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

ASMRMFV segment level that is undocumented and added 376 bytes.

VMACRMFV This caused MXG variables for Service Class, Report

Jun 8, 2012 Class, Workload, and Resource Group data extensions to be

invalid.

-ASMRMFV space message RMFV030I could have invalid values

for HARBA, HURBA, and AVAIL fields when the input RMF

Monitor III VSAM data set was non-EF and exceeded 2 GB in

size. Subsequent message RMFV031I would have invalid

percentages for USED and AVAIL fields possibly exceeding

100 percent. NOTE: Assembly and link-edit of ASMRMFV is

optional for MXG users with only EF RMF III VSAM data

sets or only non-EF data sets not exceeding 2 GB in size.

Otherwise, reinstall of ASMRMFV is strongly recommended

to resolve this issue. o Warren Cravey, Fidelity

Thanks to Warren Cravey, Fidelity, USA.

Change 30.104 -Variable SMF70CIN in TYPE70EN was incorrectly set to the

VMAC7072 value 'ZIP', but that is now SMF70CIN='IIP', consistent

Jun 8, 2012 with SMF70CIN in other MXG datasets.

-For zIIP engines, in dataset PDB.TYPE70PR, the variable

NEWWAIT contains the CPUWAITM for SMF70CIN='IIP'.

-In PDB.TYPE70EN, variable CPUWAITM is valid for CP and

IIP engine's wait time.

Thanks to Jim Horne, Lowe's, USA.

Change 30.103 -Support for z/OS 1.13-added (COMPAT) RAS (Reliability,

VMAC1415 Availability, Serviceability) segment adds flag variables

Jun 5, 2012 that are $HEX2. formatted. Because I don't expect these

new variables to be of great import, separate bit-level

variables aren't created, but these are the values:

SMF14RFG0

'1.......'B='DCBE reject flags present'

'.1......'B='PARTREL flags present'

SMF14RASDATA0

'1.......'B='DCBE invalidated because EXCP and no

foundation extension present.

'.1......'B='DCBE invalidated because DSORG is not PS,

PO OR DA.

'..1.....'B='DCBE invalidated because storage is not

addressable.

'...1....'B='DCBE invalidated because DCBE storage is

not in key of caller.

'....1...'B='DCBE invalidated because the DCBEID is not

'DCBE'.

'.....1..'B='DCBE invalidated because it is not at least

the minimum length required (56 bytes)

'......1.'B='DCBEHIARC flags set but DCBDCBE is zeros.

SMF14RASDATA1

'1.......'0=Partial release not called by CLOSE because

VIO data set

'.1......'1=Partial release not called by CLOSE because

task is abending.

'..1.....'2=Partial release not called by CLOSE because

not opened for output

'...1....'3=Partial release not called by CLOSE because

EXCP DCB but no direct access device section

present.

'....1...'4=Partial release not called by CLOSE because

even though opened for output, last I/O was

not output

'.....1..'5=Partial release had an I/O error

Thanks to Karl Lasecki, Chemical Abstracts Service, USA.
Change 30.102 These DB2xxxxx variables were not MGBYTES formatted, and

VMACDB2 had incorrect labels.

Jun 5, 2012 NDB2STOR ALLOWSTR THRDUSE QWOSDRSU QWOSDVSU

Thanks to Robert M. Dahlia, SunTrust, USA.

Change 30.101 The TITLE7 statement was repeated, so the title line that

ANALDDCN should have reported duplicate bytes DUPBYTES was not

Jun 4, 2012 printed.

Thanks to Tom White, Dell, USA.

Change 30.100 UTILEXCL in MXG 30.03 was defective and the IMACEXCL that

UTILEXCL it created did not input JMVTIMCN which caused errors,

Jun 4, 2012 usually the "NON-FIRST-TRANSACTION RECORD NUMBER" error.

Thanks to Victoria Lepak, Aetna, USA.

Change 30.099 Support for CO:Z SMF 119 Subtypes 192 and 193 records.

EXT119X2 Four new datasets are created:

EXT119X3 DDDDDD DATASET DESCRIPTION

EXT119M2 T119X2 TY119192 CO:Z SFTP SERVER LOG

EXT119M3 T119M2 TY119M92 CO:Z SFTP SERVER LOG MESSAGES

IMAC119 T119X3 TY119193 CO:Z SFTP CLIENT LOG

VMAC119 T119M3 TY119M93 CO:Z SFTP CLIENT LOG MESSAGES

VMXGINIT Jun 18: RACFUSER added to KEEP for T119M2/T119M3 as it is

Jun 3, 2012 in the BY list for both datasets.

Jun 18, 2012

Thanks to Harald Seifert, Huk-Coburg, GERMANY.
Change 30.098 New fields added to DCOLLECT BKUP records (COMPATIBLY) by

VMACDCOL z/OS 1.11 are now created in DCOLBKUP dataset:

Jun 2, 2012 UBLFS ='LARGE*FORMAT*SEQ*DATASET?'

UBNEWNM='NEWNAME*SPECIFIED?'

UBF_RETAIN_SPCD='RETAIN*SPECIFIED?'

UBF_NEVER_EXP='NEVER*EXPIRE?'

UB_RETAINDAYS='SPECIFIED*RETAINDAYS*VALUE'

Thanks to Karl Lasecki, Chemical Abstracts Service, USA.

Change 30.097 Velocity Software XAMCPUTO dataset contains totals, but

VMACXAM did not contain a count of CPUS. Variable NRCPUS is now

Jun 2, 2012 created, and forced to be an integer using

NRCPUS=FLOOR(0.1+(DURATM/(STOPTIME-STARTIME)));

Thanks to Andrew Petersen, CSC, AUSTRALIA.
====== Changes thru 30.096 were in MXG 30.03 dated May 30, 2012=========
Change 30.096 Support for ZEN OSA MONITOR PTF ZOM1322 that INCOMPATIBLY

VMACZOSA increased the length of ZOSALINK from 13 to 17 bytes.

May 29, 2012

Jun 1, 2012

Change 30.095 RMF III Enhancements.

ASMRMFV -A new table called MXG is now created which contains

VMACRMFV ASMRMFV assembly and execution data. The MXG table is

EXZRBASM created internally by ASMRMFV and is not a true RMF

VMXGINIT Monitor III table. The MXG table contains 49 variables

May 29, 2012 that describe both the assembly and the execution

environment of ASMRMFV. This data can be used either as

an audit trail or as a problem diagnostic aid. VMACRMFV

can also use this data for conditional logic decisions.

Only 1 MXG table observation will appear in the new

ZRBASM file for each run of ASMRMFV, so there is minimal

overhead.

-Two new extensions are added by ASMRMFV to the ASI

(address space) and ENC (enclave) table records with

WORKLOAD and RESOURCE GROUP information. These are in

addition to existing extensions for SERVICE CLASS and

REPORT CLASS data. VMACRMFV adds a total of 14

variables for this new information to the ZRBASI and

ZRBENC files.

-New ZRBASI variables are: ASIWNM ASIWDE ASIGNM ASIGDE

ASIGMN ASIGMX ASIGLT.

-New ZRBENC variables are: ENCWNM ENCWDE ENCGNM ENCGDE

ENCGMN ENCGMX ENCGLT.

-All ASMRMFV added data extensions now only include the

part of the information that is actually documented. In

some cases internal control block values specified a much

greater length and caused problems in VMACRMFV.

-ASMRMFV message RMFV000I now contains information on the

environment at the time of assembly.

-ASMRMFV message RMFV001I now contains information on the

environment at the time of execution.

-Corrected ASMRMFV comments to note that the optional

RMFFILT output data set can NOT be used as input to a MXG

PDB build. RMFFILT does not contain any DSH or SSH

records that would be needed.

-ASI table data extensions for very old releases of RMF

Monitor III V4.3.0 or below (pre 1994) will no longer be

attempted by ASMRMFV. ASI table data will still be

output but without the extensions in this case. The

header structure of these ancient records is not

compatible with the data extension process.

-Corrected a problem where SERVICE CLASS or REPORT CLASS

extensions for ASI or ENC table were incorrect when the

respective data indexes were zero. These values should

have been missing, but instead were populated with data

from the prior ASI entry. This was a limited condition.

-Code path length reductions were made to two subroutines

for ASI and ENC table processing in ASMRMFV by using

existing FINDxx subroutines.

-The ASIENTMX and ASIENTLN fields are now corrected in

ASMRMFV to show the true count and length in an ASI table

record.

-The ENCG3TLN field is now corrected in ASMRMFV to show

the true total length in an ENC table record.

-ASMRMFV now validates that input data sets are VSAM RRDS

with warning message RMFV017W issued if a data set is

non-VSAM or is VSAM but a non-RRDS type. In this case no

abend occurs, but final return code CC=4 is set, and

processing of the next input data set continues.

-ASMRMFV now issues warning message RMFV017W if an input

VSAM data set has a non-standard CISIZE or RECSIZE. IBM

intends that RMF Monitor III data sets be allocated with

the ERBVSDEF Clist which specifies the correct CISIZE and

RECSIZE values of 32768 and 32752 respectively. Use of

other values can result in a file that is unusable by RMF

Monitor III.

-A S0C4 Abend in ASMRMFV CPU table processing is corrected

that occurred when no other LPAR data was present. This

was most likely when a z/OS guest was running under z/VM.

-VMACRMFV did not correctly input the OSDKASID and

OSDPLIST fields from Summary Information section in the

OPD table.

-The following RMF Monitor III table records are now

blocked for efficiency in ASMRMFV output: CPD, CSR, ENT,

OPD, and SPG. This improvement results in up to 90%

reduction or more in output record count. However, this

means there are multiple data segments in each record

which can affect existing logic in any user modified

EXZRBxxx exit routines for the respective table.

VMACRMFV is upgraded to handle the additional data

entries in the blocked tables.

-Tutorial: Your tailoring logic in EXdddddd dataset exits

to control output of an MXG dataset needs this structure

to always be safe:

IF something THEN DO;

OUTPUT _Wdddddd;

END;

and can't use a DELETE, RETURN, nor "IF something;" logic

because when "something" is true, they stop the read of

this current record, skipping any un-read segments from

being tested for "something".

-Validity checking for the CPD, CSR, DVT, ENT, OPD, and

SPG tables in ASMRMFV for excessive or invalid header and

entry length is improved. When anomalies are detected

the entire table will be skipped. This should be a very

rare event.

-Prologue documentation in ASMRMFV source code has been

updated as needed including more discussion on skipped

records and entry blocking.

-NOTE: An assembly and link of each new ASMRMFV member is

ALWAYS STRONGLY recommended, keeping ASMRMFV and VMACRMFV

in sync, to create AND populate the new variables, enable

table record entry blocking, and implement other related

fixes/enhancements in this change.

-HOWEVER: Using the new VMACRMFV to process RMFBSAM data

created with the prior ASMRMFV program should not fail

unless invalid records are found, but all new variables

will have missing values.

Change 30.094 -SMF 113 counters were stored in the MXG DEFAULT=5 length,

ASUM113 but the four sets of counters can contain very large data

VMAC113 values, so variables BASICnn,PROBSTnn,CRYPTOnn,EXTNDnnne

May 28, 2012 are now stored in LENGTH 8, and the PROC MEANs in both

VMAC113 and ASUM113 now specify /INHERIT so the longer

length attribute will be preserved; comparison of short

and long length showed the longer length was needed as

there were some (smaller) values with shorter length.

-Variable SM113CST is removed from BY macro _BTY113 and

that macro matches the final sort order of PDB.TYPE113.

SM113CST was needed in the BY list for the intermediate

sorts but did not exist in the final PDB.TYPE113.

Thanks to Scott Barry, SBBWorks Inc, USA.
Change 30.093 If production jobs create reports that are sent to a data

ANALDB2R set or a report archive facility, but there was no data

May 28, 2012 matching your selections, the result is an empty file or

a nonexistent report. While MXG's SASLOG tells you that

happened (obscurely?), now, a message that no report was

produced is written to the output destination for these

reports: PMACC01 PMACC02 PMSTA02 PMAUD01 PMAUD02 PMAUD03
Change 30.092 If you have IMACEXCL in your "USERID.SOURCLIB", these new

UTILEXCL CICSTRAN duration variables were 16 times too large:

May 24, 2012 ISIOWTTM WMQGETTM MAXTTDTM T8CPUTTM MLXSSCTM JVMTHDTM

WMQASRTM RMITOTTM RMIOTHTM RMIDB2TM RMIDBCTM RMIEXDTM

RMIMQMTM RMICPSTM RMITCPTM

because UTILEXCL had generated incorrect code to convert.

Durations input as &PIB.4.6 use X=16*X;

Durations input as &PIB.8.6 use X=X/4096;

The incorrect syntax was X=16*X/4096; for 8.6.

Thanks to Victoria Lepak, Aetna, USA.

Change 30.091 Typoed character P in column one caused INPUT EXCEEDED in

VMACSVIE SVSUBTYP=27 records that had segment 21 data.

May 24, 2012

Jun 5, 2012

Thanks to Sam Knutson, GEICO, USA.
Change 30.090 Variable FERTREMI incorrectly contained the Local instead

VMACFERT of the Remote IP Address.

May 22, 2012

Thanks to Terry Back, Experian, USA.

Thanks to Jerome Vitner, Experian, USA.
Change 30.089 Support for DB2 V10 APAR PM24723 for IFCID=225 SMF ID=100

VMACDB2 Subtype=4, which COMPATIBLY adds new storage metrics that

May 22, 2012 are output in both DB2ST225 and DB2STATS datasets.

Thanks to Kerry J. Sommers, John Deere, USA.

Thanks to Ralph Baechle, John Deere, USA.
Change 30.088 MQMLOG datetime variables were on GMT clock; there is no

VMAC115 offset field, but these maximum values must be less than

May 17, 2012 the SMFTIME, so the variable GMT115TM is calculated from

SMFTIME-QJSTIOMAXIOT1 and used to adjust datetimestamps

to the local time zone.

Thanks to Joseph J. Faska, Depository Trust, USA.

Change 30.087 Variable TOTDEVHR in TYPE74CA had non-missing values that

VMAC74 were wrong when CACHIOTT was zero or missing; an MXG typo

May 14, 2012 set non-existent variable name TOTDEVNR to missing when

variable TOTDEVHR should have been set to missing value.

Thanks to Sharon Moir, JP Morgan Chase Bank, USA.
Change 30.086 Change 28.276 added BEGTIME= ENDTIME= parameters to

ANALHSM ANALHSM but, in the first step, the BEGTIME and ENDTIME

May 12, 2012 MACRO variables were set to the beginning of the data

and the end of the data so that those values could be

placed in the title lines of the reports. So all

subsequent use of those macro variables and checks for

their existence would be based on the timespan of the

actual data. To make matters worse, in REPORT 5 and

REPORT 6, the BEGTIME and ENDTIME datetime values were

being compared to TIME values for FSRTIMR and FSRTIME

so no data could ever be selected for those reports.

With this change the date/times for the report headings

are changed to BEGREPT and ENDREPT avoiding the

conflict with BEGTIME and ENDTIME selection and in

reports 5 and 6 datetime values are constructed as

they are in ASUMHSM so that the correct data can be

selected for the reports. There are also now MXGNOTEs

that will tell you when you have started processing the

data for each report and another if no data was found

for the report.

Thanks to Paul Volpi, UHC, USA.
Change 30.085 Cosmetic. UNMODSMF time is now aligned under SMFTIME to

VMACSMF make comparisons easier, and the _N_ value of LAST RECORD

May 9, 2012 IN GROUP message is no longer a missing value.
Change 30.084 ODS operator RS=NONE added to prevent wrapping of HTML

VMXGODSO statements (specific to z/OS but causes no problem for

May 9, 2012 ASCII ODS operations; only set for HTML output).
Change 30.083 -MXG 30.02, z/OS Only, SAS 9.1.3 SP4 Only:

VMAC71 ERROR: DOMAIN ERROR.

May 9, 2012 ERROR: TERMINATION DUE TO FLOATING POINT EXCEPTION

occurred in a PROC MEANS of DATA=TYPE71 in JCLTEST9.

This error did NOT occur with SAS 9.3 nor on ASCII SAS.

The Floating Point Exception resulted from MXG INPUTing

SMF71TLS field as RB4 when the field is binary (PIB4),

and a value of '00000AD1'x read as RB4 produced a value

that was a negative with E75 exponent, but it was only

when that value was subsequently read by PROC MEANS that

the error surfaced.

Thanks to John Loch, HP, USA.

CHANGE 30.082 Type 60 record with no VVR segment (for a VVDS) caused an

VMAC60 INPUT EXCEEDED RECORD LENGTH error on 3 days, and then

May 7, 2012 didn't. MXG now tests to verify a VVR segment exists (and

VVRLEN will be a missing value in these observations) but

why these records were created is unknown.

Thanks to Peter Krijger, ANZ National, NEW ZEALAND.

CHANGE 30.081 Enhancement adds rundays=mon tue wed ... to list the days

BLDSMPDB of the week when BLDSMPDB is to actually be executed; on

May 7, 2012 any other day, the program will terminate with MXGNOTEs.

The default, daily, is unchanged.

Thanks to Mynard Holloway, Spectrum Health, USA.
CHANGE 30.080 Actual GDPS records exposed wrong guesses I made when I

VMAC105 wrote code from the documentation: GDPS datetimestamps in

May 4, 2012 SM105STM/DTM/SST/SCT are reversed-SMFSTAMP8 with DATE

Jun 4, 2012 first, so simple SMFSTAMP8 format can't be used; DURATM

field is packed decimal not binary and needs divide by an

undocumented 10; the two RPO duration variables SM105SAR

and SM105SIR also needed an undocumented divide by ten;

the Product section's two variables are input and kept.

-Jun 4: IBM stored blanks for SM105SCD and SM105SCT which

caused INVALID data error. Test for blanks circumvents,

while a PMR is opened for the invalid data.

Thanks to Jeffrey A. Johns, UHC, USA.

Thanks to Paul Volpi, UHC, USA.
CHANGE 30.079 Cosmetic. Using EXEC SAS,CONFIG=CONFIMXG and MXGNAMES,

VMXGINIT DSNAMES in the //SOURCLIB concatenation are dynamically

May 4, 2012 allocated in reverse order to their concatenated order,

and dataset SASHELP.VEXTFL, which MXG reads to print the

DSNAMES on the log at initialization, is also reversed.

A PROC SORT of VEXTFL by DESCENDING LEVEL was inserted to

get the MXG list in correct order when CONFIMXG was used.

When MXGSAS93 and "static" allocation is done in JCL, the

value of LEVEL is zero and the DSNAMES are in the right

order; the sort doesn't alter that correct listing order.

Change 30.078 CICS Statistics storage variables in Change 29.221 were

VMAC110 thought to be in GB, and were multiplied by 1073741824,

May 28, 2012 but they are in MB, so they were 1024 times too large.

These variables are now multiplied by only 1048576 to

convert MB to bytes for the MGBYTES format:

SMSDSASZ SMSHWMDS SMSCSIZE SMSFSTG SMSHWMFS SMSLWMFS

SMSLFA SMSGDCUL SMSGDHWL SMSGDCUR SMSGDHWM SMTHWMPS

Thanks to Homayoun Riaza, United Health Group, USA.

CHANGE 30.077 -READDB2 internal parsing was revised to correctly process

READDB2 %LET MACKEEP=

May 3, 2012 MACRO _WDB2ACP DB2ACCTP.DB2ACCTP %

MACRO _SDB2ACP %

;

%READDB2(IFCIDS=ACCOUNT STATS,

WANTONLY=DB2ACCT DB2ACCTP DB2STATS DB2STATB,

PDBOUT=YES,LDB2ACC=DB2ACCT,LDB2STB=PDB);

to create only DB2ACCT.DB2ACCT, DB2ACCTP.DB2ACCTP,

PDB.DB2STATS and PDB.DB2STATB datasets in those LIBNAMEs.

-IFCID=STATS creates only/both PDB.DB2STATS & PDB.DB2STATB

but Change 27.322 incorrectly listed only DB2STATS so the

comments in READDB2 were clarified.

-Minor logic changes: DB2STATB routing was suppressed with

STATS, and SMF IDs to be read message only showed 101.

-Cleanup 106 message, detritus in WORK when IFCID=STATS is

used (newish STATS creates ONLY the PDB.DB2STATS dataset)

with PDBOUT=YES.

CHANGE 30.076 If USERADD= specified 102.xxx, but the length of xxx was

UTILBLDP less than 3 (e.g. 102.23 instead of 102.023), incorrect

May 2, 2012 macro token _C10223 instead of _C102023 was created which

caused unreferenced macro error. Not reported by a user.

Correction was when &LENGTH(&STRING2) was 1 or 2.
CHANGE 30.075 -Variables JOB JESNR JBL24 are now kept in TPMJBL24.

VMACTPMX -Variable READTIME is now kept in all TPMJBLxx datasets,

May 2, 2012 since READTIME as well as JOB and JESNR are required to

uniquely identify a "JOB".

-Variable TPMPCSDIFFER is now input as IB because it can

contain negative values.

Thanks to Scott Barry, SBBWorks Inc, USA.
CHANGE 30.074 Cosmetic. WARNING: MULTIPLE LENGTHS FOR BY VARIABLE LCU

ANALFIOE is removed with LENGTH DEFAULT=&MXGLEN added in the DATA

Apr 28, 2012 step that creates LCU from LCUID in TYPE78CF. There was

no error because the two lengths were 5 and 8 and LCU is

only a 2-byte binary number.

Thanks to Dan Case, Mayo Clinic, USA.

CHANGE 30.073 Support for CA Vantage Storage Resource Manager 12.6.00,

VMACSAMS INCOMPAT: SAMSLGVR='X' in new records; MXG tested GE '6'

May 29, 2012 when SAMS code was last updated. SAMSPOOL fields were

restructured, from 4 to 8 bytes, some from PD to PIB.

New variables created in SAMSPOOL dataset:

SAMSSYSP='SYSPLEX*WHERE*VANTAGE*RUNS'

SAMSLPAR='LPAR NAME*WHERE*VANTAGE*RUNS'

SAMSSUBS='SUBSYSTEM*WHERE*VANTAGE*RUNS'

SAMSHASL='LOCAL*HASH*VALUE'

SAMSHASG='GLOBAL*HASH*VALUE'

Thanks to Robert Brosnan, Goldman Sachs & Co., USA.
CHANGE 30.072 Support for RMF 74 APAR OA36831 which COMPATIBLY adds new

VMAC74 SMF74NSS='SKIPPED*SAMPS*INVALID*DELTA VALUE'

Apr 21, 2012 to TYPE74 dataset.
CHANGE 30.071 VMACTPX, written before _Vdddddd old-style macro were

VMACTPX designed, is updated to define and use those alternative

Apr 21, 2012 tokens to control what's kept in MXG datasets.

Thanks to Erling Andersen, SMT Data, DENMARK.

CHANGE 30.070 Support for CA-Spool Subtype 12 creates CNA9CX dataset,

EXCMA0CX but only creates the seven variables that were recognized

IMACCMA by format and content with the subtype 11. Additional

VMACCMA data fields will be decoded when documentation received.

VMXGINIT

Apr 21, 2012

Thanks to Orjan Lindholm, Volvo, SWEDEN.
CHANGE 30.069 These new in z/OS 1.13 TYPE72GO vars were not populated

VMAC7072 CPUPDPTM R723RTDM R723RTDC R723RTDT

Apr 20, 2012 because the test was 32 when only 24 bytes were added.

Thanks to Don Deese, (CPExpert), Computer Management Sciences, USA.

CHANGE 30.068 RMM dataset EDGRDEXT was expected, but the "D" and "V"

VMXGDSN records were combined into the "X" record, so DAILYDSN

Apr 20, 2012 was revised to use EDGRXEXT instead of EDGRDEXT.

Thanks to Jeff Dyke, USDA, USA.

CHANGE 30.067 LUNAME added where possible to SMF ID=119 example report.

ANAL119

Apr 19, 2012
CHANGE 30.066 New version of Ferret record is supported for subtype 1

VMACFERT and 4, based on test data without the vendor's DSECT.

Apr 20, 2012 -FERRET subtype 1 record with length of third triplet 80