Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Spanning a Database over Multiple Partitions

Re: Spanning a Database over Multiple Partitions

From: Bert Bear <bertbear_at_NOSPAMbertbear.net>
Date: Sat, 14 Dec 2002 23:13:27 GMT
Message-ID: <rWOK9.2204$Ih4.916887903@newssvr11.news.prodigy.com>


Hi,

Location, location, location. The worse thing for a database is bad disk location.

You might want to consider for your database LOTS of fast moderate sized drives (using RAID behind them), where you can spread the work load across in (what I like to call) three dimensions. The first two dimension is spreading the work load across the drives/strings and third dimension being striping each drive (since the third dimension should be invisible to the database/application).

Oracle has two important concepts logical data and physical data location. Tablespaces are logical. Datafiles are physical. They go hand in hand in defining your performance now and well into your future. It can really be marvious watching the two work hand in hand.

I strongly suggest you get hold of an Oracle DBA experienced in database setup (especially future performance). (S)he will want to talk to your end-users about their data and its structure and how they expect to query and add to it. Also affecting all of this is the type of environment - Data wharehousing .vs. online transactions ???

Anyway, you are at a wonderful point, the beginning. You don't have a legacy (at least you didn't tell us of one). Just remember: location, location, location.

:-)

Bertram Moshier
Oracle Certified Professional 8i and 9i DBA

http://www.bmoshier.net/bertram

"B. Kloppt" <b_kloppt_at_my-deja.com> wrote in message news:3717b54f.0212130852.28b920d1_at_posting.google.com...
> hi all,
>
> i am using oracle 9.2.0 and i need to create a very large database to
> accomodate a large influx of data.
> i have the following two partitions available on my oracle server:
> (excerpt from df -h)
> /dev/sda1 17G 33M 15G 1% /mnt/sda1
> /dev/sdb1 17G 33M 15G 1% /mnt/sdb1
>
> since i will be expecting a lot of data to be imported into this
> database, it will have to be very large, much larger than the 15GB
> free on each partition, therefore i was thinking of spanning it over
> the two partitions.
>
> if this is at all possible, my questions are:
> - how is this done? i understand that i would need to set the paths of
> the database files to point to, let's say, partition /mnt/sda1, but
> how would i set the size? i couldn't find any such options for
> database size to set in the dbca.
> - how do i make use of the second partition?
> - how do i make sure the fullest amount of the free space of these
> partitions is used for the database?
> - would setting only the paths of the control, redo and data files
> suffice or must i change the parameters, trace files and archive log
> files paths to these partitions as well?
>
> any help would be appreciated.
Received on Sat Dec 14 2002 - 17:13:27 CST

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US