[HIVE-28442] Inconsistency in schema on tables TAB_COL_STATS, PART_COL_STATS, KEY_CONSTRAINTS and SCHEDULED_QUERIES on SYSDB after upgrade from 3.1.0 to 4.1.0

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 4.1.0
• Labels: pull-request-available
• Assignee: Indhumathi Muthumurugesh
• Reporter: Indhumathi Muthumurugesh
• Created: 12/Aug/24 07:56
• Updated: 24/Sep/24 09:33
• Resolved: 23/Aug/24 02:35

Description

Fresh Install: InitSchema

1. Tables SYSDB.TAB_COL_STATS and SYSDB.PART_COL_STATS

Select * from the above tables:

 

Upgrade Hive Schema from 3.2.0 to 4.1.0:

Select * on the tables fails post upgrade

Inconsistency in columns:

   1. Missing column Engine in upgrade Path.

       Looks like missed from HIVE-22046

   2. Redundant columns which got removed as part of HIVE-27725 is not handled in the upgrade path

   TAB_COL_STATS - DB_NAME and TABLE_NAME

   PART_COL_STATS - DB_NAME, TABLE_NAME and PARTITION_NAME

2. Table SYSDB.SCHEDULED_QUERIES **

     - Missing column  ACTIVE_EXECUTION_ID - Looks like missed from HIVE-22872

After Upgrade:

 Fresh install:

 

3. SYSDB.KEY_CONSTRAINTS

• Missing column constraint PARENT_TBL_ID - Looks like missed in HIVE-22728

Fresh Path:

Upgrade Path:

Issue Links

• links to: GitHub Pull Request #5392