* copyright (C) 1984-2019 merrill consultants dallas texas usa

Thanks to Brian Cummings, Federal Reserve Information Technology USA

Thanks to Fritz Zeigler, Federal Reserve Information Technology, USA
====== Changes thru 27.229 were in MXG 27.08 dated Sep 1, 2009========
Change 27.229 Support for z/OS 1.11, COMPATIBLE.

EXTY8224 -TYPE 0. New variable:

EXTY8225 CVTTZ ='CVTTZ*TIME DIFFERENCE*LOCAL*TO GMT'

EXTY8226 -TYPE1415. NO NEW DATA, but lots of text inserts about

IMAC82 assembling the IFGSMF14 Macro with DSECT=YES, which is

VMAC0 why it is flagged in the SMF Manual as changed.

VMAC82 -TYPE30. NO NEW DATA, only one line of text updated.

VMXGINIT -TYPE82. New subtypes 24, 25, and 26.

May 2, 2009 -TYPE92. Subtype 15 now documented, but MXG supported

Sep 2, 2009 was added in Change 26.277 for APAR OA24208.

-TYPE1415 New variables DCBEEX31 XTIOTYES created from

SMF14FLGS.

-Dec 3, 2009: CHANGE 27.325 is required for z/OS 1.11.
CHANGE 27.228 Support for APAR OA30197, adds SMF30ASI, new MXG variable

VMAC30 ASID in ID Segment of all SMF 30 records. Kept in the

Sep 1, 2009 TYPE30 ASID-level datasets (_1,_4,_5,_V,_6).
CHANGE 27.227 Observations were not output because DO loop logic was

VMACINSY incorrect for the INSYAUID='D' records. Variable INSYATA

Sep 1, 2009 is now converted to EBCDIC text.

Thanks to Mark Cohen, DTS, ITALY.

CHANGE 27.226 Dataset T112DTCO, DATACOM DETAIL, variables were invalid

VMAC112. due to misalignment. Dataset T112MQCT variable MATTACH

Aug 31, 2009 and dataset T112DTCT variable DATTACH were INPUT $CHAR8,

but they are now seen to be TODSTAMP values, so they are

changed from CHAR to NUM and formatted DATETIME25.6, and

are observed to be slightly prior to the UOWTIME.

Thanks to Richard Schwartz, State Street Bank, USA.
CHANGE 27.225 Cosmetic. Long messages (listing all requested reports)

ANALRMFR are split and shifted left for readability, and the list

Aug 31, 2009 of valid INTERVAL= values was updated.
CHANGE 27.224 Eight new IBM fields starting with R7452xxx were named in

VMAC74 MXG as R7451xxx, accidentally, as all of the other fields

Aug 31, 2009 in that segment are IBM-named R7451xxx. But these eight

R7452XTY R7452XFL R7452PRO R7452PWO R7452PBR

R7452PBW R7452PRT R7452PWT

are all a second field name for the same field location,

and apparently IBM decided to note that difference by

naming them with a 2 vs a 1. In any event, to match the

SMF Manual field names in the MXG Variable names, those

eight are now named with a 2 instead of a 1.

Thanks to Jim Horne, Lowe's Companies, USA.
CHANGE 27.223 New "TOP" DB2 Report ranks resource usage by correlation

ANALDB2T id and authid, for these resource metrics:

Aug 30, 2009 CLASS*1 CPU*RANK

CLASS*2 CPU*RANK

ELAPSED*RANK

GET*PAGE*RANK

TOTAL*GET*PAGES

Thanks to Chuck Hopf, Independent Consultant, USA.

CHANGE 27.222 Support for NDM-CDI NDMRTYPE 'SC' and 'S2' records. The

FORMATS two datasets NDMSC and NDMS2 were already created, but

VMACNDM they only contained the header variables until now.

Aug 30, 2009

Thanks to Robert S. Zimmerman, HMS, USA.
CHANGE 27.221 Support for zOSEM Operating System Environment Manager

EXOSEMCM from Trident Services user SMF record creates 3 datasets:

EXOSEMRE

EXOSEMWT DATASET DATASET DATASET RECORD

FORMATS SUFFIX NAME LABEL SUBTYPE

IMACOSEM

TYPEOSEM OSEMCM OSEMCMD OSEM FEMCNTL CMD 1

TYPSOSEM OSEMWT OSEMWTO OSEM FEMWTO CMD 2

VMACOSEM OSEMRE OSEMRES OSEM JES2 EXIT44 3

VMXGINIT

Aug 30, 2009

Thanks to Larry Stahl, IBM Global Services, USA.

CHANGE 27.220 MXG 27.06-27.07. %ANALCNCR failed if INDATA= specified

ANALCNCR more than one dataset. "Cosmetic" protection was added

BUIL3005 by Change 27.165 to verify that the input dataset exists,

BUILD005 to avoid DATASET NOT FOUND errors, but it only parsed the

MNTH72 1st INDATA (to get DDNAME and DATASET to pass to VGETOBS,

TRND72 as PROC SQL which requires separate arguments). Now,

Aug 29, 2009 each token in the INDATA= argument is parsed & verified.

If all datasets have zero observations, an MXGWARN is

produced at the beginning and if there are zero obs in

the output dataset another MXGWARN will appear.

If all datasets in the INDATA= do not exist, %ANALCNCR

will stop processing, but if one or more do exist, the

output dataset will be created and populated if possible,

but MXGWARN messages will identify which INDATA= datasets

did not exist. Prior to Change 27.165, multiple datasets

were supported because the only INDATA reference was the

SET &INDATA; statement which read all of the input, but

that statement failed if any input dataset did not exist.

-The only MXG member that invokes %ANALCNCR with multiple

INDATA= datasets is ANALHTML, which exposed this error,

with its INDATA=PDB.STEPS SPIN.SPIN30_4, statement.

-In testing the new %ANALCNCR with ANALHTML, SAS printed

warnings that MSOUNITS had different lengths in those two

datasets. Investigation found MSOUNITS & SERVICE weren't

always stored in 8 bytes, so updates were made to members

BUILD005 BUIL3005 MNTH72 and TRND72 so that all instances

of variables MSOUNITS and SERVICE are now LENGTH 8.

-All INTERVAL= arguments that specify an interval token

with a fixed-length interval duration are supported.

TIMER is set to the interval and SYNCINTV=YES is used

(but that could be overridden in your %ANALCNCR).

Thanks to Tony Curry, BMC, USA.

CHANGE 27.219 TYPE11GD dataset variables CTGIAVRS and CTILAVRS were

VMAC111 reversed in their INPUT locations.

Aug 28, 2009

Thanks to Jim Poletti, Edward Jones, USA.

Thanks to Gordon E. Griffith, Edward Jones, USA.
CHANGE 27.218 Variable SMF70LAC in PDB.RMFINTRV was the maximum value

VMXGRMFI during the sub-intervals that were summarized; it is now

Aug 28, 2009 the average value, so it matches PDB.ASUMCELP values.

However, the RMF reports may print that Maximum value of

any inputted sub-interval.

Thanks to Karl Lasecki, Chemical Abstracts Service, USA.

CHANGE 27.217 PDB.TYPE70PR with SMF70CIN blank but SMF70ONT nonzero was

VMAC7072 created if the last LPAR was offline (LPARCPUX=0). That

Aug 28, 2009 caused NRCIXGT0 to be a missing value, so the second test

IF NRCIXGT0 GT 0 THEN ... was not executed. That second

test is now redundant, as SMF70CIX always exists now, so

that IF ... END was removed. Error was introduced in

MXG 27.03 by Change 27.075.

Thanks to Rudolf Sauer, T-Systems, GERMANY.

CHANGE 27.216 Change 27.126 accidentally removed SYTLPNAM from the KEEP

VMACXAM list for XAMSYT dataset, and left it in the KEEP list for

Aug 28, 2009 the XAMSYU dataset.

Thanks to Douglas C. Walter, Citigroup, USA.

Thanks to Tony Curry, BMC, USA.
CHANGE 27.215 TMON/CICS Version 3.1 INPUT STATEMENT EXCEEDED RECORD

VMACTMO2 error with 'TI' (Statistics Interval) record; the segment

Aug 26, 2009 of data starting with TIMQSOPC exist in 3.1, but MXG only

input those fields for TMMDVREL GE 3.2; the test before

that INPUT statement was changed from 3.2 to 3.1.

Thanks to Dianne Dunklau, Kohls, USA.

CHANGE 27.214 Summarized RMFINTRV/ASUM70PR output datasets could have

VMXGDUR the STARTIME/SMF70GIE values early by one INTERVAL unit

VMXGRMFI (e.g., data STARTIME midnight, INTERVAL=HALFHOUR caused

VMXG70PR output STARTIME of 23:30) after Change 27.178 used the

Aug 28, 2009 SMF70GIE vice STARTIME (because of unstable STARTIMEs).

Note Jan 21, 2010: "Unstable as in exactly one second

Nov 10, 2009 later than the expected Start of Interval."

ASUM70PR -A new argument FLORCEIL=FLOOR/CEIL is added to VMXGDUR

ASUMCAPT with FLOOR (unchanged) as the default, and the %VMXGDUR

ASUMMIPS calls in VMXGRMFI and VMXG70PR that interval the

ASUMSMFI SMF70GIE/SYNCTIME end times now specify FLORCEIL=CEIL.

No user had reported/noticed this error (yet!).

The rest of this text was REVISED when CHANGE 27.308

reset the SYNC59=NO default back in place in MXG 27.10:

-For summarizing INTERVAL data ONLY (RMFINTRV, SMFINTRV,

CICINTRV, etc datasets with existing interval start/end):

If you still specify SYNC(59) to write Interval data in

minutes 59,14,29,44, the SYNC59=YES parameter can be used

to change the DATETIME to "pretty" 00,15,30,45 startimes.

This is needed when SYNC(59) is in SYS1.PARMLIB members

SMFPRMxx or ERBRMFxx in SYS1.PARMLIB.
Using the FLORCEIL=FLOOR to calculate the interval start,

with SYNC(59) data and INTERVAL=HOUR, for time 19:59:

SYNC59=YES adds one minute to DATETIME, so 19:59 becomes

20:00, and the FLOOR(DATETIME) sets 20:00 as the Start.

This is correct, hour 20 for the 19:59-20:58 interval.

SYNC59=NO leaves the DATETIME as is, so FLOOR(DATETIME)

is 19:00 for the Start, WHICH IS THE WRONG HOUR!!!

SYNC(59) data REQUIRED SYNC59=YES for correct output.

Using the FLORCEIL=FLOOR to calculate the interval start,

consider SYNC(00) data and INTERVAL=HOUR, for time 20:00:

SYNC59=YES adds one minute to DATETIME, so 20:00 becomes

20:01, but the FLOOR(DATETIME) sets 20:00 as the Start,

which is what you wanted, even though the data is (00).

SYNC59=NO leaves the DATETIME as is, so FLOOR(DATETIME)

sets 20:00 as the Start, WHICH IS ALSO WHAT YOU WANTED.

But, using the FLORCEIL=CEIL to calculate interval end,

SYNC59 must NOT be used. With datetime 19:59 or 20:00,

the CEIL(DATETIME) returns 20:00 with SYNC59=NO. But if

SYNC59=YES added a minute here, then 20:00 would be 20:01

and CEIL(DATETIME) would then (INCORRECTLY) be 21:00.

Therefore, when using FLORCEIL=CEIL, VMXGDUR does NOT

change the value in DATETIME before the CEIL(DATETIME),

even if SYNC59=YES was specified.

Only ASUM70PR/VMXG70PR and RMFINTRV/VMXGRMFI currently

use the FLORCEIL=CEIL option to group by End Interval,

but only internally; the final sorts are BY STARTIME.

-These members were changed from SYNC59=NO to SYNC59=YES:

ASUM70PR ASUMCAPT ASUMMIPS ASUMSMFI VMXG70PR VMXGRMFI

-THE PRECEDING IS ONLY FOR SUMMARIZING INTERVAL DATA.
-FOR DETAIL TRANSACTIONS, SYNC59=NO MUST ALWAYS BE USED.

MXG 27.08 CHANGED TO SYNC59=YES IN VMXGDUR.

(This caused PDB.CICS transaction's HOUR to change.)

MXG 27.10 RESTORED THE SYNC59=NO DEFAULT IN VMXGDUR.

CHANGE 27.213 HFS and zFS EXCPs are counted in EXCPTOTL and EXCPNODD

BUILD005 (see MXG Newsletter FIFTY-FIVE, MVS Technical Note 1)

BUIL3005 but there are no DD Segment EXCP counts for HFS/zFS, so

VMAC30 you don't know how much of EXCPNODD is due to HFS/ZFS, so

Aug 21, 2009 you can't subtract the HFS/ZFS EXCPS from EXCPTOTL.

Sep 1, 2009 This change creates four new variables in TYPE30_V and

TYPE30_4 datasets:

USSEXCPS='USS-OMVS*TOTAL*EXCPS'

USSBYTES='USS-OMVS*TOTAL*BYTES'

USSCALLS='USS-OMVS*TOTAL*CALLS'

USSCPUTM='USS-OMVS*SYSCALLS*CPU*TIME'

with the sum of those counts from all OMVS segments in

the interval and step termination SMF 30 records.

Those variables are also now summed in PDB.STEPS and

PDB.JOBS when BUILDPDB is used to process SMF 30s.

-In testing this change, subtype 5 records for BPXAS STC

have all resource variables with zero value, even STEPNR,

and some zeros cause missing value calculations. Logic

was revised to prevent those harmless but confusing notes

on the SAS log for those records.

See MXG Newsletter FIFTY-FIVE, MXG Technical Note titled

1. Summary: "EXCP" counts recorded for access to HFS ....

CHANGE 27.212 Support for APAR OA30006 for DCOLLECT, adds 8-byte fields

VMACDCOL at the end of the record for the dataset storage sizes in

Aug 21, 2009 DCOLDSET dataset. The new fields replace the value in

existing variables DCDALLSP DCDUSESP DCDSCALL DCDNMBLK

when the new fields are populated.
CHANGE 27.211 Support for new-in-z/OS 1.10 DFSORT variables added to

VMAC16 existing TYPE16 dataset:

Aug 21, 2009 ICEINMRG='INTERMEDIATE*MERGES*PERFORMED'

ICEMNFLG='SIZE=MAX*IN*EFFECT?'

ICEMNVLX='SPECIFIED*OR DEFAULT*STORAGE'

ICEMNVLY='THEORETICAL*STORAGE*AVAILABLE'

ICEMNVLZ='AVAILABLE*STORAGE'

Thanks to Scott Barry, SBBWorks, Inc, USA.

CHANGE 27.210 A harmless Division By Zero when creating PDB.DB2STATS

VMACDB2 occurred in calculating variable THRDFTPT (to mimic the

Aug 20, 2009 IBM MENU2 report), added by Change 27.097, if variables

QW0225AT and QDSTCNAT were both zero. Divide protected.

Thanks to Cletus McGee, Alfa Mutual Insurance Company, USA.
CHANGE 27.209 TMON/DB2 BF record's SQL text is not quite as documented.

VMACTMDB BF0142LL, the "length" is 52, indicating there should be

AUG 19, 2009 50 bytes of text, but the next two bytes are INPUT in

new MXG variable BF0142RL, which is the original (real)

length of the original SQL text, and then, instead of the

documented 50 bytes of sql text, the record contains only

the first 48 bytes of text, kept in BF0142TT. BF0142LL

is then reduced by 2 to a value of 50 in TMDBBF dataset.

Thanks to Ernie Amador, UC Davis Health System, USA.
CHANGE 27.208 Cosmetic. Calculation of AVGWKSET caused notes on the log

VMAC30 "Missing Values were generated" if MULTIDD='Y', because

AUG 19, 2009 those records contain ONLY DD segments. Those notes are

now eliminated by only calculating for "real" records.

CHANGE 27.207 MXG 27.07. Change 27.174 stopped deaccumulation of the

VMACDB2 PDB.DB2STATS vars QSSTCONT and QSSTCRIT, incorrectly, as

Aug 18, 2009 both are accumulated and must be deaccumulated.

Thanks to Bill McDonald, Kimberly-Clark, USA.

CHANGE 27.206 The LABEL for variables STEPNAME and PROCSTEP in VMAC42

VMAC42 were the variable name, unlike all of the other labels

Aug 18, 2009 for those common variables; if you added processing of 42

records as the last physical VMAC in your BUILDPDB, those

STEPNAME='STEPNAME' and PROCSTEP='PROCSTEP' labels were

used for all BUILDPDB-created datasets. VMAC42 now has

STEPNAME='STEP*NAME' & PROCSTEP='PROCEDURE*STEP*NAME'

Thanks to Bret Hoesly, Telephone & Data Systems, Inc, USA.

CHANGE 27.205 Documentation. PDB.TYPE70PR observations for ICF engines

VMAC7072 with the LCPUPDTM (CPU Dispatch Duration) slightly larger

Aug 18, 2009 than the SMF70ONT (Online Duration) have been observed.

The maximum difference in an LPAR engine was 52 millisec,

0.052 seconds, but in that same interval, the PHYSICAL

ICF engine recorded 0.026 seconds, so the total ICFACTTM

for that interval was 78 milliseconds larger than the

online duration of exactly 900 seconds for that interval.

But some intervals do not always have exactly 900 seconds

online; in 26 intervals, 13 were exactly 900 seconds, 6

were shorter, by as much as 0.018 seconds, and 7 were

longer, by as much as 0.018 seconds. This was z/OS 1.9

on a 2094-720 with 1 ICF, 2 IFLs, and 17 CPs. The ICF

Percent Busy was 100.00008 with that largest difference;

the only circumvention is to test the PCT value and reset

it to 100% if greater than 100 (but, if greater than 101%

print a warning message and investigate further).

Thanks to Nicholas Ward, Centrelink, AUSTRALIA

CHANGE 27.204 Change 27.154 did not protect when the offset to the Path

VMAC92 Section was non-zero but there was no Path Section; it

Aug 17, 2009 caused an INPUT STATEMENT EXCEEDED RECORD LENGTH error.

Thanks to Dan Almagro, Automobile Club of Southern California, USA.

CHANGE 27.203 MXG 27.06-MXG 27.07, PDB.ASUMTALO had too few obs created

ANALCNCR due to an error introduced in ANALCNCR in Change 27.165.

Aug 17, 2009 The ANALCNCR error ONLY occurs if the TIMERNGE= argument

ASUMTALO was used; ASUMTALO is the only MXG member that invokes

%ANALCNCR with a TIMERNGE= argument. Change 27.165 added

elimination of character-to-numeric conversion messages,

but four typo's in new code stored &LOTIME into &HITIME.

Only ANALCNCR was changed in this change; ASUMTALO is

listed only to document it is impacted by this error.

-You should check your tailoring/report source libraries

to see if %ANALCNCR is used with the TIMERNGE= argument.

This ANALCNCR "utility" is inconsistently named, since it

creates an output dataset, rather than an analysis; it

is used to count the number of concurrent events across

time (like tape drives allocated each hour in ASUMTALO).

Thanks to Jon Whitcomb, Great Lakes Educational Loan Service, USA.

CHANGE 27.202 MXG 27.07. Debugging PUT statement at line 757 was

VMAC116 removed.

Aug 14, 2009

Thanks to Alfred Holtz, Medco Health Solutions, USA.

CHANGE 27.201 More DB2 Updates.

ANALDB2R

READDB2

Aug 12, 2009

CHANGE 27.200 Support for CICS/TS 4.1 GA updates/documentation notes:

EXCICMLR -Format MG110MP updated for new TCB Pool values of XPLINK,

EXCICRLR SSL, and THRD, new formats MGCICAT, MGCICRT created.

EXCICSJS -Dataset CICECC new variables:

EXCICW2R ECCCANAM ECCCAPOI ECCCAPTY ECCEBNAM ECCEVCAP ECCEVNAM

FORMATS -Dataset CICECG new variables:

IMAC110 ECGFLFAI ECGCAPFA ECGEVLCO ECGEVLOT

VMAC110 -Dataset CICECR new variables:

VMXGINIT ECRCHGAG ECRCHGDT ECRCHGUS ECREBNAM ECRGRFRM ECRINSAG

Aug 16, 2009 ECRINSDT ECRINSUS

-Dataset CICEPR new variables:

EPGSYNBK EPGEVLCO EPGEVLOT

-New Statistics Datasets created:
dddddd dataset description stid
CICRLR CICRLR CICS BUNDLE 100

CICW2R CICW2R CICS ATOMSERVICE 110

CICMLR CICMLR CICS XMLTRANSFORM 113

CICSJS CICSJS CICS JVMSERVER STATISTICS (RESID 116

Note: STID=79, STIMNR is not a Statistics DSECT, but is

a SUBTYPE=1,MNSEGCL=5 Transaction Record; support

for which was added in MXG 20.20, Change 20.200.
STID=84, STIMNT is never created in SMF records;

it is an online interface only.

-Change 27.032 noted there are 3 new TCBs in CICS/TS 4.1,

T8, EP, and TP, but only T8CPUT is recorded in CICSTRAN.

The EP TCB for Event Processing is explicitly NOT in the

CICSTRAN data, because IBM states that processing of the

new Business Event is NOT to be chargeable to users.

A TP TCB is created for every JVMSERVER resource def that

is installed and enabled, but I have not found a reason

why they are not captured in CICSTRAN.

-DOCUMENTATION. CICS Statistics datasets CICDS, CICDSPOO

have many variables starting with DSG,DS1-DS9,DSA-DSN,

and their labels were not consistent. This list is now

added at the bottom of VMAC110, and the labels revised:

DOCUMENTATION OF THE VARIABLES STARTING WITH DS.
TWO DATASETS CONTAIN VARIABLES THAT START WITH DS:

DATASET CICDS - DISPATCHER STATISTICS:

- "INTERVAL" VARIABLES STARTING WITH DSGXXXXX

- "PER TCB" VARIABLES STARTING WITH DSG DS1-DS9 DSA-DSN

DATASET CICDSPOO - TCB POOL STATISTICS:

- "PER POOL" VARIABLES STARTING WITH DSG DS1-DS5

DATASET CICDS - DISPATCHER STATISTICS:

- "INTERVAL" VARIABLES STARTING WITH DSG:

DSGNTCB DSGCNT DSGEJST DSGICVRT DSGICVSD DSGICVT

DSGMBTCH DSGPNT DSGPRIAG DSGPTCB DSGSRBT DSGSTART

DSGLSTRT DSGSTSKS DSGXSCNN DSGXSCNS DSGXTCBD
- "INTERVAL" VARIABLES ALWAYS MISSING VALUES:

DSGAMXTC DSGAMXTL DSGAMXTP DSGLRT DSGMAXOP DSGTAMXT

DSGTL DSGTOTWT

DATASET CICDS - DISPATCHER STATISTICS:

- "PER TCB" VARIABLES STARTING WITH DSG DS1-DS9 DSA-DSN

(The TCB NUMBER HAS CHANGED BETWEEN CICS RELEASES)

3.2 4.1 TCB MXG VAR

NUM NUM NAME PREFIX DESCRIPTION

-- H8 DSC H8 - NOT IN CICS/TS 3.1+

1 QR DSG* QUASI REENTRANT MODE

2 RO DS2* RESOURCE OWNING MODE

3 CO DS3* CONCURRENT MODE

4 SZ DS4* SECONDARY LU MODE

5 RP DS5* ONC/RPC MODE

6 FO DS6 FILE OWNING MODE

7 SL DS7 SOCKETS OWNING MODE (SL)

8 SO DS8 SOCKETS OWNING MODE (SO)

9 SP DSH SOCKETS PTHREAD OWNING MODE (SP)

10 EP DSM EVENT PROCESSING MODE

10 12 D2 DSD D2 - DB2 MODE

11 13 JM DSE JM - JVM CLASS CACHE MODE

11 TP DSN TP - THREADED TCB OWNING MODE

12 14 S8 DSB S8 - SOCKETS (SSL) MODE

13 15 L8 DSA L8 - OPEN MODE CICS

14 16 L9 DSI L9 - OPEN MODE USER

15 17 J8 DS9 J8 - OPEN MODE CICS

16 18 J9 DSF J9 - OPEN MODE USER

17 19 X8 DSJ X8 - OPEN MODE CICS

18 20 X9 DSK X9 - OPEN MODE USER

19 21 T8 DSL T8 - JVM MULTITHREADED

THE ABOVE TABLE MAPS EACH TCB NAME TO ITS VARIABLE PREFIX.

THESE "PER TCB" VARIABLES ARE POPULATED:

DS.ACT DS.NTCBA DS.PERCT DS.START DS.SYSW DS.TCBAF

DS.TCBAL DS.TCBCA DS.TCBCU DS.TCBDO DS.TCBDS DS.TCBDU

DS.TCBDX DS.TCBMD DS.TCBMM DS.TCBMP DS.TCBNM DS.TCBPA

DS.TCBPU DS.TCBST DS.TCT DS.TDT DS.TWT

THESE "PER TCB" VARIABLES ALWAYS HAVE MISSING VALUES:

DS.PERCT DS.TCBF1 DS.TCBF2 DS.TCBF3

DS.TCBF4 DS.SCT DS.SWT
DATASET CICDSPOO - TCB POOL STATISTICS:

- "PER POOL" VARIABLES STARTING WITH DSG DS1-DS5

DSG OPEN POOL (CONSISTING OF L8 AND L9 TCBS)

DS2 JVM POOL (CONSISTING OF J8 AND J9 TCBS)

DS3 XP POOL (CONSISTING OF X8 AND X9 TCBS)

DS4 SSL POOL (CONSISTING OF S8 TCBS)

DS5 THRD POOL (CONSISTING OF T8 TCBS)
J8 L8 S8 T8 AND X8 TCBS ARE CICS TCBS.

J9 L9 AND X9 TCBS ARE USER TCBS.

"TCB POOL" VARIABLES HAVE DSG,DS1,DS2,DS3,DS4,DS5 PREFIX:
DS.CMMWS DS.CMMWT DS.CNUAT DS.CNUUS DS.CURNW DS.CURWT

DS.MMWTM DS.MMWTS DS.MXTCB DS.NTCBL DS.PEANW DS.PMWWS

DS.PNUAT DS.PNUUS DS.TCBPN DS.TOTMT DS.TOTMW DS.TOTNW

DS.TOTWL

Thanks to Don Deese, (CPExpert), Computer Management Sciences, USA.
CHANGE 27.199 Documentation of this error message:

JCLTEST9 ERROR: PHYSICAL FILE DOES NOT EXIST, ATP1DKP.ADM.DATA.

Aug 12, 2009 occurs when a FILE ADM statement was found, but there was

no //ADM DD in the job. SAS then looks for a file named

JOBNAME.FILENAME.DATA, and fails.
CHANGE 27.198 Even when PDBOUT= was not specified, so the null default

ANALRMFR (write to //WORK) was desired, ANALRMFR still wrote the

Aug 12, 2009 TYPE7xxx datasets to //PDB instead of //WORK. %LETs were

added to protect the SPLIT70 logic.

Thanks to Kim Westcott, New York State Office of Technology, USA.
CHANGE 27.197 Change 27.175 added DB2STATB/DB2STATS/DB2GBPST but failed

ANALDB2R (VARIABLE SYSTEM NOT FOUND) if DB2STATB was redirected to

Aug 12, 2009 a separate input DDNAME.

Thanks to Scott Wiig, USBank, USA.

CHANGE 27.196 Change 27.138 heuristic test for LENGTH=412 for subtypes

VMAC85 78,79,88 did not generalize and still failed with record

Aug 12, 2009 from z/OS 1.10 subtype 78 record. The test was revised

to test R85PVRM IN (1030 1090 1100 1110) to hopefully