I am researching the Spark2 as a part of HDP 2.6.5. The most interesting part is STS, which provides SQL interface via JDBC. But I want to have a possibility to organize concurrent access to STS from different client, but STS works on Derby. Derby itself allows to have only one connection at the moment. So, is it possible to switch the installed STS to PostgreSQL instead of Derby? What is a recommended approach to organize a concurrent access to STS?
Please check the configuration. In HDP, STS doesn't use Derby by default. If the configuration is correct, STS should connect to Hive Metastore Service (which is running on MySQL/Postgres) instead.
Actually you are right, it uses my hive metastore running on top of PostgreSQL, and I can see all objects created in Hive before. But I saw in log-file mention about derby, if I rename derby jar-file, my sts installation can not start. And additionally to that, I am able to connect to the STS only via beeline installed locally on the same host with STS. All remote connections via DBeaver which is a JDBC-client are refused. At the same time I am able to connect to Hive with DBeaver. So, I strongly believe that internally it still uses derby.
This is the log entry that I am talking about:
INFO Query: Reading in results for query "org.datanucleus.store.rdbms.query.SQLQuery@0" since the connection used is closing INFO MetaStoreDirectSql: Using direct SQL, underlying DB is DERBY INFO ObjectStore: Initialized ObjectStore