Re: Why "Separating Data and Indexes improves performance" is a myth?
- From: Paul Drake <discgolfdba@xxxxxxxxx>
- To: oracle-l@xxxxxxxxxxxxx
- Date: Sat, 24 Apr 2004 20:30:18 -0700 (PDT)
--- Jared Still <jkstill@xxxxxxxxxx> wrote:
> Jeez Paul, it was just an example. :)
>
> Jared
but when I hear "5 disks", it makes me think of ...
dirty things.
I never did try a block size of 6*1024 bytes.
it would be interesting for a 6 disk, RAID 10 vol.
I would think that non-power of 2 block sizes would
subject the actor with ORA-600s beyond their wildest
dreams.
btw - I have not experienced the degredation reported
by Jonathan Lewis (~12) regarding not getting 16
blocks per read for full table scans (or 32 when using
a dbfmbrc of 32). But I only use powers of 2 for RAID
stripes.
Pd
>
> On Fri, 2004-04-23 at 18:15, Paul Drake wrote:
> > OOH - odd number of disks for RAID 0 >> BRAIN
> HURTS.
> >
> > These days, I can only do math in my head in
> powers of
> > 2.
> >
> > 2 advil plus a dose of the paper at hotsos on
> > "Aligning Oracle Blocks with Hardware Stripe
> > Boundaries".
> >
> > RAID 1 ( 2 )
> > RAID 10 ( 4, 8)
> > RAID 0 ( 1, 2, 4, 8 )
> >
> > Far better to use that odd number for a hotspare
> > (although a hot spare for RAID 0 is not really
> > worthwhile).
> >
> > RAID F ... sorry, can't do it.
> >
> > Pd
__________________________________
Do you Yahoo!?
Yahoo! Photos: High-quality 4x6 digital prints for 25¢
http://photos.yahoo.com/ph/print_splash
----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx
put 'unsubscribe' in the subject line.
--
Archives are at //www.freelists.org/archives/oracle-l/
FAQ is at //www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------
Other related posts:
