We have a reasonably large database (~ 5 - 10 Tb ) and we're trying to
tune the i/o. One of the parameters we are looking at is
db_file_multiblock_read_count. in a system that is predominantly
random i/o what affect does this value have?

On 5 Nov 2004 06:32:28 -0800, strepxe@yahoo.co.uk (championsleeper)
wrote:

>We have a reasonably large database (~ 5 - 10 Tb ) and we're trying to
>tune the i/o. One of the parameters we are looking at is
>db_file_multiblock_read_count. in a system that is predominantly
>random i/o what affect does this value have?

zero.
db_file_multiblock_read_count applies to full table scans and index
fast full scans, so non-random i/o.
Did you ever look up the explanation of this parameter in your Oracle
docs. Probably not.


--
Sybrand Bakker, Senior Oracle DBA

championsleeper wrote:

> We have a reasonably large database (~ 5 - 10 Tb ) and we're trying to
> tune the i/o. One of the parameters we are looking at is
> db_file_multiblock_read_count. in a system that is predominantly
> random i/o what affect does this value have?

Before you get there ... what is your PCTFREE and PCTUSED?

How much time are you wasting reading air?

The reason I asked is we made a huge improvement in a local phone
company's month-end reports by just teaching them the concept of
PCTFREE 0.

--
Daniel A. Morgan
University of Washington
damorgan@x.washington.edu
(replace 'x' with 'u' to respond)

strepxe@yahoo.co.uk (championsleeper) wrote in message news:<103a78f3.0411050632.15da869a@posting.google. com>...
> We have a reasonably large database (~ 5 - 10 Tb ) and we're trying to
> tune the i/o. One of the parameters we are looking at is
> db_file_multiblock_read_count. in a system that is predominantly
> random i/o what affect does this value have?

Depending on your exact version and the use of system statistics with
version 9+ larger values for the db_file_multiblock_read_count
parameter influence the cost based optimizer to favor full table scans
over using indexes.

HTH -- Mark D Powell --

In article <2687bb95.0411051826.1cb2de99@posting.google.com > Mark.Powell@eds.com (Mark D Powell) wrote:
>
>strepxe@yahoo.co.uk (championsleeper) wrote in message news:<103a78f3.0411050632.15da869a@posting.google. com>...
>> We have a reasonably large database (~ 5 - 10 Tb ) and we're trying to
>> tune the i/o. One of the parameters we are looking at is
>> db_file_multiblock_read_count. in a system that is predominantly
>> random i/o what affect does this value have?
>
>Depending on your exact version and the use of system statistics with
>version 9 larger values for the db_file_multiblock_read_count
>parameter influence the cost based optimizer to favor full table scans
>over using indexes.

Research is being done on faster units with more accurate clocks. We
can now go back farther with a higher degree of divergence confidence.

>HTH -- Mark D Powell --

Clouds gather before a storm.

--
Lady Chatterly

"So in your opinion, a bot (Chatterly) proves that women deserves to
be beaten? God, you guys are clever." -- knoxy