Re: Pool size problem -- MIDRANGE-L

At 09:17 AM 6/27/00 , you wrote:

>We have Model 720 (420 cpw batch and 35 cpw interactive) with 2 GB memory 
>and using MQ Series .
>These messages are being gotten under subsystem that we had created. ( It 
>works at Base Pool )
>Under normal conditions CPU is at 35-40%. When Mq messages starts to 
>coming so much, batch jobs start to slow down and Mq messages start 
>to  gather at the queue.
>In order to prevent the system from this situation what should we do ?
>Do we need to create a new pool or change the base pool size ?
>Now the values are ;
>                              Work with Shared 
> Pools
>Main storage size (M)  . 
>:        2048,00
> 
 >
                Defined         Max     Allocated   Pool  -Paging 
Option--
>Pool         Size (M)       Active     Size 
>(M)     ID   Defined  Current
>*MACHINE   178,01      +++++      178,01       1   *FIXED   *FIXED 
 >
*BASE         329,95             55     329,95       2   *CALC    *CALC 

*INTERACT  1506,12            92    1506,12      4   *CALC    *CALC 

*SPOOL         33,89               5       33,89      3   *CALC    *CALC 

                                                                             

Thanks ...

There REALLY isn't enough information here to solve your problem!  We need 
subsystem descriptions and most importantly, a description of the workflow 
that is being processed by this server.

Some observations ... most of your main storage is allocated to your 
interactive pool (I am assuming that you are running batch in *BASE).  The 
CPW breakout of your machine reflects a HEAVY batch orientation.  There's a 
fundamental problem here!

Appropriate "rules of thumb" would include:
    - have jobs of the same type running in the same memory pool.
    - have separate subsystems for types of jobs that you wish to control 
separately

Now, you imply that you have batch & MQ series running in the same memory 
pool.  That means that you probably have massive page stealing going on 
between these jobs & the other miscellaneous tasks that are running in *BASE.

If batch is that important to you, it should rate it's own subsystem & a 
private memory pool.  The execution class used to start jobs in the pool 
should be set to *PURGE(no) and the activity level should EXACTLY match the 
number of batch jobs that you allow to run in the subsystem.

I don't know enough about your situation to comment on where to put the MQ 
series jobs.

THIS IS A SITUATION THAT IS WORTH THE $$$ TO ENGAGE A KNOWLEDGEABLE PERSON 
TO SPEND TWO DAYS TO FIX (one day to observe / change the configuration, & 
one day to confirm the results / make corrections / teach you what they did 
& why).

The AS/400 is a simple machine to get going, but performance tuning 
requires much more specialized skills.  You seem to have a high leverage 
situation ... don't "go cheap" on this issue!

For more information, read the AS/400 Work Management manual at:

http://publib.boulder.ibm.com:80/cgi-bin/bookmgr/BOOKS/QB3ALG03/CCONTENTS

Hope this helps!


John Myers
IBM Certified Specialist - AS/400 Technical Solutions
Strategic Business Systems, Inc.
300 Lake Street, Suite B, Ramsey, NJ 07446  USA
E-mail: mailto:jmyers@sbsusa.com   Phone: +1 (201) EASY 400   x131
Web:    http://www.sbsusa.com      Fax:   +1 (201) 327-6984

Get and route intelligence from your IBM AS/400 web site - WebSurvey/400
      http://www.WebSurvey400.com

Systems supporting the distribution operations of Motor Vehicle manufacturers
     http://www.VehicleSystem.com
+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---