Hi

Everyone ..Answer given by all others are right and i am just adding a brief to it. actually what happens in backend when a tablespace is put up in backup mode is that the all datafile belonging the tablespace gets their checkpoint frozen. which means that next time the recovery should start after the checkpoint anything after the checkpoint is roll backward or roll forward to maintain consistency. At the same time database writter process writes to the datafile all the data blocks and the same data blocks are written to redo log files this is because the oracle block size and OS block sizes differs which results in fractured block to be copied by the OS utilities and for that reason when the file backuped up by hot backup required recovery and any fractured blocks are found out they are replaced by the actual block from archived log which are copy of redo logs.RMAN doesn't requires the tablespace to be put in the backup mode because oracle understands the block size and skips the block if a process is making change to it at the time of the backup.when you put the tablespace end backup the checkpoint is released and marked wih current checkpoint No and in this way oracle maintains consistency in the database.

Regards

Ritesh Raj Singh

Database Administration Team

ICICI Ltd

ICICI Corporate House

Mahalakshmi Mumbai

09322070038