Adrian Klaver <adrian.klaver@xxxxxxxxxxx> writes: > --clone I think --clone is probably contraindicated here, given that Pierre already made a copy of the data. If I understand how that works, it'll just wind up making another whole copy, but in a time-extended manner as the tables are modified. Over the long run there would still be two copies of the DB on the new disk, which doesn't seem like what he wants. regards, tom lane