| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Mailing Lists -> Oracle-L -> Re: disk_asynch_io and db_writer_processes on RAW DBFs
Mladen,
Why don't you just shut-up? I have seen your replies to other emails in a similar manner to mine.
I have been using Oracle since 1987 and very well know what Metalink is. FYI, I have posted this same question on Metalink before here. I thought this list is made of some knowledgeble and helpful people, except for a few stubborns.
I did do my homework, by looking at Steve Adams web-site, metalink, google and what not, but never could find anything related to ODM and these parameters, all in context of async i/o, hence this question to the list.
Finally, I think my question broke no rules as far as posting to this list is concerned. It is a very legitimite question, and thankfully to Steve, I got my answer.
Hope you understand, and before replying to anyone, show little concern.
Thanks,
Deepak
>
> On 02/19/2006 10:56:11 PM, Deepak Sharma wrote:
> > OS: Sun Solaris 2.8
> > Database: Oracle 10.1.0.4
> > CPU_COUNT: 32
> >
> > On a database using RAW datafiles, what values
> > should be specified for "db_writer_processes"
> > and/or "filesystemio_options", assuming that
> > "disk_asynch_io=TRUE" (also async i/o is available
> at
> > OS-level) ?
>
> There is Oracle's on-line support page called
> Metalink. Its URL is
> http://metalink.oracle.com. My understanding is that
> this list is
> not a plug-in replacement for metalink. Having said
> that and given
> your repeatedly demonstrated unwillingness to do
> your own homework,
> let me answer the question. There is a metalink note
> 47328.1 with
> the following title: Reference Note for Init.Ora
> Parameter "DISK_ASYNCH_IO".
> To some people, me included, it would look like a
> clue for searching
> the answers. Here is what it says:
>
> "DISK_ASYNCH_IO controls whether I/O to datafiles,
> control files, and logfiles is asynchronous (that
> is, whether parallel server processes can overlap
> I/O requests with CPU processing during table
> scans). If your platform supports asynchronous I/O
> to disk, Oracle Corporation recommends that you
> leave this parameter set to its default value.
> However, if the asynchronous I/O implementation is
> not stable, you can set this parameter to false to
> disable asynchronous I/O. If your platform does not
> support asynchronous I/O to disk, this parameter has
> no effect."
>
> and then:
>
> "If DISK_ASYNCH_IO is set to FALSE, then typically
> either
> <Parameter:DBWR_IO_SLAVES> or
> <Parameter:DB_WRITER_PROCESSES> should
> be set unless the system is low throughput."
>
> My understanding of that is that if you have
> asynchronous I/O, it will be used
> automatically. DB_WRITER_PROCESSES should be
> increased only if you experience
> extensive waits on buffer LRU latch. IF you're using
> dynamic SGA, this is automatically
> computed for you, as per Metalink note 148495.1
> Finally, the filesystemio_options do not matter for
> raw devices as raw devices
> are not filesystem. You automatically have
> asynchronous and direct I/O on raw
> devices.
-- http://www.freelists.org/webpage/oracle-lReceived on Mon Feb 20 2006 - 13:44:30 CST
 |
 |