[BUG] Connector write/read opens multiple connections

[Aryex]

Problem Description

It has been reported that the connector is creating multiple jdbc sessions. The connector should only create a constant number of JDBC connections on the driver node to prepare for read/write. The number of connections should not scale with executors created.

---

Environment

• Spark version: N/A
• Hadoop version: N/A
• Vertica version: N/A
• Vertica Spark Connector version: 3.2.1
• Java version: N/A
• Additional Environment Information: pyspark on Jupyter client

[jeremyprime]

FYI, when testing the connections to Vertica you can run any of the following queries in vsql to monitor the number of open sessions (be aware the query itself will create 1 session):

$ docker exec -it docker_vertica_1 vsql
> SELECT COUNT(*) FROM v_monitor.sessions;
> SELECT session_id, user_name, current_statement, last_statement, statement_start FROM v_monitor.sessions;
> SELECT session_id, node_name, user_name, client_hostname, login_timestamp, statement_start, current_statement, last_statement, client_type, client_os FROM v_monitor.sessions ORDER BY user_name;