| 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: (((Re))): db_file_mutliblock_read_count and physical IO
That's correct. Each db file scattered read wait event represents a =
readv() system call. It's essentially Oracle saying "Hmmmm...I need =
this and it's not in the buffer cache. O/S, please get it for me." For =
the duration of the readv() call, the server process waits on a db file =
scattered read event. As Cary mentioned previously, the O/S may find =
the data in a variety of places: a filesystem buffer cache (if you're =
not running raw or directio), or in a disk subsystem cache (EMC, NetApp, =
etc, most have large caches), or in a disk controller cache, or possibly =
in the disk's local cache, or, ultimately, on the physical disk platter =
itself.
-Mark
-----Original Message-----
From: oracle-l-bounce_at_freelists.org
[mailto:oracle-l-bounce_at_freelists.org]On Behalf Of ryan gaffuri
Sent: Wednesday, August 18, 2004 5:43 PM
To: oracle-l_at_freelists.org
Subject: (((Re))): db_file_mutliblock_read_count and physical IO
so how do i determine by looking at trace files whether changing the
db_file_multiblock_read_count makes a difference? I am assuming I should =
see
a reduction in the sheer number of 'db file scattered reads' in my trace
file correct?
this is more for general knowledge. I am not trying to tune by strictly
focusing on physical IO.
----- Original Message -----=20
From: "Wolfgang Breitling" <breitliw_at_centrexcc.com>
To: <oracle-l_at_freelists.org>
Sent: Wednesday, August 18, 2004 12:39 PM
Subject: Re: db_file_mutliblock_read_count and physical IO
> tkprof summarizes and for "disk IO" it counts the number of blocks =
read,
> not the number of physical I/O. So if you have 1 scattered read that =
read
> 128 blocks, or 16 scattered read that each read 8 blocks, tkprof will
> report 128 "disk IO"
> That is one of the (many) reasons Cary advocates to look at the raw
-- Archives are at http://www.freelists.org/archives/oracle-l/ FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html ----------------------------------------------------------------- ---------------------------------------------------------------- Please see the official ORACLE-L FAQ: http://www.orafaq.com ---------------------------------------------------------------- To unsubscribe send email to: oracle-l-request_at_freelists.org 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 -----------------------------------------------------------------Received on Wed Aug 18 2004 - 16:57:43 CDT
 |
 |