Howdy Tim,
We run all that here. Not sure what you're looking for but I can
tell you how we use WLM here in conjunction with them. As you can
imagine there's all sorts of schools of thoughts that work. Take this as
it works here but not maybe not somewhere else.
Importance of 1, execution velocity of 40 for ADABAS, GBP
Importance of 1, execution velocity of 17 for COMPLETE, ETB ( broker )
Importance of 1, execution velocity of 8 for RPC servers
Resource group based on SU's - Day - 5k, Night - 7k for ADABAS, GBP
Resource group based on SU's - 4k COMPLETE, ETB
Resource group based on SU's - 3k
On 2/17/2017 2:57 PM, Neil Duffee wrote:
Caveat: list digestion leads to delayed responses...
I'm surprised Lizette hasn't piped up already but I'll pass along that you
would probably get a better response via sa...@listserv.uark.edu which is where
the SAG clients hang out. I've cc:'d them in this message.
Since we were DB2, Adabas (for Predict) was not a specialized WLM case for us.
Perhaps others have done more detailed work. We were using Transaction Goals for
CICS especially to weed out the long-running transactions ie. OmegaMon, & Cics
system tasks, so they didn't pollute the response times collection. They were
given 5-10% CPU goals but got carried along by the Transaction Goals since CICS is
adjusted by WLM as an entire region. I seem to recall there was some discussion in
the SysProg WLM Guide Redbook to this effect. [1] I re-read the guide many times
and found new insights [2] even after a few years had elapsed.
[1] I created a 'medium' Goal with more relaxed response times intended for
the lessor-preferred transactions but we never got around to making any
distinction.
[2] The biggest being that the goals should not be specified as 'What you want'. Instead, they should be
more, "when the stuff starts flying, I'd like these (acceptable) minimums, plz." For example,
instead of "I'd like this to get 80% CPU", you should say, "if things are bad, this could
survive on 40% CPU. I won't like it but we'll stay processing." If you're not constrained, WLM will
let you over-perform but if you artificially constrain it, it can't steal from other processes to smooth
things out. I especially like how WLM will temporarily boost a low-goal process that's holding
up/enqueing/locking a higher process. It shows up in the RMF report as below:
--PROMOTED--
BLK0.000
ENQ0.502
CRM0.000
LCK 230.133
SUP0.000
> signature = 8 lines follows <
Neil Duffee, Joe Sysprog, uOttawa, Ottawa, Ont, Canada
telephone:1 613 562 5800 x4585 fax:1 613 562 5161
mailto:NDuffee of uOttawa.ca http:/ /aix1.uOttawa.ca/ ~nduffee
"How *do* you plan for something like that?" Guardian Bob, Reboot
"For every action, there is an equal and opposite criticism."
"Systems Programming: Guilty, until proven innocent" John Norgauer 2004
"Schrodinger's backup: The condition of any backup is unknown until a restore is
attempted." John McKown 2015
-Original Message-
From: Tim Brown [mailto:tbr...@cenhud.com]
Sent: February 15, 2017 19:58
Subject: Wlm and adabas\complete
Does anyone have experience with Adabas 8.3.3 and Sags complete product (Cics
like) And wlm ?
--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
--
Brian W. France
Systems Administrator (Mainframe)
Pennsylvania State University
Administrative Information Services - Infrastructure/SYSARC
Rm 25 Shields Bldg., University Park, Pa. 16802
814-863-4739
b...@psu.edu
"To make an apple pie from scratch, you must first invent the universe."
Carl Sagan
--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
