There are different kinds of bad sector recovery, so the bad sector's data isn't always automatically moved to a known good sector. Only during write is a persistently bad sector removed from use with certainty.
smartmontools should be used to do periodic extended (long) tests, which will show the first LBA with a read error. As soon as such errors appear the disk should be zero'd (or ideally ATA Secure Erased), which will cause persistent bad sectors to be removed (i.e. no corresponding LBA). The disk can then be used again, with vigilant SMART testing for additional errors. If it continues to be a recurring problem, I'd have it swapped out under warranty.