I think that one of the most enjoyable comments I've read in an IBM
document was Kathy's comment in the White Paper: "A caution about shared
ICFs is in order. Because a function is provided does not mean an
installation should employ this function."
For that matter, Kathy only partially describes some of the pitfalls with
using shared ICFs. For example, LPAR CPU Busy can be significantly
understated and LPAR PR/SM overhead can be understated. While APAR II10549
also describes some of these effects (and attributes the cause to hardware
problems), the problems described by APAR II10549 apply equally to shared
ICFs.
I have data from pdbs sent to me by CPExpert users with shared ICFs that
show 12-14% understated LPAR busy time. Kathy did describe this effect,
but the implications were not clear to me until I started seeing large
negative numbers with SMF70PDT-SMF70EDT in my data (what a deal, 'eh? Use
shared ICF and you have negative overhead ;-)).
As Kathy described, RMF reports these negative numbers as "****" but that
happens only when the numbers are negative. It is not necessary that the
numbers be negative to be incorrect; RMF will dutifully report very small
positive numbers (e.g., 0.1%) for PR/SM overhead even though these very
small positive numbers are incorrect. When you see "****" for PR/SM
overhead in RMF reports, you can be certain that the reported CPU Busy is
incorrect for the LPAR.
The bottom line here is to be wary of using shared ICFs, even in a test
environment, and you probably should NOT use shared ICFs in a production
environment.
Regards,
******
Don Deese, Computer Management Sciences, Inc.
Voice: (703) 922-7027 Fax: (703) 922-7305
http://www.cpexpert.org
******
At 10:13 AM 2/6/2006, you wrote:
Re.
http://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/W99037
>And I thought 9672s were (soon to be?) unsupported....
Well, I was as intrigued as you are, John.
So I contacted Kathy Walsh of the WSC, who replied to me:
"There is no revision. The revision came about because of a "flaw" in the
database design of Techdocs. As you state the information is as
appropriate for z's as 9672."
Jan
--
No virus found in this outgoing message.
Checked by AVG Anti-Virus.
Version: 7.1.375 / Virus Database: 267.15.0/249 - Release Date: 2/2/2006
----------------------------------------------------------------------
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
