Java使用hive连接kyuubi

一、Maven依赖

<dependency><groupId>org.apache.hive</groupId><artifactId>hive-jdbc</artifactId><version>2.3.9</version>
</dependency>

二、相关配置信息

驱动类：org.apache.hive.jdbc.HiveDriver连接URL:jdbc:hive2://xxx:2181,xxx:2181/;principal=hadoop/xxx@EMR-GMxxx;serviceDiscoveryMode=zooKeeper;zooKeeperNamespace=kyuubi;?spark.app.name=test-kyuubi;kyuubi.engine.share.level=USER;spark.submit.deployMode=cluster;kyuubi.session.engine.check.interval=PT5M;kyuubi.session.engine.idle.timeout=PT30M;
【注意】：
principal 配置的信息为账号鉴权信息
zooKeeperNamespace 需要指定为 kyuubi
?后的参数为kyuubi配置参数

三、 Kerberos 身份验证

1、需要准备两个文件，可以从集群kerberos配置里拷贝
krb5 /app/data/config/kerberos/krb5.conf
krb5Keytab /app/data/config/kerberos/emr.keytab2、获取账号
krb5hadoop hadoop/xxxx@EMR-QX9xxxx3、加载代码
/*** 加载hive krb 配置*/public void loadingHiveKrbConfig() {try {log.info("加载hive krb 配置");System.setProperty("java.security.krb5.conf", krb5);Configuration configuration = new Configuration();configuration.setBoolean("hadoop.security.authorization", true);configuration.set("hadoop.security.authentication", "Kerberos");UserGroupInformation.setConfiguration(configuration);if (UserGroupInformation.isLoginKeytabBased()) {log.info("UserGroupInformation.isLoginKeytabBased() is ture");UserGroupInformation.getLoginUser().reloginFromKeytab();} else {log.info("UserGroupInformation.isLoginKeytabBased() is false");UserGroupInformation.loginUserFromKeytab(krb5hadoop, krb5Keytab);}log.info("ticketCache=====>" + UserGroupInformation.isLoginTicketBased());} catch (Exception e) {log.error("加载hive krb 配置 失败：", e);e.printStackTrace();}}

四、连接代码

Connection connection = null;HiveStatement statement = null;HiveQueryResultSet resultSet = null;try {loadingHiveKrbConfig();Class.forName("org.apache.hive.jdbc.HiveDriver"));connection = DriverManager.getConnection(url);statement = (HiveStatement) connection.createStatement();resultSet = (HiveQueryResultSet) statement.executeQuery(sql);} catch (Exception e) {e.printStackTrace();} finally {try {if (null != resultSet) {resultSet.close();}if (null != statement) {statement.close();}if (null != connection) {connection.close();}} catch (final Exception e) {e.printStackTrace();}}