RE: separate tablespaces for tables and indexes

  • From: "Knight, Jon" <jknight@xxxxxxxxxxxxxx>
  • To: Oracle-L Freelists <oracle-l@xxxxxxxxxxxxx>
  • Date: Mon, 13 Dec 2004 13:41:52 -0600

Sorry if I missed this already, but ...

Say session A has already read the index and is doing some datafile access,
when along comes session B with an index range scan.  My understanding is
that there's only one read/write head on the disk and contention results.
Of course 10s or 100s of sessions would compound the issue.  Is this
reasoning flawed?

Thanks,
Jon Knight

 -----Original Message-----
From:   oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx]
On Behalf Of David Sharples
Sent:   Monday, December 13, 2004 1:31 PM
To:     Oracle-L Freelists; thomas.mercadante@xxxxxxxxxxxxxxxxx
Subject:        Re: separate tablespaces for tables and indexes

well a table and its index were never accessed at the same time anyway
- always serially., always one then the other - so it doesnt matter if
they were in the same tabespace / disk anyway.

Back in the old todays, before SAN's and striping and all that placing
hot segments on different disks was a good idea, but a process using a
table and its index would never benefit from it

So doing it for performance was / is a 'myth'.  For management sure
splitting them could be a good idea

see
http://asktom.oracle.com/pls/ask/f?p=4950:8:::::F4950_P8_DISPLAYID:901906930
328
for a good explanation

Dave
On Mon, 13 Dec 2004 14:19:56 -0500, Mercadante, Thomas F
<thomas.mercadante@xxxxxxxxxxxxxxxxx> wrote:
> David,
> 
> Please describe your definition of "never".  :)
> 

> -----Original Message-----
> From: David Sharples [mailto:davidsharples@xxxxxxxxx]
> Sent: Monday, December 13, 2004 2:08 PM
> To: oracle-l@xxxxxxxxxxxxx
> Subject: Re: separate tablespaces for tables and indexes
> 
> never was any need
--
//www.freelists.org/webpage/oracle-l
--
//www.freelists.org/webpage/oracle-l

Other related posts: