Thanks a lot, Barry. I read this part of note in MXG sourcelib,too. It is
very useful.
However, I'm a little bit confuced when I sum up CPITCBTM & CPISRBTM by date
for the 2 lpars (one is DDCONS=YES, the other is NO), the result is not what
I expected to see:
1. The lpar with DDCONS=NO, then total CPITCBTM+CPISRBTM approachs
0. because all the jobs should bypass DD consolidations.
But what I got is, still many many jobs spend a few seconds on
CPITCBTM. So, how to explain?
2. The lpar with DDCONS=YES, should have much bigger value for
CPITCBTM+CPISRBTM than the other lpar.
But in fact the difference varies from 100 seconds to 350 seconds per
day.
regarding the EMPTYEXCPSEC(SUPPRESS) , we're still in 1.9.
On Mon, Dec 21, 2009 at 5:40 AM, Barry Merrill <[email protected]> wrote:
> In 1987, Diane Eppestine at Southwestern Bell saw that whenever their
> SAR job (a SYSOUT processing subsystem) was cancelled, the CPU went to
> 100% busy for 30-60 minutes. Instruction traces found the "loop" was in
> DD Consolidation. SAR dynamically allocates a DD for each SYSOUT file
> it processes; by the end of the week that step had over 75,000 DD
> entries! DD consolidation reads the first DD segment, scans the
> remaining 74,999 segments for a match, reads the second and scans the
> remaining 74,998 for a match, etc. etc., etc., all at DPRTY=FE! In
> response to Diane's discovery, Bill Richardson, IBM SMF Development,
> subsequently provided a new SMF option, DDCONS(NO), specified in
> SYS1.PARMLIB(SMFPRMxx), so that you can disable this very unwise (in my
> opinion) algorithm, and thereby eliminate its wasted CPITCBTM and CPISRBTM
> (MXG names for SMF30ICU and SMF30ISB, the "Initiator" CPU time that
> occurs before Step Program Load and after Step Program Termination).
>
> All of those empty SYSOUT DD segments can now be removed in z/OS 1.10
> with the EMPTYEXCPSEC(SUPPRESS) option:
>
>
>
> 1. APAR Identifier OA29582 adds new EMPTYEXCPSEC(SUPPRESS) option in
> SYS1.PARMLIB to suppress all empty EXCP entries.
>
> The option prevents the creation of null segments in SMF 30 records
> for SMS Candidate Volumes, and could significantly reduce the size
> of your step and job termination records, especially if your site
> has a default of (SYSDA,5) for every allocation!!
>
> The MVS Initialization and Tuning Reference, under the SMFPRMxx
> parmlib parameter definitions, EMPTYEXCPSEC:
> SUPRESS specifies that the system suppress the creation of empty
> EXCP sections. Empty sections can be the result of non-dataset
> allocations, such as DD DUMMY, or for spool file allocations
> (i.e., SYSIN, SYSOUT JES DDs), or for non-allocated candidate
> volumes in the SMS storage group.
>
> One MICS site reported a 28% reduction in CPU time with removal of
> all of their empty EXCP segments.
>
> New EMPTYEXCPSEC option in PARMLIB is z/OS 1.10 only.
> While EMPTYEXCPSEC option is documented in the z/OS 1.9 SMF
> manual in Section 13.34.2.7 (Execute Channel Program (EXCP)
> Section) of z/OS MVS System Management Facilities (SMF) Document
> SA22-7630-16, for z/OS 1.9, testing on a z/OS 1.9 system
> resulted in
>
> IEE945I UNRECOGNIZABLE OPTION 'EMPTYEXCPSEC' IN PARMLIB INPUT
> IEE947I '/* DEFAULT RETRY */
> EMPTYEXCPSEC(SUPPRESS)' SKIPPED DUE TO PREVIOUS ERROR
>
> IBM has confirmed that the option only exists in z/OS 1.10, where
> it is listed in the Release Guide as a 1.10 enhancement
>
>
>
> Barry Merrill
>
> ----------------------------------------------------------------------
> 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
>
--
Cobe Xu
Best Regards
-----------------------------------------------------------
zOS Performance & Capacity Analyst
E2E Performance Analyst
Email: [email protected]
-----------------------------------------------------------
----------------------------------------------------------------------
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
