Is there any advantage to spreading table fragments across distinct
dbspaces if bulk storage is managed via SAN?

Simply? Yes. More parallelism within the engine and most queries are
processed against memory anyway. ALso improved flush times up to the limits
of the array.

Art

Art S. Kagel
Advanced DataTools (www.advancedatatools.com)
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 Fri, Jul 22, 2011 at 2:03 PM, red_valsen <red_valsen (AT) yahoo (DOT) com> wrote:

On Jul 22, 3:02*pm, Art Kagel <art.ka... (AT) gmail (DOT) com> wrote:

Oh, yes. As far as parallelism is concerned, no difference.

Note, however, that dirty page flushes at checkpoint time happen with one
thread per chunk, so you have to balance the ease of a single dbspace with a
single huge chunk versus many chunks in one or more dbspaces.

The main downside to ganging all of the fragments of a large table in to a
single large dbspace is that as extents are added to the fragments they will
be interleaved between all of the fragments. Dedicating a dbspace to each
fragment will ensure that each new extent is contiguous with the existing
extent in that chunk and will be coalesced into a single larger extent which
means less head movement - potentially. Of course the uncertainlty of a SAN
environment limit how much this will actually improve overall performance,
but we can try.

Art

Art S. Kagel
Advanced DataTools (www.advancedatatools.com)
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 Fri, Jul 22, 2011 at 4:13 PM, red_valsen <red_valsen (AT) yahoo (DOT) com> wrote: