作者：櫰木

1、集群规划

2、zookeeper kerberos主体创建

在生产中zk服务端和客户端票据可以设置成不通名称或相同名称，本次设置成同一个票据名称zookeeper/HOST@DTSTACK.COM格式表示zk服务端和客户端票据。

在hd1.dtstack.com主机上root权限下创建kerberos主体，使用zookeeper系统用

在每台机器上执行生成脚本，执行此脚本可自动生成principal（在每个节点执行）

bash /root/bigdata/getkeytabs.sh /etc/security/keytab/zookeeper.keytab  zookeeper

3、zookeeper 安装

修改配置文件

root@hd2.dtstack.com ~]# cd /root/bigdata && tar -xzvf apache-zookeeper-3.7.1-bin.tar.gz -C /opt
[root@hd2.dtstack.com ~]# ln -s /opt/apache-zookeeper-3.7.1-bin/ /opt/zookeeper
[root@hd2.dtstack.com ~]#cd /opt/zookeeper
[root@hd2.dtstack.com conf]#cat >zoo.cfg<<EOF
tickTime=2000
initLimit=10
syncLimit=5
dataDir=/data/zookeeper/data/
dataLogDir=/data/zookeeper/log/
clientPort=2181
maxCnxns=20000
maxClientCnxns=2000
minSessionTimeout=4000
maxSessionTimeout=60000
autopurge.purgeInterval=24
autopurge.snapRetainCount=5
quorum.cnxn.threads.size=20
#zk集群服务地址配置
server.1=hd1:2888:3888
server.2=hd2:2888:3888
server.3=hd3:2888:3888
#zk kerberos配置
authProvider.1=org.apache.zookeeper.server.auth.SASLAuthenticationProvider
jaasLoginRenew=3600000
kerberos.removeHostFromPrincipal=true
kerberos.removeRealmFromPrincipal=true
requireClientAuthScheme=sasl
quorum.auth.enableSasl=true
quorum.auth.learner.saslLoginContext=Learner
quorum.auth.server.saslLoginContext=Server
quorum.auth.kerberos.servicePrincipal=zookeeper/hd1.dtstack.com@DTSTACK.COM
4lw.commands.whitelist=mntr,conf,ruok,cons
EOF

说明：

• 改配置文件中ssl配置（标红色部分）目的是解决ranger 配置hive等组件在连接测试过程出现无权限问题
• 创建SSL认证文件zookeeper-jaas.conf、java.env

root@hd2.dtstack.com conf]# cat >zookeeper-jaas.conf<<EOF
Server {
  com.sun.security.auth.module.Krb5LoginModule required
  useKeyTab=true
  keyTab="/etc/security/keytab/zookeeper.keytab"
  storeKey=true
  useTicketCache=false
  principal="zookeeper/hd2.dtstack.com@DTSTACK.COM";
};
Client {
  com.sun.security.auth.module.Krb5LoginModule required
  useKeyTab=true
  keyTab="/etc/security/keytab/zookeeper.keytab"
  storeKey=true
  useTicketCache=false
  principal="zookeeper/hd2.dtstack.com@DTSTACK.COM";
};
Learner {
  com.sun.security.auth.module.Krb5LoginModule required
  useKeyTab=true
  keyTab="/etc/security/keytab/zookeeper.keytab"
  storeKey=true
  useTicketCache=false
  principal="zookeeper/hd2.dtstack.com@DTSTACK.COM";
};
EOF

说明：

• principal用具体票据名，不能用hadoop/host_name@DTSTACK.COM或hadoop/_HOST@DTSTACK.COM,否则报错
• zk服务端和客户端票据在kerberos中已经创建，此处省略，客户端和服务端票据是不一样的

root@hd2.dtstack.com conf]# cat >java.env<<EOF
export JVMFLAGS="-Djava.security.auth.login.config=/opt/zookeeper/conf/zookeeper-jaas.conf"
export JAVA_HOME="/opt/java"
EOF
 
[root@hd2.dtstack.com conf]# cat >/data/zookeeper/data/myid<<EOF
2
EOF

同步到其他机器

root@hd2.dtstack.com conf]# cd /opt/
[root@hd2.dtstack.com software]# scp -r apache-zookeeper-3.7.1-bin root@hd1.dtstack.com:/opt/
[root@hd2.dtstack.com software]# scp -r apache-zookeeper-3.7.1-bin root@hd3.dtstack.com:/opt/

在其他机器修改对应配置

hd1.dtstack.com主机root权限：

[root@hd1.dtstack.com conf]# cat >/data/zookeeper/data/myid<<EOF
1
EOF

[root@hd1.dtstack.com ~]# cd /opt/apache-zookeeper-3.7.1-bin/conf
[root@hd1.dtstack.com ~]# sed -i 's#hd2.dtstack.com#hd1.dtstack.com#g' zookeeper-jaas.conf
[root@hd1.dtstack.com ~]# sed -i 's#hd2.dtstack.com#hd1.dtstack.com#g' zoo.cfg

hd3.dtstack.com主机root权限：

[root@hd3.dtstack.com conf]# cat >/data/zookeeper/data/myid<<EOF
3
EOF
[root@hd3.dtstack.com ~]# cd /opt/apache-zookeeper-3.7.1-bin/conf
[root@hd3.dtstack.com ~]# sed -i 's#hd2.dtstack.com#hd3.dtstack.com#g' zookeeper-jaas.conf
[root@hd3.dtstack.com ~]# sed -i 's#hd2.dtstack.com#hd3.dtstack.com#g' zoo.cfg

4 zookeeper集群启停

zk集群启停脚本zk_cluster.sh内容如下：

[root@hd1.dtstack.com apache-zookeeper-3.7.1-bin]# cat >zk_cluster.sh<<EOF
#!/bin/bash
 case $1 in
"start"){
for i in hd1.dtstack.com hd2.dtstack.com hd3.dtstack.com
do
        echo ---------- zookeeper $i 启动 ------------
ssh $i "source /etc/profile;/opt/zookeeper/bin/zkServer.sh start"
done
};;
"stop"){
for i in hd1.dtstack.com hd2.dtstack.com hd3.dtstack.com
do
        echo ---------- zookeeper $i 停止 ------------    
ssh $i "source /etc/profile;/opt/zookeeper/bin/zkServer.sh stop"
done
};;
"status"){
for i in hd{1..3}
do
        echo ---------- zookeeper $i 状态 ------------    
ssh $i "source /etc/profile;/opt/zookeeper/bin/zkServer.sh status"
done
};;
esac
EOF

修改整体目录文件权限

[root@hd1.dtstack.com apache-zookeeper-3.6.3-bin]# chown -R zookeeper:zookeeper /opt/apache-zookeeper-3.7.1-bin

启动命令：

sh zk_cluster.sh start

停止命令：

sh zk_cluster.sh stop

查看状态命令：

sh zk_cluster.sh status

5 zookeeper集群验证

可通过执行sh zk_cluster.sh status查看

端口查看，命令netstat -an|grep 2181

进程查看，命令jps

至此，zk三节点集群搭建完成

更多技术信息请查看云掣官网https://yunche.pro/?t=yrgw