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

-APAR OA25559 adds two new variables S42DSRDD,S42DSRDT to

subtype 5 and subtype 6 records, and these variables are

are kept in TYPE42DS, TYPE42SR, and TYPE42VT datasets:

S42DSRDD='AVERAGE*READ MS*DISCONNECT*TIME'

S42DSRDT='READ*OPERATIONS*TO*DATASET'

Change 27.061 MXG 27.01 only. A DEBUGging PUT statement at line 235

VMAC88 PUT _N_= SUBTYPE= SMFTIME= SMF88LTD= SMF88LIT= GMT88OFF=;

Apr 8, 2009 is now deleted by this change.

Thanks to Michael Creech, Lender Processing Services, Inc, USA.

Change 27.060 Cosmetic. Labels for TYPE23 BUFSZMAX and WRNSZMAX said

VMAC23 "PERCENT", but BUFSZMAX has always been the MAX VALUE.

Apr 8, 2009 WRNSZMAX was the Warning PERCENT thru z/OS 1.5, but in

z/OS 1.6 the description was changed to "VALUE", but it

was NOT documented in the changes in the 1.6 SMF Manual.

Now, both labels say VALUE rather than PERCENT.

Thanks to Chuck Hopf, Bank of America, USA.
Change 27.059 SMF 102 IFCID=108 variables starting at QW0108PR were not

VMAC102 input because QWT02R1N, instead of QWT02R1L (length) was

Apr 6, 2009 tested for GE 160. Additionally, these new variables are

now input and kept in T102S018 dataset:

QW0108DY QW0108DP QW0108RO QW0108KD QW0108OH QW0108IW

QW0108SC QW0108CC

Thanks to Tony Anderson, Blue Cross Blue Shield of Alabama, USA.
Change 27.058 Change 27.015 enhanced VMACSMF to test IF SUBTYPE GT 255

VMACSMF because some user SMF records store their subtype in the

Apr 6, 2009 left byte of the two byte field. MXG printed an ERROR

message, and used the left byte as the subtype number,

but this ERROR message was not documented in that text.

I also failed to protect the optional BMC CICS Statistics

SMF 110 records that are written as subtype=2818 or 47874

('BB02'x,'0B02'x) are really subtype=2 STID=200 to 205.

IF you had enabled the optional IMACICBB member to decode

these optional STID's, then their CICSBBxx datasets would

have had no observations, plus the new MXG ERROR message.

Thanks to Lorena Ortenzi, UniCredit Global Info Services, ITALY.

Change 27.057 MXG 27.01. DCOLLECT DCOLVOLS the volume size variables

VMACDCOL DCVALLOC DCVFRESP DCVLGEXT DCVVLCAP are small by 1024.

Apr 3, 2009 Change 27.034 for EAV support mislocated the code that

multiplies the KB value by 1024 so it was ONLY executed

if the record is the longer, EAV-updated, volume record.

Thanks to Frank Lund, EDB, NORWAY.

Thanks to Chris A. Hughes, University of Georgia, USA.
Change 27.056 Support for Control-D "Decollating" SMF record, which is

EXTYCTCD a new, different SMF record than the Control-D TYPECTLD

IMACCTCD support. This new record creates dataset CONTROCD.

TYPECTCD

TYPSCTCD

VMACCTCD

VMXGINIT

Apr 2, 2009

Thanks to Davide Marone, SGS S.p.a. ITALY

Thanks to Carlo Gavinelli, SGS S.p.a. ITALY

Thanks to Gianvittorio Negri, SAS Institute, ITALY.
Change 27.055 OBJECTSTAR (a/k/a HURON) INPUT STATEMENT EXCEEDED RECORD

VMACHURN ERROR for short Subtype 13 record; record should have

Apr 2, 2009 3123 bytes for 26 segments, but LENGTH is only 3110.

MXG prints message and hex dump of first 3 instances.

Thanks to David Stone, TESCO, ENGLAND.
Change 27.054 MXG 27.01 ONLY, CRITICAL ERROR FOR SMF 42 SubTypes 1-20.

VMAC42 I really botched this one. A partial statement was left

Mar 31, 2009 (IF SUBTYPE GT 20 THEN) ahead of IF OFFPROD GT 0 THEN DO,

preventing INPUT of STARTIME, ENDTIME & DURATM variables

for these datasets that are created from subtypes 1-20:

TYPE4220 TYPE4237 TYPE42AD TYPE42AU TYPE42CC TYPE42CU

TYPE42CV TYPE42D1 TYPE42D2 TYPE42DS TYPE42EX TYPE42L1

TYPE42L2 TYPE42NF TYPE42NU TYPE42P1 TYPE42P2 TYPE42P3

TYPE42S1 TYPE42S2 TYPE42SC TYPE42SR TYPE42TO TYPE42VL

TYPE42VS TYPE42VT TYPE42X1 TYPE42X2 TYPE42XR TYPE42XV

Not only are STARTIME, ENDTIME, and DURATM missing values

but also variables that are calculated from DURATM, such

as, in TYPE42DS, CACHRATE, DASDMPL, and DASDRATE also are

missing values.

Thanks to Paul Volpi, UHC, USA.
Change 27.053 -CICS Statistics STID=108 in CICS/TS 3.2 added 48 bytes as

FORMATS part of the 284 bytes INPUT in MXG 27.01 for CICS/TS 4.1.

UTILEXCL Now, the first 48 are INPUT then the other 236 are INPUT.

VMAC110 STID=108 DFHSORTS data is output in dataset CICTCPIP.

Mar 31, 2009 -CICS Statistics STID=31 inconsistent length/LDBNRDSN was

Apr 1, 2009 only partially circumvented in Change 26.052, as it used

Apr 2, 2009 re-calculated LDRNSKIP=(STILEN-104)/44 for segment count

APR 4, 2009 only if LDRNSKIP was GT LDBNRDSN, causing the prior count

to be used when LDRNSKIP=LDBNRDSN. Since it is now clear

that LDBNRDSN is the number of POPULATED segments, but is

NOT the number of segments, LDRNSKIP=(STILEN-104)/44 is

always used to control the number of segments read in.

Existing test for LDBDSNAM GT ' ' causes only populated

segments to be output in the CICLDBDS statists dataset.

-In CICSTRAN dataset, back in 1994, Change 12.166 noted

that TASKNR (IBM TRANNUM) is not always a packed decimal,

and can contain four-character-in-EBCDIC values:

bytes old TASKNR description

1st 2-4 (decimal)

'00'X,'III' 13,224,393 SYSTEM INITIALIZATION TASK

'00'X,'J00' 13,758,704 JOUR CONTROL, JOURNAL=00

'00'X,'J99' 13,761,017 JOUR CONTROL, JOURNAL=99

'00'X,'JBS' 13,746,914 JOURNAL CONTROL

'40'X,'TCP' 14,926,807 TERMINAL CONTROL

but that change only stored the right 3-bytes as a large

decimal value when a character value was found, without

decoding or mapping that strange TASKNR value. And, that

change was not implemented in UTILEXCL logic, causing the

TASKNR to be a missing value for these character values

when IMACEXCL built by UTILEXCL was used. Now, if TASKNR

is not a valid packed decimal value, the full four-byte

field is input as binary, and TASKNR is FORMATed with new

MG110TN format that decodes those binary values to print

the actual 'III', 'Jnn', 'JBX', or 'TCP' text value,

right justified to align with numeric values of TASKNR.

And UTILEXCL was updated with this revised logic.

Because TASKNR is the only packed decimal field in the

CICSTRAN segment the "vanilla" VMAC110 tested IF TASKNR=.

to detect misalignment (that is how MXG finds EXCLUDEd

fields or optional data segments), but with this change,

TASKNR can never be a missing value, so the new test is

IF TASKNR EQ 0 OR

(TASKNR GT 99999 AND COMPRESS(PUT(TASKNR,MG110TN.))

NOT IN : ('III','TCP','JBS','J') ) THEN DO;

The original test for TASKNR=. could not be used when the

IMACEXCL was in use, since the site might have EXCLUDEd

the TASKNR field, but since this test is now valid even

if TASKNR is excluded, this misalignment test was moved

so it is always applied, even when IMACEXCL is used.

Thanks to Paul Volpi, UHC, USA.

Thanks to Jon Keister, Caremark, USA.
Change 27.052 Updates to PMACC01 DB2PM-like report had a number of

ANALDB2R blank values introduced in MXG 26.26, which also changed

Mar 28, 2009 the format of some durations (Class 7 for example) to

match the DB2PM printed time format.

Thanks to Paul Billick, QVC Inc, USA.
Change 27.051 -%VMXGOPTR is used internally in MXG to alter the user's

VMXGOPTR options, which are then reset to their original value.

VMXGSUM For example, in VMXGSUM, we must have OBS=MAX for the

Mar 27, 2009 internal parsing DATA steps, so we store your OBS value,

Apr 2, 2009 change it to OBS=MAX for our purpose, and then restore

Apr 13, 2009 your ORIGINAL option value. But if %VMXGOPTR was then

Apr 29, 2009 invoked a second time for that same option, we used your

ORIGINAL value instead of your CURRENT value. Now, each

time %VMXGOPTR is invoked with NEWVALUE not ORIGINAL, we

store the current value of that option, so that when the

resetting %VMXGOPTR invocation with ORIGINAL, we then can

reset that option to the LAST stored value.

-After ORIGINAL has been reset, the text "MXG" is stored.

If a subsequent ORIGINAL invocation finds that "MXG", a

"MXGWARN STORED VALUE FOR xxxxxx IS MXG" prints to alert

you that you have un-paired %VMXGOPTR invocations. This

might not be a problem, depending on your intended use of

%VMXGOPTR.

-VMXGSUM required change because it did not have matched

pairs of %VMXGOPTR invocations.

-It is VERY unlikely any MXG users have actually used the

%VMXGOPTR macro, and even less likely that anyone used

that retained value of the original %VMXGOPTR design,

but, technically, this is an UNCOMPATIBLE change.

-Note that if you should use %VMXGOPTR, that you need to

use it exclusively to change options; if you interleave

%VMXGOPTR and OPTIONS statements, your results might be

unexpected.

This change was precipitated by a simple user program:

OPTIONS OBS=0;

%INCLUDE SOURCLIB(TYPEDB2);

RUN;

OPTIONS OBS=MAX;

%INCLUDE SOURCLIB(ASUMUOW,ASUMCICX);

that causes ASUMCICX to fail, because the internal

%VMXGOPTR in VMXGSUM reset the option OBS=0, which was

the last value it stored, and so PDB.CICS built by the

ASUMCICX had 0 OBS.

Thanks to Paul Volpi, UHC, USA.
Change 27.050 Support for LPAR NMON data for RAWLPAR and RAWCPUTOTAL

EXNMONRC objects create two new datasets:

EXNMONRL DDDDDD DATASET DESCRIPTION

IMACNMON NMONRC NMONRAWC NMON RAW CPU

VMACNMON NMONRL NMONRAWL NMON RAW LPAR

VMXGINIT

Apr 13, 2009

Thanks to Peter Turner, Saint George Bank, AUSTRALIA.

Change 27.049 %VGETSYSI for z/OS only, returns global macro variables

VGETSYSI MXGSYSID and MXGSUSEC which contain the SYSTEM on which

Mar 26, 2009 SAS is executing, and the SU_SEC value for that system.

Thanks to Chuck Hopf, Bank of America, USA.

Change 27.048 Variable TAPEDRVS in PDB.ASUMTALO is the cross product of

ASUMTALO allocated tape drives times their duration, so it can be

TRNDTALO a large and confusing number. New variable AVGDRIVE is

Mar 26, 2009 now added to PDB.ASUMTALO with the average number of tape

drives that are allocated. TRNDTALO was also updated with

the new variable.

Thanks to Graham Cornwell, DDS, UK.

Thanks to Craig Heron, DDS, UK.

Change 27.047 -MXG 27.01 only. The EAV support added in Change 27.034

VMACDCOL was misaligned, i.e., it was incorrect. This change has

Mar 25, 2009 now been validated with an EAV volume's DCOLLECT record.

-The above error also mis-located the previous tests for

DCVFLAG1 and DCVERROR, causing all of the flag variables

created from them to be blank.

Only the DCOLVOLS dataset was in error in 27.01.

But see also Change 27.057.

Thanks to Dr. H. Pat Artis, Performance Associated, USA.

Thanks to Brian Currah, Independent Consultant, CANADA.

Thanks to Dan Case, Mayo Clinic, USA.
Change 27.046 Major updates to D,V,X records created by EDGHSKP utility

VMACEDGR add new variables to EDGRDEXT, EDGRVEXT, and EDGRXEXT

Mar 25, 2009 dataset. These extensive changes have been validated

Apr 2, 2009 with z/OS 1.7, 1.8, 1.9, and 1.10 RMM records.

Apr 4, 2009 -The RMM EDGRDEXT final segment test for GE 95 should have

been GE 31. RVMVDSN1,RVDSTBIN,RVDSTMED,RVVVOL1 do not

exist in the DEXT record; these fields were added in 2003

in Change 21.158, but I must have misread the doc!

Also, the RDEXPDT and RDEXPDTO date fields were not

converted to SAS date values.

-Support for APAR OA22132 adds variables to D/V/X records:

RDFACTOR='SPACE/SIZE*FACTOR' (MB in EBCDIC)

RDSIZE ='RDSIZE'

Variable RDSIZE is converted by RDFACTOR and formatted

with MGBYTES to show size in KB/MB/GB/ etc Note that

RDSIZE will be a missing value when that 10-byte field

is blanks, which does occur in many (old?) records.

Thanks to Paul Volpi, UHC, USA.

Change 27.045 -These QISTWFxx variables in DB2STATS were wrong because

VMACDB2 MXG de-accumulated them, but are not accumulated values:

Mar 20, 2009 QISTW04K QISTW32K QISTWF04 QISTWF32 QISTWFCK QISTWFCU

Mar 24, 2009 QISTWFMU QISTWFMX

Apr 11, 2009 -Variables QWACIXLT and QWAXIXLT were incorrectly divided

Apr 19, 2009 by 4096; they are in microseconds, not TOD clock units.

-Variables QSSTGETR, QWSDARIR, Q9STCTX1 were deaccumulated

twice, QSSTGETE was not deaccumulated.

-Dataset DB2STATS variables for Star Joins are marked in

the DSECT as "SERVICEABILITY", but are added

-Star Join statistics variables are added to DB2STATS:

QISJTRY ='ALLOCATION*REQUESTS IN*STAR JOIN*POOL'

QISJFAIL='FAILURES*CAUSED BY*FULL*STAR JOIN*POOL'

QISJSIZE='CURRENT*SIZE OF*STAR JOIN*POOL (MB)'

QISJMAX ='MAXIMUM*SIZE OF*STAR JOIN*POOL (MB)'

They are marked "SERVICEABILITY" in the current DSECT.

Thanks to Lori Masulis, Fidelity Systems, USA.

Thanks to Rachel Holt, Fidelity Systems, USA.

Thanks to Terry L. Berman, DST Systems, USA.

Thanks to Steve Wood, DST Systems, USA.

Change 27.044 Variable CPCFNAME was defined as $10 in VMXG70PR but it

VMXG70PR contains only $8, and its length in RMFINTRV is also $8.

Mar 20, 2009 Since RMFINTRV is created by default in BUILDPDB, I chose

to change the length in the ASUM70PR datasets to avoid

conflicts.

Thanks to Jim Horne, Lowe's Companies, USA.

Change 27.043 -HP OpenView on AIX variables RELEASE, SOFTWARE, SYSID

ADOCMWAI were wrong or had no values with the current REPORT macro

ADOCMWSU (which is now in ADOCMWAI, so you can use that command to

VMACMWAI extract the data in the format expected by the revised

VMACMWSU VMACMWAI, and similarly, VMACMWSU/ADOCMWSU members are

Mar 18, 2009 updated for HP OpenView on Sun, with ADOCMWSU containing

Apr 10, 2009 the REPORT macro to match the VMACMWSU code.

Thanks to Albert Jacobs, KBC ICT Services NV, BELGIUM.

====== Changes thru 27.042 were in MXG 27.01 dated Mar 17, 2009=========
Change 27.042 Support for ASG TMON for CICS V3.2 was in MXG since 25.11

VMACTMO2 for important MONITASK dataset, and most other datasets,

Mar 17, 2009 although MONIIWT & MONITDQ datasets require this change,

as their offsets were relocated by insertion of the new

fields for the MONISYST and MONITDQ datasets.

-Dataset MONITASK, new variables:

TADFHAP1 TADFHAP2

-Dataset MONIPA, new variables:

PACMDLYC PACMDLYT PADFHAP1 PADFHAP2 PAICSCC

PAICSCD PAICSRC PAICSRD PAL9CPUC PAL9CPUT

PAPCDCC PAPCDLL PAPCDRL PAPCLCC PAPCRCC

PAPCRCL PAPCXCC PAPGBCC PAPGCCC PAPGCTC

PAPGGCC PAPGGCL PAPGMCC PAPGPCC PAPGPCL

PASTDLYC PASTDLYT PAWBBOC PAWBI1C PAWBIRC

PAWBIWC PAWBO1C PAWBOSC PAWBPRC PAWBRDL

PAWBROC PAWBWDL PAWBWOC PAX8CPUC PAX8CPUT

PAX9CPUC PAX9CPUT PAXTDLYC PAXTDLYT

-Dataset MONISYST, new variables:

TIMQSCLC TIMQSCLT TIMQSGTC TIMQSGTT TIMQSIQC

TIMQSIQT TIMQSOPC TIMQSOPT TIMQSP1C TIMQSP1T

TIMQSPTC TIMQSPTT TIMQSSTC TIMQSSTT TIRECICT

-Dataset MONITDQ, new variables:

TDRECICT
Change 27.041 The example to summarize PDB.SMFINTRV into PDB.ASUMSMFI

ASUMSMFI was enhanced to specify the SYNC59 argument, with the

Mar 17, 2009 default SYNC59=NO, so there would be no change to the

summarized output (until you change NO to YES!).

Thanks to Larry Stahl, IBM Global Services, USA.
Change 27.040 Dataset TYPE78PA (Private Area for Monitored Jobs) ID=78

VMAC78 Subtype 2 INPUT STATEMENT EXCEEDED RECORD LENGTH ERROR

Mar 17, 2009 was due to MXG miscoding the new-in-z/OS 1.10 Memory

Objects. COMOHWM,LGMOHWM,SHMOHWM,TOFRHWM,TOMOHWM vars

were INPUT but never existed and are removed. Labels for

their MIN and MAX values are changed to FRAMES instead of

BYTES, and their MGBYTES formats are removed.

Thanks to Scott Chapman, American Electric Power, USA.

Change 27.039 Change 26.183 is wrong; it removed the QBAC and QTXA vars

ASUMDB2P from the summary in ASUMDB2P and TRNDDB2 code, claiming

TRNDDB2P those variables were not in the IFCID=239 (101 subtype 1)

VMACDB2 Package Data records starting in V7.1. But they all DO

Mar 16, 2009 exist in DB2ACCTP dataset in DB2 V8.1 or later, so they

are now restored in ASUMDB2P and TRNDDB2P summary code.

The QBAC and QTXA fields are only populated in DB2ACCTP

dataset if DB2 Accounting Trace Class 10 is enabled.

Those variables have ALWAYS been kept in the DB2ACCTP

dataset, so this only impacts users of MXG's summaries.

Thanks to Giuseppe Giacomodonato, E.P.V. Technologies, ITALY.
Change 27.038 Change 26.183 SMF/Performance Sentry Release 3.1.4.

VMACNTSM A major enhancement in NTSMF adds IDPROCES and SQLVERSN

Mar 15, 2009 to 72 SQLSERVER/MSSQL dataset, so that the individual

SQL Server Process can be identified. There were also

37 new objects that are now supported by these new exit

members:

exntdtos exntm8ca exntm8co exntm8dm exntm8dp exntm8lo

exntm8md exntm8me exntm8pa exntm8pc exntm8pi exntm8pr

exntm8se exntm8th exntmlca exntmlco exntmldm exntmldp

exntmllo exntmlmd exntmlme exntmlpa exntmlpc exntmlpi

exntmlpr exntmlse exntmlth exntqlbl exntqldf exntqlrp

exntqlwg exntrpsv exntsqbt exntsqdf exntsqpl exntsqrp

exntsqwg
Change 27.037 The example in the comments to execute did not have the

ADOC70PR //SPIN DD UNIT=SYSDA,SPACE=CYL(1,1)

Mar 15, 2009 that is required by that example program. It can be

a temporary dataset, because this example builds a

single output PDB library for analysis. The SPIN DD

is only used in the RMFINTRV program to create SPINRMFI,

used to keep the last four hours of CPU busy for the old

MXG MSU4HRAV variable when RMFINTRV is created on a daily

basis. MSU4HRAV is actually archaic, predating the IBM

addition of their SMF70LAC, which should be used in place

of MXG's MSU4HRAV.

Thanks to MP Welch, SPRINT, USA.

Change 27.036 -Fields DBHCEUID, DBHCEUTX, DBHCEUWN are now kept in the

VMACTMDB TMON for DB2 TMDBBF,TMDBBG,TMDBBH, and TMDBBI datasets,

Mar 14, 2009 and all three have any '00'X translated to blanks.

-Variable HSIID now formatted with MGDB2II.

Thanks to Ernie Amador, UC Davis Health System, USA.
Change 27.035 Previously, you could not leave SYSTEM blank in TIMETABL,

VMXGTIME but with this revision, if ALL of your SYSTEMS are to be

Mar 11, 2009 changed the same way, you do NOT need to fill in SYSTEM,

and the same for SYSPLEX. If you have systems on

different time zones, then each SYSTEM or SYSPLEX must be

in the TIMETABLE (but only those that differ from base).

So if you have 3 systems, 2 on local, and one on local+2,

to convert all systems to local, only the local+2 system

needs to be listed in TIMETABL.

Thanks to Richard Schwartz, State Street Bank, USA.

Change 27.034 Support for EAV, Extended Address Volume, devices with

VMACDCOL more than 65520 cylinders, introduced in z/OS 1.10, with

Mar 11, 2009 HUGE volume sizes, 236GB (256K Cylinders) first release.

EAV devices are formally documented as 3390-A devices.

Dataset DCOLVOLS: new variables:

DCVCYLMG='EAV*INDICATOR*FLAG'

DCVDPTYP='PHYSICAL*DEVICE*TYPE'

DCVEAVOL='EAV*INDICATOR*FLAG'

DCVTRALC='ALLOCATED*TRACK*MANAGED'

DCVTRFRG='FRAGMENTATION*INDEX*TRACK*MANAGED'

DCVTRFRX='FREE*EXTENTS*TRACK*MANAGED'

DCVTRFSP='FREESPACE*TRACK*MANAGED'

DCVTRLGE='LARGEST*EXTENT*TRACK*MANAGED'

DCVTRPCT='PCT FREESPACE*TRACK*MANAGED'

DCVTRVLC='TOTAL CAPACITY*TRACK*MANAGED'

The new Track-Managed Space or Base Addressing Space

is the first 65520 cylinders (54GB), and space above

that is the Extended Addressing Space.

Disk allocation in the EAS is in units of 21 cylinders.

Initial implementation is VSAM only, KSDS,RRDS,ESDS and

linear, so DB2 and zFS can use EAV volumes.

Dataset DCOLSC: new variables:

DSCBAKUP='ACC*BACKUP*PARM*VALUE'

DSCBUSP ='ACC*BACKUP*PARM*SPECIFIED'

DSCDSSEP='DATASET*SEPARATE*PROFILE'

DSCFLAG3='DSC*FLAG*BYTE*3'

DSCFOSL ='OAM*SUBLEVEL*1'

DSCTIER ='MULTI-TIER*SG*VALUE'

DSCTIERS='MULTI-TIER*SG*SPECIFIED'

DSCVERSN='ACC*VERSION*PARM*VALUE'

DSCVERSP='ACC*VERSIONING*PARM*SPECIFIED'

Dataset DCOLDC new variables:

DDCFSDB ='SDB*IS*SPECIFIED?'

DDCFVORD='OVERRIDE*JCL*SPECIFIED?'

Dataset DCOLBKUP new variables:

UBFLAG3='INFORMATION*FLAG*3'

UBNOSPH='SPHERE(NO)*PROCESSED?'

UBGVCN ='GENVSAMCOMPNAMES*PROCESSED?'

-Variable UMNMIG, number of times migrated, is reserved in

dataset DCOLMIGS.

-Doc notes:

Length of VOLUME RECORD structure still 112.

-Type 1415 enhancement:

-MXG Change 25.071 added variable EADSCBOK was decoded

from SMF14EADSCB, in the type 5 extended information

segment of the SMF type 14 and 15 records. It means

that the user program provided a valid DCBE that has

EADSCB=OK. It does NOT mean that OPEN needed this

option.

If this bit is zero, then the program did not specify

EADSCB=OK on the DCBE macro. IBM suggests upgrading

the program to handle 28-bit cylinders and coding

EADSCB=OK. This is to help plan for when EAV might

support more types of data sets.

-MXG Change 25.071 also decoded variable SMF14EAV from

SMF14EXCPBAM. This is 'Y' if the user program issued

one or more instances of the EXCP or XDAP macro since

the BSAM, BPAM or QSAM DCB was opened.

-SMF19 enhancements added by Change 26.071, documentation:

These existing two-byte counters describe free space for

the entire volume. If the value exceeds the capacity of

the counter, it will be set to the maximum of 65535. The

correct values will always be in SMF19EVF and SMF19TMF.

MXG var IBM name

NRDSCBS SMF19NDS Number of DSCBs in VTOC

NRDSCB0 SMF19DSR Number of format 0 (unused) DSCBs

in VTOC

NRALTRKS SMF19NAT Number of unused alternate tracks

CYLAVAIL SMF19SPC Numbers of free complete cylinders

TRKAVAIL SMF19??? Number of free tracks

MAXCYLS SMF19LEX Numbers of complete cylinders in

largest free extent

MAXTRKS SMF19??? Numbers of complete tracks in

largest free extent

NRUNEXTS SMF19NUE Number of free extents

New SMF19CYM variable will mean that the volume has

cylinder-managed space.

The following will be new four-byte unsigned counters at

the end of the record:

SMF19SDS Number of DSCBs in VTOC

SMF19SL0 Number of format 0 (unused) DSCBs in VTOC

The following 5 new counters apply to the whole volume.

SMF19SUC Number of free complete cylinders