Re: Redo logs only useful if archiving on

No, you can use the redo log files themselves in recovery IF they have not been overwritten since the cold backup.

You can do this by using more redo log files, bigger redo log files, or more bigger redo log files. With no guarantees!

Archive log mode guarantees that no redo log file will be overwritten until it has first been copied (archived).

Because Oracle's DBWR writes according to block LRU stats and not FIFO, your transaction may not be written to disk until a checkpoint. It WILL be written to the redo log file before you receive a return from a commit. We mirror the redo log files to be sure we do not have a single point of failure. A redo log switch always triggers a checkpoint and the archiving of the redo log file. Now our two copies of the data are the database and the archived redo log file.

Frank Calfo wrote:

> Seems to me that the redo logs can only help you recover
> data up to the point of failure if the database is in archive log mode.
> Is that true?
>
> Also we have carefully planned our database to place the redo logs
> on separate drives in case of media failure. But again, unless
> we're running in archive log mode, this step doesn't really seem to
> buy us much in terms of data protection. True ?
>
> (I assume we would still want to consider which drives the redo logs
> are on even without archiving for performance reasons.)
>
> For those of you running in archive log mode, is it working well or
> is it so cumbersome that its not worth the trouble? Would cold-backups
> be sufficient?
Received on Thu Jun 18 1998 - 23:09:48 CDT