Re: Linux File system and Oracle setup

Fabrizio apparently said,on my timestamp of 11/06/2005 8:45 PM:

> data=journal mode provides full data and metadata journaling. All new data
> is written to the journal first, and then to its final location.
> In the event of a crash, the journal can be replayed, bringing both data
> and metadata into a consistent state.

Which is completely and totally redundant for a database like Oracle. Only metadata (writeback mode) needs to be journaled in a file system being used by datafiles of an Oracle database: the data blocks are taken care of by the database's own redo/undo mechanism). And then only if and when you are creating the data files or expanding them: no other metadata is touched for normal I/O ops. Tempfiles being an exception.

Journaling data blocks in a file system used by Oracle database datafiles is a total waste of resources as the f/s journaling will never ensure a consistent database in the event of a crash - unless you run ALL database files, INCLUDING redo logs and control files, in the SAME file system! Do you? I don't think anyone does, then again... ;)

> The last mode, data=journal, requires a larger journal for reasonable
> speed in most cases and therefore takes longer to recover in case of
> unclean shutdown, but is sometimes faster for certain database operations.

Yeah, but the M$ question is still: *how much* larger? And what exactly is the nature of the "sometimes" and the "certain" db operations? What about all "other" ops? :)

-- 
Cheers
Nuno Souto
in sunny Sydney, Australia
wizofoz2k_at_yahoo.com.au.nospam