RE: Memory Performance - When is RAM over committed ?

Scott ...

Your suggesting is a great one! This is the kind of insight I was hoping
to get. There is a limitation with using SETOBJACC though... You can't
use it with the *INTERACT or *MACHINE pools ...

Don't you wish that WRKSHRPOOL had a column labeled: "% Utilized" ???

Work with Shared Pools

System:
S02
Main storage size (M) . : 31456.53



Type changes (if allowed), press Enter.



Defined Max Allocated Pool -Paging Option--

Pool Size (M) Active Size (M) ID Defined Current

*MACHINE 7483.55 +++++ 7483.55 1 *FIXED *FIXED

*BASE 3640.54 394 3640.54 2 *CALC *CALC

*INTERACT 10375.95 17 10375.95 6 *CALC *CALC

*SPOOL 6.96 5 6.96 8 *CALC *CALC

*SHRPOOL1 6474.13 37 6474.13 3 *CALC *CALC

*SHRPOOL2 31.45 8 31.45 4 *CALC *CALC

*SHRPOOL3 31.45 5 31.45 5 *CALC *CALC

*SHRPOOL4 3412.47 40 3412.47 7 *CALC *CALC


Kenneth
Kenneth E. Graap


-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx
[mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Ingvaldson, Scott
Sent: Monday, July 20, 2009 10:59 AM
To: Midrange Systems Technical Discussion
Subject: RE: Memory Performance - When is RAM over committed ?

The problem is the faulting rates only have "guidelines" which can
sometimes be hard to communicate to management. A hard number can be
easier to communicate.

Try using SETOBJACC to see exactly what amount of memory is free in any
pool at the current time:

SETOBJACC OBJ(FILE) OBJTYPE(*FILE) POOL(*BASE) or SETOBJACC OBJ(FILE)
OBJTYPE(*FILE) POOL(*SHRPOOL1)

Will return a message CPC1140: 189K of FILE brought to pool with
5509556K unused. The second number is the amount of free RAM in the
pool at the time. If that number is routinely zero or very small your
memory is overcommitted.

I've found that "real" numbers are much more persuasive than guidelines.

It would be fairly easy to put this into a batch program that ran at
regular time frames and create a memory usage chart similar to your CPU
usage charts. Remember that the performace adjuster will most likely be
changing the pool sizes continuously as well. Don't forget to purge the
object after you check the memory usage.


Regards,

Scott Ingvaldson
Senior IBM Support Specialist
Midwest Region Data Center
Fiserv.


-