原文链接: http://gudaoyufu.com/?p=1395
zookeeper工作方式
ZooKeeper 是一个开源的分布式协调服务,由雅虎创建,是 Google Chubby 的开源实现。 分布式应用程序可以基于 ZooKeeper 实现诸如数据发布/订阅、负载均衡、命名服务、分布式协 调/通知、集群管理、Master 选举、分布式锁和分布式队列 等功能。
在使用中,通常以集群的方式部署,Zookeeper节点部署越多,服务的可靠性越高,建议部署奇数个节点,因为zookeeper集群是以宕机个数过半才会让整个集群宕机的,集群节点数为奇数最佳。
zookeeper也可以以单实例或伪集群的方式运行,只不过这种方式不适用高并发的环境。下面记录一下部署zookeeper的过程,包括单实例和伪集群,分布式集群的部署。
安装JDK
zookeeper是由JAVA开发,运行需要有JAVA环境,安装前先安装JDK。
JDK下载:http://www.oracle.com/technetwork/java/javase/downloads/jdk8-downloads-2133151.html
下载完成后直接yum安装即可
单机实例部署
- 下载
各版本可以在官方网站下载 : https://mirrors.tuna.tsinghua.edu.cn/apache/zookeeper/
这里下载的版本是 3.4.12
- 解压至指定路径
[root@zookeeper ~]# tar zxf zookeeper-3.4.12.tar.gz -C /data/
- 修改解压目录名称,做伪集群或分布式集群最好将目录标明白,容易看清
[root@zookeeper ~]# mv /data/zookeeper-3.4.12/ /data/zookeeper1
- 修改配置文件
[root@zookeeper ~]# cd /data/zookeeper1/conf/[root@zookeeper conf]# mv zoo_sample.cfg zoo.cfg
- 单机实例部署配置如下
[root@zookeeper conf]# vim zoo.cfg tickTime=2000 #2000毫秒=2秒initLimit=10syncLimit=5dataDir=/opt/zookeeper/datadirdataLogDir=/opt/zookeeper/logdirclientPort=2181
- 创建数据目录
[root@zookeeper conf]# mkdir -pv /opt/zookeeper/{datadir,logdata}mkdir: 已创建目录 "/opt/zookeeper"mkdir: 已创建目录 "/opt/zookeeper/datadir"mkdir: 已创建目录 "/opt/zookeeper/logdata"
配置参数说明
ickTime这个时间是作为zookeeper服务器之间或客户端与服务器之间维持心跳的时间间隔,也就是说每个tickTime时间就会发送一个心跳。 initLimit : 配置项是用来配置zookeeper接受客户端(这里所说的客户端不是用户连接zookeeper服务器的客户端,而是zookeeper服务器集群中连接到leader的follower 服务器)初始化连接时最长能忍受多少个心跳时间间隔数。当已经超过10个心跳的时间(也就是tickTime)长度后 zookeeper 服务器还没有收到客户端的返回信息,那么表明这个客户端连接失败。总的时间长度就是 102000=20秒。
- syncLimit :配置项标识leader与follower之间发送消息,请求和应答时间长度,最长不能超过多少个tickTime的时间长度,总的时间长度就是5*2000=10秒。
- dataDir :是zookeeper保存数据的目录,默认情况下如果不定义dataLogDir,zookeeper将写数据的日志文件也保存在这个目录里,最好分开定义
- clientPort : 客户端连接Zookeeper服务器的端口,Zookeeper会监听这个端口接受客户端的访问请求;
server.n=ipA:B:C :定义集群节点号,ip,监听端口,选举通信端口,n是一个数字,表示这个是第几号服务器,A是这个服务器的IP地址,B第一个端口用来集群成员的信息交换,表示这个服务器与集群中的leader服务器交换信息的端口,C是在leader挂掉时专门用来进行选举leader所用的端口。
启动zookeeper
[root@zookeeper ~]# /data/zookeeper1/bin/zkServer.sh startZooKeeper JMX enabled by defaultUsing config: /data/zookeeper1/bin/../conf/zoo.cfgStarting zookeeper ... STARTED
- 查看状态
[root@zookeeper ~]# /data/zookeeper1/bin/zkServer.sh statusZooKeeper JMX enabled by defaultUsing config: /data/zookeeper1/bin/../conf/zoo.cfgMode: standalone #单机模式
- 查看端口 2181已经启动
zookeeper指令
zookeeper启动后,可以先连接测试是否正常
[root@zookeeper ~]# cd /data/zookeeper1/bin/[root@zookeeper bin]# ./zkCli.sh -server 127.0.0.1WATCHER::WatchedEvent state:SyncConnected type:None path:null[zk: 127.0.0.1(CONNECTED) 0] #登录成功
zookeeper启动后会在当前用户的家目录生成状态日志zookeeper.out,平时服务状态出现问题可以从里面找到问题所在
连接上zk后,随意输入一个字符,通常是?号,会给出zookeeper的指令帮助
[zk: 127.0.0.1(CONNECTED) 0] ?ZooKeeper -server host:port cmd args stat path [watch] set path data [version] ls path [watch] #查看节点 delquota [-n|-b] path ls2 path [watch] setAcl path acl setquota -n|-b val path history redo cmdno printwatches on|off delete path [version] sync path listquota path rmr path get path [watch] create [-s] [-e] path data acl #创建节点 addauth scheme auth quit getAcl path close connect host:port #连接指令
- 创建节点
[zk: 127.0.0.1(CONNECTED) 1] create /test "test"Created /test[zk: 127.0.0.1(CONNECTED) 2] ls /[zookeeper, test]
- 获取节点内容
[zk: 127.0.0.1(CONNECTED) 3] get /test testcZxid = 0x8 :该节点是由哪个事务ID产生ctime = Sat Sep 01 21:41:06 CST 2018mZxid = 0x8 :最近更新了该节点的事务IDmtime = Sat Sep 01 21:41:06 CST 2018pZxid = 0x8 :该节点的子节点列表被修改的事务IDcversion = 0 :子节点版本号dataVersion = 0 : 数据版本号aclVersion = 0 : ACL版本号ephemeralOwner = 0x0 dataLength = 4 : 数据长度numChildren = 0 : 子节点个数
乐观并发访问控制和悲观并发访问控制 —延伸
- 更新节点
[zk: 127.0.0.1(CONNECTED) 4] set /test "test path"
- 删除节点
[zk: 127.0.0.1(CONNECTED) 6] delete /test[zk: 127.0.0.1(CONNECTED) 7] ls /testNode does not exist: /test#如果删除有子目录的节点,使用rmr指令
临时节点不能有子节点
ookeeper四字命令
使用telnet可以连接zookeeper发送4个字符的命令。
[root@zookeeper bin]# telnet 127.0.0.1 2181Trying 127.0.0.1...Connected to 127.0.0.1.Escape character is '^]'.ruok #探测是否存活imokConnection closed by foreign host.
[root@zookeeper bin]# telnet 127.0.0.1 2181Trying 127.0.0.1...Connected to 127.0.0.1.Escape character is '^]'.stat #查看zk版本信息Zookeeper version: 3.4.12-e5259e437540f349646870ea94dc2658c4e44b3b, built on 03/27/2018 03:55 GMTClients: /127.0.0.1:52636[0](queued=0,recved=1,sent=0)Latency min/avg/max: 0/2/308Received: 248Sent: 247Connections: 1Outstanding: 0Zxid: 0xcMode: standaloneNode count: 5Connection closed by foreign host.
[root@zookeeper bin]# telnet 127.0.0.1 2181Trying 127.0.0.1...Connected to 127.0.0.1.Escape character is '^]'.conf #查看节点配置信息clientPort=2181dataDir=/opt/zookeeper/datadir/version-2dataLogDir=/opt/zookeeper/logdir/version-2tickTime=2000maxClientCnxns=60minSessionTimeout=4000maxSessionTimeout=40000serverId=0Connection closed by foreign host.
zookeeper伪集群部署
zookeeper伪集群是部署在一台服务器上,通过创建多个工作目录配置项等,实现多实例的运行
在上面的基础上,再添加两个实例
- 将原来/data/zookeeper1目录复制
[root@zookeeper ~]# cp -ar /data/zookeeper1/ /data/zookeeper2[root@zookeeper ~]# cp -ar /data/zookeeper1/ /data/zookeeper3[root@zookeeper ~]# ls /data/zookeeper1 zookeeper2 zookeeper3
- 修改zookeeper1配置文件
[root@zookeeper ~]# vim /data/zookeeper1/conf/zoo.cfg dataDir=/opt/zookeeper/datadir #各伪节点目录不能相同dataLogDir=/opt/zookeeper/logdirclientPort=2181 #各伪节点端口不能相同#下面的节点定义各伪节点要相同server.1=192.168.214.171:2888:3888 server.2=192.168.214.171:2889:3889server.3=192.168.214.171:2890:3890
- 修改zookeeper2配置文件
[root@zookeeper ~]# vim /data/zookeeper2/conf/zoo.cfg dataDir=/opt/zookeeper2/datadirdataLogDir=/opt/zookeeper2/logdirclientPort=2182server.1=192.168.214.171:2888:3888server.2=192.168.214.171:2889:3889server.3=192.168.214.171:2890:3890
- 修改zookeeper3配置文件
dataDir=/opt/zookeeper3/datadirdataLogDir=/opt/zookeeper3/logdirclientPort=2183server.1=192.168.214.171:2888:3888server.2=192.168.214.171:2889:3889server.3=192.168.214.171:2890:3890
- 创建数据存放目录
[root@zookeeper ~]# mkdir -pv /opt/zookeeper{2,3}/{datadir,logdir}mkdir: 已创建目录 "/opt/zookeeper2"mkdir: 已创建目录 "/opt/zookeeper2/datadir"mkdir: 已创建目录 "/opt/zookeeper2/logdir"mkdir: 已创建目录 "/opt/zookeeper3"mkdir: 已创建目录 "/opt/zookeeper3/datadir"mkdir: 已创建目录 "/opt/zookeeper3/logdir"
- 为每个节点设置节点ID号
[root@zookeeper ~]# echo 1 > /opt/zookeeper/datadir/myid[root@zookeeper ~]# echo 2 > /opt/zookeeper2/datadir/myid[root@zookeeper ~]# echo 3 > /opt/zookeeper3/datadir/myid[root@zookeeper ~]# cat /opt/zookeeper/datadir/myid version-2/ [root@zookeeper ~]# cat /opt/zookeeper/datadir/myid 1[root@zookeeper ~]# cat /opt/zookeeper2/datadir/myid 2[root@zookeeper ~]# cat /opt/zookeeper3/datadir/myid 3
- 启动集群
[root@zookeeper ~]# /data/zookeeper1/bin/zkServer.sh startZooKeeper JMX enabled by defaultUsing config: /data/zookeeper1/bin/../conf/zoo.cfgStarting zookeeper ... STARTED[root@zookeeper ~]# /data/zookeeper2/bin/zkServer.sh startZooKeeper JMX enabled by defaultUsing config: /data/zookeeper2/bin/../conf/zoo.cfgStarting zookeeper ... STARTED[root@zookeeper ~]# /data/zookeeper3/bin/zkServer.sh startZooKeeper JMX enabled by defaultUsing config: /data/zookeeper3/bin/../conf/zoo.cfgStarting zookeeper ... STARTED
- 查看状态
[root@zookeeper ~]# /data/zookeeper1/bin/zkServer.sh statusZooKeeper JMX enabled by defaultUsing config: /data/zookeeper1/bin/../conf/zoo.cfgMode: leader[root@zookeeper ~]# /data/zookeeper2/bin/zkServer.sh statusZooKeeper JMX enabled by defaultUsing config: /data/zookeeper2/bin/../conf/zoo.cfgMode: follower[root@zookeeper ~]# /data/zookeeper3/bin/zkServer.sh statusZooKeeper JMX enabled by defaultUsing config: /data/zookeeper3/bin/../conf/zoo.cfgMode: follower
- 测试master切换
关闭leader节点
[root@zookeeper ~]# /data/zookeeper1/bin/zkServer.sh stopZooKeeper JMX enabled by defaultUsing config: /data/zookeeper1/bin/../conf/zoo.cfgStopping zookeeper ... STOPPED#zookeeper2节点立即变成了leader[root@zookeeper ~]# /data/zookeeper2/bin/zkServer.sh statusZooKeeper JMX enabled by defaultUsing config: /data/zookeeper2/bin/../conf/zoo.cfgMode: leader
- 再启动zookeeper1节点,会以follower角色工作
[root@zookeeper ~]# /data/zookeeper1/bin/zkServer.sh startZooKeeper JMX enabled by defaultUsing config: /data/zookeeper1/bin/../conf/zoo.cfgStarting zookeeper ... STARTED[root@zookeeper ~]# /data/zookeeper1/bin/zkServer.sh statusZooKeeper JMX enabled by defaultUsing config: /data/zookeeper1/bin/../conf/zoo.cfgMode: follower
关于zookeeper的单机模式和伪集群就写到这,分布式集群部署