Resolution:
The schema files in the Repo manual include operations like drop table/index/sequence and create table/index/sequence. The user id used in dbinit script file needs to have these capabilities.
      
Once the tables are created with the init script, the only privileges needed are table updates.  We add rows, delete rows and modify cell values. The one exception is that if the schema needs to be modified because of a version update, then the userid used by the server would need authority to add, drop and alter tables.

Even for Repository Manager, delete/add repository instances operation will only need "table update" privileges. That is because the information for all instances is stored in the same tables. So there is no need to add/drop tables when you create/delete instances.  

Based on the previous discussion, the user id used in the dbinit script needs the create/drop table/index/sequence, whereas the user id used in the server.ini file for routine repository operation will only need the "update table" privilege.  So these two users can be different.