Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
![]() |
![]() |
Home -> Community -> Mailing Lists -> Oracle-L -> Re: disk_async_io vs. multiple db_writer_processes
We moved to ODM last weekend and saw a drop of avg cpu
usage from 90% to 60%. Didn't modify anything at
database-level (I had a thread related to that change
earlier). Currently, we have disk_async_io=true,
filesystemio_options=asynch and db_writer_processes=6
(default, based on 48 cpus).
Deepak
>
> You can also look at the option of using quick i/o
> or ODM.
> I normally set both disk_async_io to TRUE and
> value of db_writer_processes > 1 after ensuring i/o
> is balanced using some type of logical/physical
> striping with proper configuration of I/O sizes.
>
>
> Deepak Sharma <sharmakdeep_oracle_at_yahoo.com> wrote:
> Following is an excellant place to start:
> http://www.ixora.com.au/notes/direct_io.htm
>
>
>
> > Hello all:
> > I would like to ask for your expert opinions on
> > the topic of using disk_async_io, and the use of
> > multiple db_writer_processes. We're running
> > Oracle E-business suites 11.5.10(9i-9205-32bit)
> > on Sun E6900, 12 CPUs, 48GB RAM, Solaris 9/64bit.
> >
> >
> > By default, disk_async_io, filesystemio_options;
> > db_writer_processes are default to TRUE, ASYNC,
> > and 3. The database files(redologs, datafiles,
> > controlfiles) are allocated on Sun 3510 storage
> > arrays(Hardware RAID10). The UFS filesystems are
> > mounted with forcedirectio option.
> >
-- http://www.freelists.org/webpage/oracle-lReceived on Sat Mar 04 2006 - 14:54:00 CST
![]() |
![]() |