RE: Why separate pools?

[Alexei Pytel <pytel@xxxxxxxxxx>]

Jobs in any pool can page-in *some* pages of the same object - e.g. jobs 
reading different records of the same file, might bring in different parts 
of the same object to different pools.
The intended behaviour of SETOBJACC is to preload object to one pool.
To keep it from being paged out, one wants to keep all pages of this 
object in one pool. This is why you want to use SETOBJACC to purge all 
pages of the object from all other pools - before loading it in the 
designated pool.
By the way, if entire object does not fit in its designated pool during 
SETOBJACC, the remaining pages will be paged-in in a usual way by whoever 
touches them in whatever pool.

Anyway, everything I've said is just a direct implication of single level 
storage.

    Alexei Pytel
always speaking for myself only




"Joe Pluta" <joepluta@xxxxxxxxxxxxxxxxx> 
Sent by: midrange-l-bounces@xxxxxxxxxxxx
11/23/2004 04:17 PM
Please respond to
Midrange Systems Technical Discussion


To
"'Midrange Systems Technical Discussion'" <midrange-l@xxxxxxxxxxxx>
cc

Subject
RE: Why separate pools?






> From: Alexei Pytel
> 
> With all respect... This DOES work across memory pools.
> As soon as page of a vurtual memory is addressable by machine, it does
not
> matter, which pool it is in.

I don't know about that, Alexei.  If you read the documentation on
SETOBJACC, you specify the pool to bring the object into.  It seems you
can bring the same object into multiple pools.  *PURGE will purge the
object "from all pools".

Why would you bring an object into multiple pools if one pool can access
the same object loaded in a different pool?

Maybe an IBMer can help us.

Joe


--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing 
list
To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.