Thread: PG_RESTORE and database size
Here's a basic question. On working with backing up and restoring it seems like if I keep on doing it in a row, the backup file size keeps on increasing in size. It almost doubles in size for every backup/restore. I want the restore the backup to overwrite the database and not add to it. Is there another procedure I'm missing or a parameter I'm not using right with pg_restore and pg_dump?
Thank you for any help.
Marcus Couto wrote: > Here's a basic question. On working with backing up and restoring it > seems like if I keep on doing it in a row, the backup file size keeps > on increasing in size. It almost doubles in size for every > backup/restore. I want the restore the backup to overwrite the > database and not add to it. Is there another procedure I'm missing or > a parameter I'm not using right with pg_restore and pg_dump? If I were to ask you about vacuum-ing would you know what I was talking about? If not, check the manuals and read up on the topic. -- Richard Huxton Archonet Ltd
"Marcus Couto" <marcus_couto@hotmail.com> writes: > Here's a basic question. On working with backing up and restoring it = > seems like if I keep on doing it in a row, the backup file size keeps on = > increasing in size. It almost doubles in size for every backup/restore. = > I want the restore the backup to overwrite the database and not add to = > it. Is there another procedure I'm missing or a parameter I'm not using = > right with pg_restore and pg_dump? Why are you restoring into a live database? It sounds to me like you are probably ignoring a lot of object-already-exists errors from pg_restore and then having it add on duplicate data to the tables. Usually people drop the database and recreate it empty before running pg_restore. There is a switch named --clean or something like that to make pg_restore emit DROP commands before recreating objects, but hardly anyone uses it because it's usually slower than dropping the whole database at once. regards, tom lane