Subject | Best way to archive |
---|---|
Author | Bob Murdoch |
Post date | 2004-02-17T13:15:42Z |
I've got a 10GB database that contains four years of information. We
really only need to have two years available in this database for online
use. While we should, theoretically, be just fine keeping all four years
in the same DB, we have found the file size to be difficult to
handle. What I would like to do is split this DB into two, making one a
read only archive of years 1 and 2, and keeping years 3 and 4 online.
I'm looking for the best approach to accomplishing this. I have started by
making a copy of the database, from which I will delete years 3 and 4. I
also plan on deleting years 1 and 2 from the 'online' copy. My plan is to
make the deletions, immediately do a backup without garbage collection,
then do a restore.
Is there any reason why I should delete in increments, committing after
each? Or is it safe to delete a million records in one transaction?
Any other tips appreciated,
Bob M..
really only need to have two years available in this database for online
use. While we should, theoretically, be just fine keeping all four years
in the same DB, we have found the file size to be difficult to
handle. What I would like to do is split this DB into two, making one a
read only archive of years 1 and 2, and keeping years 3 and 4 online.
I'm looking for the best approach to accomplishing this. I have started by
making a copy of the database, from which I will delete years 3 and 4. I
also plan on deleting years 1 and 2 from the 'online' copy. My plan is to
make the deletions, immediately do a backup without garbage collection,
then do a restore.
Is there any reason why I should delete in increments, committing after
each? Or is it safe to delete a million records in one transaction?
Any other tips appreciated,
Bob M..