Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
Home -> Community -> Usenet -> c.d.o.server -> Re: Enable 32K Block in 8K Block DB
David Williams wrote:
> "Frank van Bortel" <fvanbortel_at_netscape.net> wrote in message
> news:c4otl4$3fg$1_at_news3.tilbu1.nb.home.nl...
>
>>Bottom line (still) is - don't use buffered IO; use direct or raw >>(why do you think every benchmark of oracle still uses raw?!?)
Well , the ones I read - yes.
But it may have shifted towards fs
(just as ext2 used to outperform ext3,
but not anymore, it seems. Still something I'd like
to test).
OK - latest from tpc.org; full disclosure doc, App. C:
mknod /home/oracle/dev/ocr c 162 200
mknod /home/oracle/dev/quorum c 162 201
raw /home/oracle/dev/ocr /dev/sdgf1 raw /home/oracle/dev/quorum /dev/sdav1 chown oracle:oracle /home/oracle/dev/ocr chown oracle:oracle /home/oracle/dev/quorum
Looks like raw is still used (although I did not check all 206 pages to find out how, where and what)
> "according to Oracle raw devices are not faster, in fact filesystems are
> quicker"
>
> FROM MY CUSTOMERS than?
> ^^^^^^^^^^^^^^^^^^^^^^
>
> Used to really annoy me...so Oracle have come around to my way of
> thinking than?
> As it's faster? What a suprise...
>
>
>
>>Now - how do I convince the SA to go raw on his HP machine with VX7100, >>where he found a benchmark that proofs he was right to create the >>filesystem with 8k blocks?
-- Regards, Frank van BortelReceived on Tue Apr 06 2004 - 14:23:49 CDT