THIS IS A TEST INSTANCE. ALL YOUR CHANGES WILL BE LOST!!!!
...
- Name → local name to be able to reference the connector with. This name is shown in "show connectors" and will be used in connector DDLs like drop/alter and also in "create remote database .." statements.
- TYPE → "HIVEJDBC" so Hive Metastore knows that Connector class to use.
- URL → JDBC URL for the remote HiveServer instance.
- DCPROPERTIES → This is a freeform list that contains other info like credentials and other optional properties. These properties will be passed onto the table definitions for the databases created using this connector.
How do I use it?
- Create a connector first.
CREATE CONNECTOR hiveserver_connector TYPE 'hivejdbc' URL 'jdbc:hive2://<maskedhost>:10000' WITH DCPROPERTIES ("hive.sql.dbcp.username"="hive", "hive.sql.dbcp.password"="hive");
- Create a database of type REMOTE in hive using the connector from Step1. This maps a remote database named "
default
" to a hive database named "hiveserver_remote
" in hive.
CREATE REMOTE DATABASE hiveserver_remote USING hiveserver_connector WITH DBPROPERTIES ("connector.remoteDbName"="default");
3. Use the tables in REMOTE database much like the JDBC-storagehandler based tables in hive. One big difference is that the metadata for these tables are never persisted in hive. Currently, create/alter/drop table DDLs are not supported in REMOTE databases.
0: jdbc:hive2://localhost:10000> USE hiveserver_remote;
...