Export – Import Very Large Database
Export – Import Very Large Database
Step 1:
EXP FULL=Y ROWS=N from the source-db then IMP FULL=Y ROWS=N INDEXES=N CONSTRAINTS=N into the target-db
The idea is to lay down the tablespaces and table definitions in the target-db. Depending on similarities (or the lack thereof) between the two dbs, you may or may not be able to create tablespaces using IMP, but the goal is to import tables DDL and the DDL for other objects (but not indexes and constraints) using IMP
Don’t forget to disable all DML triggers created in target-db before going on to the next step…
Step 2:
Disable Logging on All segment Level (Table & Table Space)
Step 3:
Run as many concurrent exp/imp streams as you can between the source-db and target-db to copy the table data across Goal is to keep as many streams running concurrently as possible. For smaller tables, one exp/imp stream per table. For larger tables, one exp/imp stream per partition or one exp/imp stream per section of the table specified using EXP QUERY=
Step 4:
Create indexes and constraints using previously-obtained EXP FULL=Y ROWS=N in step #1 Usually I just extract the DDL using IMP INDEXFILES= and create SQL*Plus scripts for this, so I can break the task up into many concurrently running jobs for speed…
Step 5:
Enable Logging and Trigger
