Hi,

I have a database of 250 GB of user data. How many dbspaces is
suitable for optimal performance:
- For OLTP transactions ?
- For queries ?
- For backup using onbar ?
Currently, I have 30 dbspaces configured.
Is there any strategy or tip that I can use?

Please, provide any information according your experience.
Thanks in advance

Peter

More important than dbspaces are the underlying disk structures. You need
to keep the rootdb space independent of the physical log, the physical log
independent of the logical logs, the logical logs independent of the high
access tables, high access tables apart from each other and from their
indexes, etc. If it's all on the same SAN array, it really doesn't matter
that much.

That said, if you are using dbspaces for fragmenting tables to take
advantage of more parallelism or less contention during massive updates,
then the more dbspaces the better up to the limits of the parallelism you
can muster on your machine. So if you only have a single 1GHZ CPU core you
won't get much parallel processing done compared to a 4GHZ 256 core
machine..

All in life and databases is balance. There are no hard and fast rules.
Your workload and equipment will be the deciding factors in determining the
optimum configuration. Personally, given my preferences, I want to
configure a machine with as many CPU cores as I can cram into the box and
attach as many small drive spindles as I can. Give me a machine with 128
cores and 256 disk drives and I'll build you a database box that will
SCREAM! Alas, I mostly get to deal with 3 cores and a SAN with a single
array built from 6 2TB drives that's being shared with a bunch of Windows7
filesystems.

Art

Art S. Kagel
Advanced DataTools (www.advancedatatools.com)
IIUG Board of Directors (art (AT) iiug (DOT) org)
Blog: http://informix-myview.blogspot.com/

Disclaimer: Please keep in mind that my own opinions are my own opinions and
do not reflect on my employer, Advanced DataTools, the IIUG, nor any other
organization with which I am associated either explicitly, implicitly, or by
inference. Neither do those opinions reflect those of other individuals
affiliated with any entity with which I am affiliated nor those of the
entities themselves.



On Wed, Nov 3, 2010 at 8:04 PM, Peternt <roger_vilca (AT) yahoo (DOT) es> wrote:

On 11/3/2010 8:23 PM, Art Kagel wrote:
And the screaming is of a different type?

JWC

Fewer spindles, less screaming.

Art

Art S. Kagel
Advanced DataTools (www.advancedatatools.com)
IIUG Board of Directors (art (AT) iiug (DOT) org)
Blog: http://informix-myview.blogspot.com/

Disclaimer: Please keep in mind that my own opinions are my own opinions and
do not reflect on my employer, Advanced DataTools, the IIUG, nor any other
organization with which I am associated either explicitly, implicitly, or by
inference. Neither do those opinions reflect those of other individuals
affiliated with any entity with which I am affiliated nor those of the
entities themselves.



On Mon, Nov 8, 2010 at 6:57 PM, John Carlson
<jwcarlson1 (AT) yahoo (DOT) com.invalid>wrote: