Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
![]() |
![]() |
Home -> Community -> Mailing Lists -> Oracle-L -> Re: ODM, disk_asynch_io and db_writer_processes
> OS: Sun Solaris 2.8
> Database: Oracle 10.1.0.4
>
> We plan on implementing the ODM (Oracle Disk
> Manager),
> on our database in a few weeks. The datafiles
> currently are on vxfs-based filesystems. Apparently
> no
> change needs to happen on the datafiles itself, but
> I
> have a few questions ;
>
> disk_asynch_io should be set to TRUE (default), in
> order to make use of Solaris asynch i/o capability
> once ODM is enabled.
>
> The questions are;
>
> 1) What value should be set for db_writer_processes
> ?
> We have 24 cpus and the value (if not set
> explicitly)
> that Oracle would choose is 3 (cpu_count/8). Or,
> should we need to set db_writer_processes=1, since
> we
> are specifying disk_asynch_io=TRUE, and O/S would
> take
> care of async operations.
>
> 2) What value should the filesystemio_options be set
> to? By default we have it set to 'asynch' currently.
I didn't get any response to the above, so let me rephrase the question.
Let's say we are using RAW Filesystem as Oracle database files, then what values should be specified for "db_writer_processes" and "filesystemio_options", assuming that "disk_asynch_io=TRUE" ?
Thanks,
Deepak
-- http://www.freelists.org/webpage/oracle-lReceived on Thu Feb 16 2006 - 10:55:29 CST
![]() |
![]() |