I would tinker with the policy but not use resource groups. I would only use 
resource groups to cap unruly tasks not guarantee. Are you running WLM Managed 
Initiators? Do you have control over the initiators and batch work or are 
people resetting service classes and starting inits along the way? Have you 
tried to tune the policy to your problem time frame? If that problem batch work 
has unattainable goals I'd be hard pressed to see how it will get any CPU.
   
  I ended up tuning to our calendar month end process that could hit mid week 
and run all the next day. I used WLM Managed Initiators but used time of day 
rules to limit concurrence by job class based on requirements. A UK piece had 
to start early, 14:00, on the last day of the month, EST. Then the rest of the 
batch, 2 different cycles, would kick at around 18:00. My bottom feeders would 
see about 85% wait on CPU during prime time the next day but still get some 
service. I ended up doing a multi-period imp 4/5 service class setup with low 
velocities for this. The developers knew that during this time frame long 
runners would be just that or real long runners. 
   
  Or maybe consider multi period batch with imp's 3, 4, and 5. Three to get 
them started and run a bit, 4 throttle down and 5 just let them run all day. 
   
  "McKown, John" <[EMAIL PROTECTED]> wrote:
  > -----Original Message-----
> From: IBM Mainframe Discussion List 
> [mailto:[EMAIL PROTECTED] On Behalf Of Tom Schmidt
> Sent: Monday, December 17, 2007 11:36 AM
> To: IBM-MAIN@BAMA.UA.EDU
> Subject: Re: WLM question.
> 



> 
> Sorry to see your sysplex disappear over this... ;) 
> 
> -- 
> Tom Schmidt 

The parallel sysplex disappeared when nobody wanted to pony up for the
cost of a CFL. The basic sysplex is now endangered due to the CPU cost
of running two z/OS images for no apparent benefit over "tinkering" with
WLM. I don't really mind much. I was not convinced that any type of
sysplex, in our environment, was all that much of a positive. There are
some positives for tech services in the area of better testing before
production. There really aren't any for applications or production
control that I can see.

--
John McKown
Senior Systems Programmer
HealthMarkets
Keeping the Promise of Affordable Coverage
Administrative Services Group
Information Technology


----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Reply via email to