Why oracle having two options here ?
It's a feature called Real Time Apply, which is also the base for Data Protection Mode other than MAXIMUM PERFORMANCE.
LGWR process on the primary sends the redo data. RFS processes receive this redo that is coming in and here is where the two options differ:
1) With Standby redo log files: RFS process writes the redo in a continuous manner, directly from the primary into the standby log files. Then, the MRP process keep reading and applying that redo as it is being written. This is called Real Time Apply, because the propagation delay of redo stream from primary to standby is minimal.
2) Without standby redo log files: RFS process writes the redo into the archivelog of the standby. This will only happen at each redo log switch of the primary, so the propagation delay is always one sequence behind. RFS writes the redo to the archivelog, and after that MRP will read the archivelog and apply it.
2) In physical standby directly applies redo to its data files. - Is this really true ?
It's a continous recovery process. MRP stands for Managed Recovery Process, in that regard it doesn't use the buffer cache for processing any DML. It applies the change vectors directly to the data files, yes.
Hi alavaro ;
Thanks for your comments.
1) Could you please post any official link to explain briefly regarding 'what you said' ?
2) Transferring redo information from primary to standby - Is this called as redo stream ?
3) Is this called as real time apply ?
In standby site , if oracle applies redo data only with standby redo log file.
4) Is there any option to find where the redo data is applied in stand by site ?
i mean ( in standby redo log files or archive redo logs ) ?
Thanks in advance.