[SPARK-10810] Improve session management for SQL

• Type: Improvement
• Status: Resolved
• Priority: Blocker
• Resolution: Fixed
• Fix Version/s: 1.6.0
• Component/s: SQL
• Target Version/s: 1.6.0
• Assignee: Davies Liu
• Reporter: Davies Liu
• Created: 24/Sep/15 19:13
• Updated: 02/Nov/15 05:18
• Resolved: 09/Oct/15 00:34

Description

Currently, we try to support multiple sessions in SQL within a Spark Context, but it's broken and not complete.

We should isolate these for each session :

1) current database of Hive
2) SQLConf
3) UDF/UDAF/UDTF
4) temporary table

For added jar and cached tables, they should be accessible for all sessions.

Issue Links

• links to: [Github] Pull Request #8909 (davies)