RE: QQRYDEGREE

A QQRYDEGREE setting  of *OPTIMIZE or *MAX will also allocate significantly
more resource to index rebuilds on the system. 

This can be a good thing or a not so good thing. In my experience we had an
application that created several work files when a user signed on in the
morning. These work files were preloaded with some data and logical views
were build over them. The original design point was to take the time to get
all this stuff ready before the user needed it so when they were talking to
a customer on the phone everything would be ready to go.... Several 100
users sign on at about 8AM each day. When I set the QQRYDEGREE system value
to *OPTIMIZE our CPU utilization climbed to 100% (over 8 processors!)
resulting in a resource queuing effect that increased response time
significantly during this time of the day... It wasn't pretty.

I now change the value to *IO during the day and at 5PM change it back to
*OPTIMIZE. CPU utilizations peaks aren't a problem in the morning anymore
and response time is much more consistent. 

We have also installed a tool that through policies, exit programs and
validity checking programs allow us to modify the "parallel processing
degree" using the CHGQRYA command for individual jobs as they use query and
SQL processes. This has worked out very well for us. These processes also
allow us to utilize the *NBRTASKS value for parallel processing degree
setting, which isn't an available option for the QQRYDEGREE setting. This
lets us allocate the number of parallel tasks the system should allow for
individual jobs.

The parallel processing degree can also be controlled for groups of jobs or
individual jobs via a setting in the QAQQINI (QUERY OPTIONS) file. 

 Kenneth

****************************************
Kenneth E. Graap
IBM Certified Specialist 
AS/400e Professional System Administrator
NW Natural (Gas Services)
keg@xxxxxxxxxxxxx
Phone: 503-226-4211 x5537
FAX:    603-849-0591
****************************************


-----Original Message-----
From: Vern Hamberg [mailto:vhamberg@xxxxxxxxxxxxxxxxxxxxxxxxx]
Sent: Friday, April 18, 2003 7:01 AM
To: Midrange Systems Technical Discussion
Subject: Re: QQRYDEGREE


Most of the benefit of SMP applies to SQL & other query-ish activity. It 
appears it can be used for CPYFRMIMPF, as well. But not if you need RRNs to 
be the same in the copy.

Native I/O is not assisted much, if at all, IIRC. Please correct me if this 
is not the case.

A good reference is the Performance Mnagement Resource Library at 
<http://www-1.ibm.com/servers/eserver/iseries/perfmgmt/resource.htm>. Also, 
the iSeries DB2 page at <http://www-1.ibm.com/servers/eserver/iseries/db2/>

HTH

Vern