zookeeper分布式安装
1.先把zookeeper-3.5.2-alpha.tar上传解压到/usr/local/下重命名为zookeeper
2 ############# zookeeper env ################# 3 export ZOOKEEPER_HOME=/usr/local/zookeeper 4 export PATH=$JAVA_HOME/bin:$ZOOKEEPER_HOME/bin:$PATH
2.配置zookeeper各配置文件
1 [root@hadoop1 conf]# cd /usr/local/zookeeper/conf 2 [root@hadoop1 conf]# cp zoo_sample.cfg zoo.cfg 3 [root@hadoop1 conf]# vim zoo.cfg 最后追加 4 server.1=hadoop1:7000:7001 5 server.2=hadoop2:7000:7001 6 server.3=hadoop3:7000:7001 7 [root@hadoop1 conf]# mkdir -p /tmp/zookeeper # 所有机器,创建指定的dataDir 8 [root@hadoop1 conf]# touch /tmp/zookeeper/myid # 所有机器,/tmp/zookeeper是刚才zoo.cfg里指定的dataDir,如果变更了这里也要变 9 [root@hadoop1 conf]# echo "1">/tmp/zookeeper/myid # 所有机器操作,与zoo.cfg最后追加的对应,hadoop2为2,hadoop3为3
3.修改日志记录方式,将原本的一直累计追加变更为每日一份
1 [root@hadoop1 conf]# vim log4j.properties 2 zookeeper.root.logger=INFO, CONSOLE变更为zookeeper.root.logger=INFO, ROLLINGFILE 3 log4j.appender.ROLLINGFILE=org.apache.log4j.RollingFileAppender变更为log4j.appender.ROLLINGFILE=org.apache.log4j.DailyRollingFileAppender 4 [root@hadoop1 conf]# vim ../bin/zkEnv.sh 5 ZOO_LOG4J_PROP="INFO,CONSOLE"变更为ZOO_LOG4J_PROP="INFO,ROLLINGFILE" 6 [root@hadoop1 local]# scp -rp zookeeper/ hadoop2:/usr/local/ 7 [root@hadoop1 local]# scp -rp zookeeper/ hadoop3:/usr/local/
4.启动并验证
1 [root@hadoop1 conf]# zkServer.sh start # 所有机器 2 [root@hadoop1 conf]# jps 3 18628 Jps 4 18484 QuorumPeerMain 5 [root@hadoop1 conf]# zkServer.sh status 6 ZooKeeper JMX enabled by default 7 Using config: /usr/local/zookeeper/bin/../conf/zoo.cfg 8 Client port found: 2181. Client address: localhost. 9 Mode: follower # 会有一台是leader,剩下都是follower,因为zookeeper是完全对等的,所以不一定是当前操作的hadoop1就是leader