin# bin/zkServer.sh status
ZooKeeper JMX enabled by default
Using config: /conf/zoo.cfg
Client port found: 2181. Client address: localhost.
Mode: leader
root@zookeeper-server:/apache-zookeeper-3.5.6-bin#

centos-hebe:

root@zookeeper-server:/apache-zookeeper-3.5.6-bin# bin/zkServer.sh status
ZooKeeper JMX enabled by default
Using config: /conf/zoo.cfg
Client port found: 2181. Client address: localhost.
Mode: follower
root@zookeeper-server:/apache-zookeeper-3.5.6-bin#
⚠️ [注意事项]：
1.Zookeeper的集群服务采用的是Leader-Follower线程模型，如上所述。
2.Zookeeper使用了ZAB(Zookeeper Atomic Broadcast)协议，保证了leader,follower的一致性，leader 负责数据的读写，而follower只负责数据的读，如果follower遇到写操作，会提交到leader。
版权声明：本文为博主原创文章，遵循相关版权协议，如若转载或者分享请附上原文出处链接和链接来源。
0 阅读 1.8k
推荐阅读
《工具》centos 7.x 下安装 JDK，超详细总结一看就会 环境变量_JAVA_LAUNCHER_DEBUG，它能给你更多的JVM信息 Java的EOF标识？ ARM64架构下，OpenJDK的官方Docker镜像为何没有8版本 在docker上编译openjdk8 微软宣布加入 OpenJDK，打不过就改变 Java 未来！
...
mazhilin
关注数
0
文章数
9
关注作者 作者主页
目录
极术微信服务号
关注极术微信号
实时接收点赞提醒和评论通知
安谋科技学堂公众号
关注安谋科技学堂
实时获取安谋科技及 Arm 教学资源
安谋科技招聘公众号
关注安谋科技招聘
实时获取安谋科技中国职位信息