Re: Disable logging in tablespace vs using hidden parameter _disable_logging

> That is correct, and that is why the only way
> the _disable_logging feature affects things is
> by not doing the final write from buffer to file.
For some cases event that bit can be significant I believe. During some upgrades, implementations, transitions phases then disk IO is bottle neck for particular process disabling logging can speed up thinks tremendously.

Of course gentleman who sad that if an application code written efficiently enough than disabling logging have little help. But we not all ways have power on application code.

J.

On 2/3/06, Jonathan Lewis <jonathan_at_jlcomp.demon.co.uk> wrote:
>
> That is correct, and that is why the only way
> the _disable_logging feature affects things is
> by not doing the final write from buffer to file.
>
> It does introduce the interesting question of
> how the forward change vectors for global
> temporary tables are "avoided".
>
>
> Regards
>
> Jonathan Lewis
>
> http://www.jlcomp.demon.co.uk/faq/ind_faq.html
> The Co-operative Oracle Users' FAQ
>
> http://www.jlcomp.demon.co.uk/cbo_book/ind_book.html
> Cost Based Oracle: Fundamentals
>
> http://www.jlcomp.demon.co.uk/appearances.html
> Public Appearances - schedule updated 10th Jan 2006
>
> ----- Original Message -----
> From: "Alex Gorbachev" <gorbyx_at_gmail.com>
> To: <jonathan_at_jlcomp.demon.co.uk>
> Cc: <oracle-l_at_freelists.org>
> Sent: Thursday, February 02, 2006 11:23 PM
> Subject: Re: Disable logging in tablespace vs using hidden parameter
> _disable_logging
>
>
> Jonathan,
> If I am not mistaken Oracle first generates change vector writing it
> to the log buffer and then applies it. At least, this is what I've
> been taught.
> If this stands correct than there is no way to really bypass redo log
> generation inside an Oracle instance. Is that correct?
> Thanks,
> Alex
>
> 2006/2/2, Jonathan Lewis <jonathan_at_jlcomp.demon.co.uk>:
> >
> >
> > For a large import, the point I was making about
> > commits is pretty irrelevant. I was only thinking
> > about the usual potential for latch collision and
> > redo wastage for highly concurrent systems, and
> > 'row at a time' batch processing.
> >
> > Everything that could normally cause a redo-related
> > problem still can cause a problem with _disable_logging
> > set to true apart from one little detail, which is the
> > log file write time.
> >
> > Regards
> >
> > Jonathan Lewis
>
>
> --
> http://www.freelists.org/webpage/oracle-l
>
>
>

--
Jurijs
+44 7738 013090 (GMT)
============================================
http://otn.oracle.com/ocm/jvelikanovs.html
--
http://www.freelists.org/webpage/oracle-l