RE: Interactive activity approaching capacity of installed featur e

[Jim Damato <jdamato@xxxxxxxxxxxxxxxxx>]

There's an good point in here somewhere (and it's not fast/400).  I'm always
one of the premier whiners about Interactive Feature, but the technical
solution here does not require the usual foot-stomping about the whole
concept.

The last time we went got the "Interactive activity approaching capacity of
installed feature" we just went in and monitored QINTER activity for a
while.  We found that programmers were running interactive queries, RPG
compiles, and CPYF.  Stopping these practices easily brought the situation
under control.  As Jim Franz suggested, you might also look for application
programs that process a lot of data online, or other batch-type work that
has not been submitted to batch.  Software packages sometimes can be
configured to give (or take away) the ability for users to choose online vs.
batch processing.

I did, however, find it interesting that the measure of Interactive Feature
is so arbitrary.  A few dozen users busily running well-architected OLTP
work won't run up the meter nearly as quickly as a couple of interactive
compiles or file copies.

-Jim

-----Original Message-----
From: Leif Svalgaard [mailto:leif@xxxxxxxx]
Sent: Wednesday, June 11, 2003 5:05 PM
To: Midrange Systems Technical Discussion
Subject: Re: Interactive activity approaching capacity of installed
feature


From: James Rich <james@xxxxxxxxxxx>
> On Wed, 11 Jun 2003, Andy Nolen-Parkhouse wrote:
> > There really are only two things to do.  First, you can find a way to
> > decrease the interactive workload by scheduling your work differently or
> > finding the user who is running batch jobs interactively.  If you have
> > programmers who do interactive compiles for instance.  The second option
> > is to purchase additional interactive capacity, which can be very
> > expensive.
> 
> There are more than two things to do, one is fast/400 and another is
> change platforms.  Those last two options may not be desireable, however.
> 

what an interesting thought, James, fast400 might help keep people on
the platform and thus be good for the AS/400 community at large.