> I've been asked to prove why the Redos shouldn't be RAID 10, because the
> striping should allow for faster writes. I remember reading that this was
> nullified by the sequential nature of the writes, but I can't remember
where
> I saw this.

Well, if you have all the logs on since RAID 10, then the IOs caused by
archiving a log will be disturbing LGWR writes (causing lot's of extra head
movement for disks). This means occassional performance degradation for both
ARCH and LGWR processes.
But if you have spread your odd redos to one disk set and even ones to
another set, then the archiver can copy away the logs without disturbing
LGWR at all.

This is feasible approach for environments where disk IO on redolog spindles
is the bottleneck (btw, your SANs cache may relieve this issue a lot if
configured properly).

Tanel.


----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to: oracle-l-request@(protected)
put 'unsubscribe' in the subject line.
--
Archives are at http://www.freelists.org/archives/oracle-l/
FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------