Restore is done from media or tape to Teradata AMPs, meaning archive data is transferred to AMPs again. The target system could be different to that in which data is archived.
The following is the pictorial representation of the restore process:
There are a few things that we need to consider when doing a restore:
- Do not drop the database after you archive the object you want to restore. Use COPY if this occurs.
- Restore will drop any new object created since archiving.
- Space on targeted machines should be more or similar to the restoring object.
- JOIN INDEXES all need to be dropped and recreated if it's on any restored object ...