Re: Different backup procedure.

Sure, you can put all the tablespaces into backup mode, execute the backup and then end backup mode for all the tablespaces.

The only bad effect is this:

While a tablespace is in backup mode, LGWR logs full database blocks for any insert, update or delete affecting that tablespace. If all tablespaces are in backup mode, then all logging is full block. If only one tablespace is in backup mode, then only logging of that tablespace is full block.

Is this important? If you are doing hot backup during heavy update time, you will feel it. If updating is light during the hot backup, it will not make any real difference.

So, it is better to do one tablespace at a time; but, if that is not available to you, mass backup mode does work.

David Shi wrote:

> In my dept, a software does all the backup automatically, it is not possible to
> have script to do "begin backup" "end backup" sequence. My question is, is it ok
> to do an "begin backup" on all the tablespaces and then do a whole datafile
> backup?
>
> I think I read somewhere this is not recommended, can any one give me a little
> insight into this? Thanks.
>
> David
>
> On Fri, 07 Aug 1998 22:40:46 -0500, "John P. Higgins" <jh33378_at_deere.com> wrote:
>
> >You MUST use archivelog mode to do hot backup.
> >
> >When you do hot backups, the database is up and users are updating!
> >
> >The basic hot backup script is:
> >
> > for each tablespace
> >
> > alter tablespace _______ begin backup:
> > for each datafile of the tablespace
> >
> > tar, cpio, dd, .... the file
> >
> > next file
> > alter tablespace _______ end backup;
> >
> > next tablespace
> >
> >Since the users are updating while the backup is running, the backup is
> >'fuzzy'. That is, the backup files are not from a single, consistent
> >point-in-time. So, if you restore all the database files from your hot
> >backup, you will need to apply the redo logs (recovery) that were created
> >during the backup. (You may need to apply even more logs, depending on the
> >point-in-time to which you are recovering.)
> >
> >This is why you need archivelog mode for hot backups: archivelog mode
> >ensures that you still have the redo log data that you need for recovery.
> >
> >_____________________________________________________________________
> >
> >Backups enable recovery. If you don't need recovery, you don't need backup!
> >
> >Cold backup enables recovery to the point-in-time of the cold backup.
> >Without archivelog mode, no further recovery is possible. With archivelog
> >mode, you can recover to any point-in-time up to current time.
> >
> >Hot backup enables recovery to the point-in-time where the hot backup
> >terminated. Since this requires archivelog mode, you can recover to any
> >point-in-time up to current time.
> >
> >Archivelog mode enables recovery of the work since the last hot or cold
> >backup. If you can afford to lose all work since the last backup, you don't
> >need archivelog mode!
> >
> >Archivelog mode also enables hot backup. If you can shut down your database
> >long enough for a cold backup, you don't need hot backup.
> >
> > John
> >
> >giau45_at_yahoo.com wrote:
> >
> >> I keep reading on Oracle backup procedure. And I'm still fuzzy about
> >> hot backup and archive mode.
> >>
> >> Cold backup: shutdown all DB and do a complete copy of dbf files (as I
> >> understand it)
> >>
> >> How is backup work and archive log mode work ? Thank you.
> >
Received on Mon Aug 10 1998 - 22:07:49 CDT