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

Change 23.018 Invalid ID=9 record caused ARRAY SUBSCRIPT OUT OF RANGE,

VMXGGETM when UTILGETM was used by JCLTEST8. The SMF 9 record

Feb 18, 2005 does NOT contain a subtype, but the first byte was '7F'x,

with bit 1 (the '40'x bit) on, and that caused MXG to

read the bytes 19-20 as a two-byte SUBTYPE value, which

contained '0200'x, so SUBTYPE=512. Subtype was originally

a one-byte field, so only values of 0-511 were valid

subtypes. But as subtypes can now be two bytes, and

because the array is only used in VMXGGETM to tabulate

the count of SMF IDs and their SUBTYPEs, MXG only

intended to count subtypes 0-511 (and it protected known

larger subtypes, like CICS records from Boole & Babbage

by resetting them). The array error was because MXG's

protection was incorrect, the statement IF SUBTYPE GT 512

should have been IF SUBTYPE GT 511, and that is the

statement corrected by this change. The 02 value in byte

19 that caused SUBTYPE=512 is SMF9VPC, which identifies

the issuer of the Vary Path command that caused this SMF

9 to have been written, and I now see a new value

described:

2=Enterprise System Connection Manager

previously, only 0=Operator was written, which explains

why we are now correcting this problem in VMXGGETM. BUT:

The real error is that Bit 1 in SMF9FLG is on, which says

there's a valid subtype, which is not true for ID=9.

Thanks to Charles Bellamy, SCANA, USA.

Change 23.017 If %INCLUDE SOURCLIB(ASUMUOW) is used (JCLTESTV8/9 do),

VMXGUOW but IMACUOW was not updated to tell MXG that you want us

Feb 17, 2005 to create observations in PDB.ASUMUOW, your job will fail

ERROR: NO BY STATEMENT USED OR NO BY VARIABLES SPECIFIED

WARNING: DATASET WORK.UOWSORT MAY BE INCOMPLETE.

You were not supposed to have to update IMACUOW, but the

changes we made to add MQ-Series data to PDB.ASUMUOW were

always tested with a tailored IMACUOW; I failed to test

with the default IMACUOW in MXG 22.22, and we always

build the PDB.ASUMUOW dataset in our QA streams, to make

sure it is created without error.

The source of the problem is arguably a SAS design error;

we added the creation of a second dataset in a data step

whose first dataset was created as a VIEW, but the

default IMACUOW sets OPTIONS OBS=0; to prevent ASUMUOW

from sorts of both CICSTRAN and DB2ACCT. Unfortunately,

it turns out that that second dataset is NEVER created

when OBS=0 is in effect. While SAS Technical Support

investigates to see if this is a feature or a defect, we

circumvent the error by always creating that dataset

(_UOWCIC) with zero obs, before the data step with the

VIEW, so that it always will exist. If you enable IMACUOW

to create obs, the dataset is re-created with obserations

when the VIEW is executed.

Thanks to Alan Chenoweth, EDS, USA.

Thanks to Linda Piekarski, National Grange Mutual Insurance Co., USA.

Change 23.016 A VM 4.4 system's MONWRITE control block data contained a

VMACVMXA value of 00028709X instead of the normal 00008709X;

Feb 17, 2005 change line 5071 to test for both values:

IF IPARMLF1 NE 00008709X AND IPARMLF1 NE 00028709X THEN DO;

while we await a reply from IBM VM Support as to whether

this is an error or an undocumented feature.

Thanks to Mike Salyer, Citicorp, USA.
Change 23.015 If you create a WEEK.SMFINTRV or MONTH.SMFINTRV, your job

MONTHASC may get ERROR: BY VARIABLES NOT SORTED MON.SMFINTRV. The

MONTHBL3 sort order of PDB.SMFINTRV was changed in MXG 22.13 for

MONTHBLD Change 22.320 consolidation of MULTIDD='Y' observations.

MONTHBLS The original BY list in BUILDPDB was:

MONTHDSK READTIME JOB JESNR INTBTIME INTETIME

WEEKBL3 and the new sort order internally is

WEEKBL3D READTIME JOB JESNR INITTIME INTBTIME

WEEKBL3T In the cases that failed, there were multiple instances

WEEKBLD of the same STC JOBNAME that was an "early ASID" that had

WEEKBLDD started before SMF (i.e., it did NOT have a JESNR), and

WEEKBLDT both interval records had exactly the same values for

Feb 16, 2005 READTIME JOB JESNR and INITTIME, but the INTBTIMEs were

.01 seconds out of order, causing the NOT SORTED.

The solution is simple: in your WEEKBLD/MONTHBLD members,

change the BY list for the PDB.SMFINTRV dataset to only:

READTIME JOB JESNR

to eliminate the exposure.

-While using %INCLUDE SOURCLIB(TYPS30) will create dataset

PDB.SMFINTRV, that is no longer a recommended way,

because that dataset still contains the multiple

MULTIDD='Y' obs. See the example in the text of Change

22.320 if you want to create PDB.SMFINTRV directly from

SMF without BUILDPDB.

Thanks to Kevin Fletcher, CONSECO, USA.
Change 23.014 Consoldiated into Change 23.012.
Change 23.013 -Variable LPARCPUX is now kept in TYPE70PR, because it is

VMAC7072 the count of logical CPs assigned to the LPAR, and that

Feb 13, 2005 includes the number of "initial" CP engines, the number

of "reserved" CPs (which also includes ICFs, IFLs, and

IFAs that are assigned to the LPAR). This is important so

that enough "reserved" CPs can be known for each LPAR.

If too few reserved CPs were defined, you cannot

non-disruptively use Dynamic Capacity Upgrade on Demand

to implement an upgrade to an LPAR.

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

Change 23.012 -Typos in MXG 22.22 JCL examples:

COVERLTR -JCLTEST9: Line 130, WORKVOL=5 needs a comma.

JCLTEST8 Lines 166,168 need // in col 1 and 2.

JCLTEST9 Line 392 should be MXGSASV9

MXGSASV9 -JCLTEST8 and JCLTEST9:

MXGSASV9 //TESTQAPM step needs //QAPMJSUM DD DUMMY.

README -MXGSASV9: Lines 58 and 60, JCL needs //.

UPRINDOC -MXGSASV8 and MXGSASV9: the INSTREAM DD should be:

Feb 12, 2005 //INSTREAM DD UNIT=SYSDA,SPACE=(CYL,(1,20)),

Feb 22, 2005 // RECFM=FB,LRECL=80,BLKSIZE=0

Feb 23, 2005 The change to LRECL=72 caused unnecessary error

May 23, 2005 when UTILS2ER was run, and there's no value in

changing the historic LRECL from 80 to 72.

-Coverltr: Ref to Change 22.133 should be 22.123 for V9.

-README File on CD and the README member was missing an

equal sign: SPACE=(80,2,1)) SPACE=(CYL,(25,5)).

The PUT should be PUT 'c:\mxgcd\ieb....'.

-UPRINDOC: JCL example was missing ending comma, line 11.

Comments updated for example ASCII execution.

The program works, but produces an error; the

lines after the RUN; after %INCLUDE INSTREAM;

were debugging lines that should be deleted.

Thanks to Linda Piekarski, National Grange Mutual Insurance Co., USA.

Thanks to Sam Bass, McLane Company, USA.

Thanks to Jack Basile, PCH, USA.

Thanks to Francisco Ojeda, SAS Cary, USA.

Thanks to John Mattson, EPSON, USA.

Thanks to David Klein, DOITT, City of New York, USA.

Thanks to Sam Bass, McLane Company, USA.

Thanks to Brian Felix, Wachovia Corporation, USA.

Thanks to Jim Horne, Lowe's Companies, Inc, USA.
Change 23.011 MXG 22.22-22.12. DB2 variables QWHT../QWHU../QWHD./QWHA..

VMACDB2H from DB2 Trace, CPU, Distributed, Data Sharing Header may

Feb 11, 2005 be blank or missing; the extraneous statement at line 130

Feb 25, 2005 LENLEFT=LENGTH-COL+1;

May 20, 2005 must be deleted. Of course, they can also be blank when

those headers don't exist in a particular record, which

is completely normal! In addition, for SMF records that

are created from TMON/DB2 data, that statement caused the

INPUT STATEMENT EXCEEDED RECORD LENGTH, error: TMON SMF

records have headers at the start, rather than the end of

the SMF record (which is okay, since they are relocatable

sections with a "triplet" location pointers); it was MXG

recalculation of LENLEFT that was the cause of the error.

May 20,2005: This error also causes THREADTY variable to

be blank when it shouldn't be!

Thanks to Marty Pruden, Nestle Purina PetCare Co, USA.

Thanks to C. C. VanHook, DST Systems, USA.

Thanks to Steve Cunningham, DST Systems, USA.

Thanks to Craig Gifford, City of Lincoln NE, USA.
Change 23.010 Some fields that could contain '00'x are translated to a

VMAC119 blank value. RACFUSER, TTPOL, TTPROF, FCHOSTNM, FCM1,

Feb 8, 2005 FCRS, FCSUSER, and UDLPHIGH.

Thanks to Bruce Widlund, Merrill Consultants, USA.

Change 23.009 The RACF Database "Installation data" filed, INSTDATA,

VMACRACF can be $255, but only $40 was input.

Feb 8, 2005
Change 23.008 -Invalid ESS data that cannot be corrected in MXG caused

IMAC6ESS INPUT STATEMENT EXCEEDED. 0038x segment had Length=127,

VMAC6 but only 27 bytes of data, 0027x segment had length=0.

Feb 3, 2005 While IBM is contacted to resolve their errors, the only

circumvention is to insert MACRO STOPOVER MISSOVER %

as the first statement after //SYSIN. You will still see

MXGWARN: UNDECODED KEYS IN EXTENDED SMG 6 ESS DATA notes.

-A 0047x segment had 64 bytes when MXG expected 62, so the

size of FSSDATA field was increased to 99 bytes in this

change.

-The 0044x segment is supported creates ESSOFSTY variable.

Thanks to Dr. Alexander Raeder, Itellium, GERMANY.

Change 23.007 Five CPU variables in QAPMJOBL were incorrectly input as

VMACQACS milliseconds (&PD8.3) instead of microsedonds (&PD.8.6).

Feb 3, 2005 The "8.3" was changed to "8.6" for these variables in

the VMACQACS (with Change 22.371 as the last change):

JBCPU 5672

JBRUT 5736

JBSZWT 5757

JBTCPU 5780

JBMTXT 5792

Thanks to Paul Gillis, Pacific Systems Management Pty, AUSTRALIA.

Change 23.006 -Support for RACF Events 32-34,38-39,42-45,48,50,54-55,57,

EXTY8032 59,61-62, and 64 creates these

EXTY8033 these 18 new TYPE80A datasets:

EXTY8034 dddddd Dataset Description

EXTY8038 TY8032 TYPE8032 RACF EVT 32 CHANGE DIRECTORY

EXTY8039 TY8033 TYPE8033 RACF EVT 33 CHANGE FILE MODE

EXTY8042 TY8034 TYPE8034 RACF EVT 34 CHANGE FILE OWNERSHIP

EXTY8043 TY8038 TYPE8038 RACF EVT 38 INIT OF PROCESS

EXTY8044 TY8039 TYPE8039 RACF EVT 39 TERM OF PROCESS

EXTY8045 TY8042 TYPE8042 RACF EVT 42 MAKE DIRECTORY

EXTY8048 TY8043 TYPE8043 RACF EVT 43 MAKE NODE

EXTY8050 TY8044 TYPE8044 RACF EVT 44 MOUNT FILE SYSTEM

EXTY8054 TY8045 TYPE8045 RACF EVT 45 OPEN NEW FILE

EXTY8055 TY8048 TYPE8048 RACF EVT 48 REMOVE DIRECTORY

EXTY8057 TY8050 TYPE8050 RACF EVT 50 SET EFFECTIVE UID

EXTY8061 TY8054 TYPE8054 RACF EVT 54 UNLINK

EXTY8062 TY8055 TYPE8055 RACF EVT 55 UNMOUNT FILE SYSTEM

EXTY8064 TY8057 TYPE8057 RACF EVT 57 CHECK PRIVILEGE

IMAC80A TY8059 TYPE8059 RACF EVT 59 RACLINK

VMAC80A TY8061 TYPE8061 RACF EVT 61 IPCGET

VMXGINIT TY8062 TYPE8062 RACF EVT 62 IPCCTL

Feb 9, 2005 TY8064 TYPE8064 RACF EVT 64 CHKOWN2

-SMF 80 records with SMF80EVT=0 were discovered and were

being output to TYPE80CM dataset (which should always

have zero observations, since it is used only for events

that are not supported in MXG, and once you have one of

those events and send the SMF data, that event will be

supported!). IBM APAR OA03336 documents on condition

that causes SMF80EVT=0 and installing the PTF for that

APAR eliminated the RACFEVNT=0 observations.

Thanks to Peter Schubert, CITEC, AUSTRALIA
Change 23.005 An old VMXGSUM member in "USERID.SOURCLIB" caused message

VMXGSUM ERROR: APPARENT MACRO VGETENG NOT RESOLVED in BUILDPDB.

VMXGVERS (The INSTALL instructions remind you to remove all of the

Feb 2, 2005 VMXGxxxx & VMACxxxx members from your tailoring library

when you install a new version.) But even if you enable

diagnosic OPTIONS SOURCE SOURCE2 MACROGEN MPRINT; we can

only see members %INCLUDEd from your library; %VMXGxxxx

members are AUTOCALLed %MACROs, and they do not print the

source text on the log. To prevent these errors due to

a back-level %VMXG.... or %VGET.... macro, each of them

now invoke the %VMXGVERS macro that compares the member's

internal version number with the current MXG version, and

if they are inconsistent, prints this message:

MXGERROR: VMXGxxxx IS BACK-LEVEL AT NN.MM, ....

reminding you to remove that back-level VMXGxxxx member.

Thanks to Bill Whitehead, Experian, USA.

Change 23.004 Support for SAMS Vantage "LSPACE" record subtype, creates

EXSAMLSP SAMSLSPC dataset.

IMACSAMS

VMACSAMS

VMXGINIT

Feb 1, 2005

Thanks to Christelle Abily, GICM, FRANCE.
Change 23.003 Unused Change.
Change 23.002 Comments only. Description of the arguments for SYSTEM=

ANALAVAI and APPn= were confusing, and have been clarified.

Feb 1, 2005

Thanks to Allen O. Homonyom, Department of the Army, USA.

Change 23.001 NDM Subtype 'SY' caused INPUT STATEMENT EXCEEDED RECORD

VMACNDM error; the length in NDMSYOPL included itself, which was

Jan 30, 2005 not documented, causing MXG to read one byte too many.

After the @; after the INPUT of NDMSYOPL, insert

IF NDMSYOPL GT 0 THEN NDMSYOPL=NDMSYOPL-1;

Thanks to Victor Chacon, Verizon Wireless, USA.

LASTCHANGE: Version 23.

=========================Member=CHANGE22================================

/* COPYRIGHT (C) 1984-2005 MERRILL CONSULTANTS DALLAS TEXAS USA */
MXG Version 22.22 is dated Feb 1, 2005, thru Change 22.378

Early MXG Version 22.22 was dated Jan 22, 2005, thru Change 22.366

MXG Version 22.13 was dated Jan 13, 2005, thru Change 22.350

First MXG Version 22.13 was dated Jan 12, 2005, thru Change 22.345

Final MXG Version 22.12 was dated Dec 11, 2004, thru Change 22.316

Second MXG Version 22.12 was dated Dec 10, 2004, thru Change 22.313

First MXG Version 22.12 was dated Dec 2, 2004, thru Change 22.308

MXG Version 22.11 was dated Oct 26, 2004, thru Change 22.277

MXG Version 22.10 was dated Oct 13, 2004, thru Change 22.264

MXG Version 22.09 was dated Aug 25, 2004, thru Change 22.227

Final MXG Version 22.08 was dated Aug 20, 2004, thru Change 22.219

Second MXG Version 22.08 was dated Aug 13, 2004, thru Change 22.207

First MXG Version 22.08 was dated Aug 12, 2004, thru Change 22.205

Final MXG Version 22.07 was dated Jul 25, 2004, thru Change 22.180

MXG Version 22.07 was dated Jul 23, 2004, thru Change 22.177

MXG Version 22.06 was dated Jun 15, 2004, thru Change 22.129

MXG Version 22.05 was dated May 22, 2004, thru Change 22.108

MXG Version 22.04 was dated May 2, 2004, thru Change 22.099

MXG Version 22.03 was dated Apr 5, 2004, thru Change 22.066

First MXG Version 22.03 was dated Apr 2, 2004, thru Change 22.063

MXG Version 22.02 was dated Mar 24, 2004, thru Change 22.055

MXG Version 22.01 was dated Mar 11, 2004, thru Change 22.036

First MXG Version 22.01 was dated Mar 10, 2004, thru Change 22.034

Final MXG Version 21.21 was dated Feb 11, 2004, thru Change 21.320

MXG Newsletter FORTY-FOUR was dated Feb 11, 2004.
Contents of member CHANGES:
Member NEWSLTRS (and the Newsletters frame at http://www.mxg.com) now

contain the current MXG Technical Notes that used to be put in member

CHANGES between Newsletters. New Technical Notes are now added (and

now dated!) in NEWSLTRS/Newsletters with each new MXG Version.

I. MXG Software Version 22.22 is now available upon request.

II. Incompatibilities and Installation of MXG 22.13.

III. Online Documentation of MXG Software.

IV. Changes Log

=======================================================================

I. MXG Software Version 22.22 is available upon request.

Major enhancements added in MXG 22.22
BUILDPDB 22.365 BUILDPDB now sets Condition/Return code of zero.

ASUMMIPS 22.354 Interval Capacity by Workload, MIPS and MSU used.

ASMTAPEE 22.366 MXGTMNT ML-32, has MEXIT=ON,XMEMF=ON,ARCV=ON

TYPE110 22.359 Support for CICS/TS 3.1 with no EXCLUDEd fields.

UTILEXCL 22.347 New CICS/TS 3.1 WBREPRDL/WBREPWDL/PGCRECCT supported.

TYPE7072 22.349 Negative PCTMVSBY/CPUMVSTM/SHORTCPs, SMF70CNF bit 6.

TYPE30 22.375 IBM error in CPUIFATM, MXG error in SRVTCBTM.

TYPETPF 22.374 Support for MQ Series in TPF operating system.

TYPEQACS 22.371 Support for OS/400 5.3.0, new QAPMSYS dataset.

TYPEVMXA 22.369 Support for z/VM 4.4 and 5.1 new data (INCOMPATIBLE).

Major enhancements added in MXG 22.13
BUILDPDB 22.342 TYPE115/TYPE116 are now in BUILDPDB, will cause error

if you have already tailored MXG to add 115/116s.

You must remove your 115/116 tailoring from EXPDBxxx

or you will get this error message if you miss this:

ERROR: DATA SET WORK.MQMLOG IS ALREADY OPEN ....

BUILDPDB 22.320 MULTIDD='Y' obs now combined in PDB.SMFINTRV.

BUILDPDB 22.326 Variable CPUCEPTM now deaccum in PDB.SMFINTRV.

TYPE110 22.312 Support for CICS/TS 3.1 (INCOMPATIBLE).

UTILEXCL 22.350 Support for CICS/TS 3.1 new fields, errors fixed.

ASUMUOW 22.336 MQMACCT/MQMACCTQ data can be added to PDB.ASUMUOW

TYPE70 22.325 "Short CPs" variable SHORTCPS created in TYPE70.

TYPE70PR 22.325 "Short CPs" variable SHORTCPS created in TYPE70PR.

TYPETNG 22.339 Major TNG enhancement - array sizes dynamically set.

TYPE74 22.334 Support for APAR OA06476 type 74 subtype 5 and 8.

ONLYINTV 22.326 Example to build only PDB.SMFINTRV/PDB.TYPE30_6.

TYPE6 22.321 Support for second format type 6 PrintWay record.

TYPE7072 22.340 Revision to support for varying IFAs online/offline.

IMAC6ESS 22.332 Support for GEPARMKY 0036, 0041, 0043, fix 0034x.

TYPEHIOM 22.331 Support for hIOmon File I/O Performance Monitor.

This is a Windows environment monitor that tracks

I/O activity to the filename and user.

MONTHDSK 22.343 "MONTHBLD" program to build MONTHLY PDB on disk.

Major enhancements added in MXG 22.12
MXG 22.12 corrects IRD errors introduced in MXG 22.10/22.11.

MXG 22.12 corrects TYPE6 errors introduced in MXG 22.10/22.10.

TYPE6 22.309 Final correction for type 6 INPUT EXCEEDED errors.

TYPE6 22.298 SMF 6 STOPOVER on PrintWay section - missing @;

TYPE7072 22.307 Negative CPU values for IRD - Required CHANGE.

TYPEHURN 22.304 Support for ObjectStar subtype 45 Page Sweep.

TYPE6 22.302 Support for VPS V1 R8.0 VPS-FAX data

TYPE102 22.294 Support for APAR PQ73385,PQ91101 for IFCIDs 217, 225

TYPE102 22.294 Support for APAR PQ87848 for IFCID 173

TYPECIMS 22.314 Support for Mainview IMS IMF 4.1.00 (NO CHANGES!).

UTILEXCL 22.313 Support for APPLNAME,CANDEXNM,CANDEXTY CICS segments.

TYPENSPY 22.312 Support for NetSpy Version 7.0 (COMPATIBLE).

TYPEQACS 22.311 Support for OS/400 5.3.0 CONF/DISK/POLL/JOBL data.

ASMRMFV 22.316 Enhanced support for RMF III VSAM files.

TYPETNG 22.291 Support TNG NT Platforms NTW400I, WNS502I, ZPP501I.

VMACSMF 22.300 Use of FTP access to read SMF MVS-to-MVS supported.

ASUM70PR 22.293 LP0xxxxx variables now populated with PHYSICAL's data

CICINTRV 22.288 Comments show how to create PDB.CICINTRV from SMF.

VMXGPRAL 22.287 Enhancement to use PROC FREQ, example for 102 "who".

TYPE80A 22.286 Numerous enhancements, multiple RACF segments, etc.

ANALSIZE 22.276 Utility to analyze size of SAS data libraries.

ASUM70PR 22.274 Vars TOTSHARE/TOTSHARC kept for orig/current weights.

Major enhancements added in MXG 22.11
MXG 22.11 is NOW required for z/OS 1.6 with IFA/zAAPs, and it HAS

been tested with SMF 30, 70, and 72s from a system with real IFAs!

MXG 22.09 and MXG 22.10 do correctly support z/OS 1.6 without any

zAAP engines, but actual test data uncovered MXG errors and IBM

undocumented fields (Changes 22.272, 22.262, 22.265) that caused

most of the IFA/IFE fields to contain invalid values.

Additional enhancements in 22.11:
TYPE30 22.272 Support for zAAP IFA engines.

TYPE7072 22.272 Support for zAAP IFA engines.

TYPE30 22.265 Support for APAR OA09118, adds CPUCOEFF to SMF 30s.

TYPE94 22.268 Support for VTS R7.3 additional statistics.

TYPECMF 22.266 Support for CMF Version 5504 User SMF (INCOMPAT).

VMACDB2H 22.270 22.08-22.10 only. DB2 V8.1 INPUT STATEMENT EXCEEDED.

TYPE71 22.269 22.07-22.10 only. LPAxxxx variables missing values.
Major enhancements added in MXG 22.10
MXG 22.10 supports z/OS 1.6, but only if there are no IFA engines.
Additional enhancements in 22.10:
TYPE7072 22.260 Support for z/OS 1.6 WITH IFA engines.

TYPEVMXA 22.240 Support for z/VM 4.4, INCOMPAT.

TYPENTSM 22.246 Support for NTSMF Release 2.4.7 COMPATIBLE.

TYPEXAM 22.245 Support for XAMAP Release 3.4 INCOMPAT.

TYPETDSL 22.249 Support for TDSLink product's user SMF record.

TYPEBETA 22.250 Support for BETA93 Release 3.5 subtypes 0-5.

TYPETMDB 22.235 Support for ASG/Landmark TMON for DB2 V4.0 (COMPAT)

SYSLOG 22.238 Preliminary support for SYSLOG file.

ANALGART 22.242 Example analysis for Gartner Group requests.

ASUMCIMS 22.241 Example summarization of the four IMF datasets.

ERRORASC 22.239 ASCII platform errors when incorrect SMF download.

ANALFLSH 22.236 New member tracks concurrent flash copies executing.

TRND.... 22.258 Symbolics &TRENDINP,&TRENDNEW,&TRENDOLD added.
Major enhancements added in MXG 22.09
MXG 22.09 supports z/OS 1.6, but only if there are no IFA engines.
See Change 22.221 and especially MVS Technical Notes in NEWSLTRS.
Major enhancements added in MXG 22.08
MXG 22.08 is required for safe execution with SAS V9.1.2 or V9.1.3.
While MXG 22.07 had critical revisions for SAS 9.1.2, more design

changes were discovered in V9.1.2 that required more MXG changes.

In addition, Syncsort's add-on product PROC SYNCSORT was found to

corrupt INFORMATs, causing fatal errors in BUILDPDB, but because

I could remove all MXG INFORMAT statements faster than they could

even examine their error, I believe you can safely use that add-on

with MXG code (but you'll need to check your own code for INFORMAT

and watch for their eventual revised version that will work with

SAS V9.1.2). Note, the errors are in the PROC SYNCSORT add-on

product (which prints "PROC SYNCSORT" on your SAS log if used).

We have not seen these errors with the Syncsort Sort product.

The Syncsort ticket number # SR387805 is open for PROC SYNCSORT.

The details of the MXG changes that support V 9.1.2 and V 9.1.3 are

in the change text of the below MXG changes, but for execution of

MXG under "MVS", the only critical changes required are to:

- Install MXG 22.08 and use MXGSASV9 and CONFIGV9 from 22.08, and

- Run the UTILS2ER utility against all of your source libraries to

see if any lines of your SAS programs conflict with S2=72 option

that was required to replace the previous S2=0 option in MXG.
Changes related to SAS V9.1.2 and MXG execution: