Hive客户端访问

适用模块

hive

具体说明

通过beeline访问Hive：
可将以下语句创装为shell脚本，通过有数离线开发script节点进行调度；
如需在用户自定义客户端执行，需添加kerberos认证，配置详见
https://study.sf.163.com/documents/read/service_support/dsc-p-a-0006

使用示例

## kerberos认证，其中keytab_dir提替换为具体keytab路径，keytab_principal需根据klist -kt  xxx.keytab结果获取
## kinit -kt keytab_dir/xxx.keytab   keytab_principal

## -u参数后接Hiveserver2链接串，有数已实现通过ZK保障Hiveserver2服务高可用，链接信息可咨询运维人员或网易技术支持；
## -e 参数后接要执行的SQL语句

$HIVE_HOME/bin/beeline
-u "jdbc:hive2://demo1.jd.163.org:2181,demo2.jd.163.org:2181,demo3.jd.163.org:2181/;serviceDiscoveryMode=zooKeeper;zooKeeperNamespace=hive-cluster10;principal=hive/_HOST@HADOOP2.HZ.NETEASE.COM" 
-e "set mapreduce.job.queuename=root.xxx.default;select a.employee_id,a.employee_name,18,null,null,a.ds from ods_employee_info;"
连spark
./current/bin/beeline -u "jdbc:hive2://demo01.jdlt.163.org:2182,demo02.jdlt.163.org:2182,demo03.jdlt.163.org:2182/;spark.yarn.queue=root.bdms.sla;serviceDiscoveryMode=zooKeeper;zooKeeperNamespace=kyuubi-cluster;principal=hive/_HOST@BDMS.COM" -e "use bdms;show table extended like 'xh';"
或
./current/bin/beeline -u "jdbc:hive2://dsc-demo19.jdlt.163.org:2182,dsc-demo17.jdlt.163.org:2182,dsc-demo18.jdlt.163.org:2182/;serviceDiscoveryMode=zooKeeper;zooKeeperNamespace=kyuubi-cluster;principal=hive/_HOST@BDMS_DEMO.COM;#spark.yarn.queue=root.support.default" -e "show databases;"

java 连spark
"jdbc:hive2://dsc-demoxx.jdlt.163.org:2182,dsc-demoxx.jdlt.163.org:2182,dsc-demoxx.jdlt.163.org:2182/;serviceDiscoveryMode=zooKeeper;zooKeeperNamespace=kyuubi-cluster;principal=hive/_HOST@BDMS_DEMO.COM;#spark.yarn.queue=root.support.default"

作者：林帅