* copyright (C) 1984-2019 merrill consultants dallas texas usa
Yüklə 28,67 Mb.
|
|
set to the maximum SMF time found. Variable OPENTM is created only for a match-up. Close-without-open will have OPENTM=. and ID=64, while open-without-close will have OPENTM=. and ID=62 in the VSAMOPEN dataset. There are often four close records for each open; I guess the Data and Index are closed, and then when the Cluster is closed, there is a second pair for the Data and Index. This is work in progress, and the efficiency of the merge will be enhanced. One consideration is using the output dataset to measure availability of VSAM production datasets, so there will likely be revisions to ANAL6264. Thanks to Rachel Quiroz Holt, Fidelity Systems, USA.
UTILBHEX build a file of readable SMF records from an email of the May 4, 1998 SYSOUT print file of the SAS log that contains the hex dump of SMF record(s) that were printed by the "LIST" statement, or were printed when "INPUT STATEMENT EXCEEDED RECORD" errors occur. That automatic hex dump on the SAS log has been invaluable in resolving errors, incompatible records and new product validation, and I still rely on the dump, but debugging a problem in CICS type 110 SMF records with excluded fields plus optional segments was made much easier by writing this utility to create the record and then enabling the debugging logic in MXG code. While member SENDDATA provides instructions by which you can re-read the SMF data and create an SMF file on MVS to then be downloaded, zipped, and emailed, now, if you have an error and I need to see the record, you can simply save the SYSOUT as a file on MVS, download it and zip it on your PC, and then email it to me for diagnosis. Also, member SENDDUMP shows how to produce a hex dump of select SMF records, and now its output can also be converted to readable records by this utility.
IMACIMS the additional datasets created by JCLIMSLG's TYPEIMSA, VMACIMSA so you can control the contents of all IMS log datasets. Apr 30, 1998 New exit members EXIMSTRN, EXIMSBMP, EXIMSUNM, EXIMSMRG EXIMS07I, EXIMSFAS, and EXIMS598 (and similarly named _L and _K macros) exist for the IMSTRAN, BMPS, UNMATCHED, IMSMERGE, IMS07, IMSFASTP, and IMS5938 datasets.
VMXGINIT created and GLOBALed to continue the externalization of Apr 30, 1998 DDnames for datasets created by BUILDPDB. I have decided Jun 25, 1998 on the naming conventions for dataset LIBNAME macros: Exit Member Name Dataset Name Dataset &Macro EXTYxxxx TYPExxxx &PDBxxxx EXdddddd All others &Pdddddd All VMACs that process SMF records will have their &PDBxxxx or &Pdddddd "Dataset &Macro" defined and set with a %LET in VMXGINIT member. All resolve to "PDB" by default, except for the CICSTRAN and DB2ACCT (high volume) datasets that are not sorted by MXG's BUILDPDB. See Change 15.320 for further discussion.
IMACICDA CICS TS 1.2 have two new optional segments added to their IMACICOC type 110 record, CANMQ with MQ Series counts and duration IMACICOM and CANWLMSC with WLM Service Class names. These new IMACICOW segments are now supported in new members IMACICOM and VMAC110 IMACICOW respectively, and member IMACICDA was updated to Apr 30, 1998 add their include in the default order (MQ before WLMSC), and the new variables were added to the KEEP= list in the VMAC110 member (but no new variables are created unless you remove the comment block in IMACICOM or IMACICOW). As always, you need to run UTILCICS to find out which of the segments exist and in what order in each region, and and might have to update IMACICDA if UTILCICS shows the order is different in your records than MXG expects. Since I now have the DSECTS, I found that there are two bytes of flags in the OMEGBSC segment that are now two variables, OMDBWRN (DB Warnings) and OMGENWRN (General warnings) added to CICSTRAN if you enable IMACICOC. Thanks to Melanie Floyd, UPS, USA. Thanks to Jon Taf, UPS, USA. Thanks to Dennis Hancock, UPS, USA.
CONFIG used so that sites that have their own macro libraries in Apr 28, 1998 their SASAUTOS collection can be found under MXG. The earlier option was just SASAUTOS=SOURCLIB, which prevented you from finding your macros in your SASAUTOS. (Many sites concatenated their macro libraries and/or the SAS.MACRO library to their //SOURCLIB DD.) Thanks to Alfred Mueller, ABN - Amro, USA. Change 16.065 The NPM format MG028VA decoded an undocumented '0A'x FORMATS value. IBM has confirmed the '10'x documented for the Apr 28, 1998 LVMATYP field should be '0A'X, so FORMATS now contains: 10='0AX:PCT TIME IN BLOCKED RATE' instead of 16='10X:PCT TIME IN BLOCKED RATE'. Thanks to Paul Hill, Midland Bank, ENGLAND. Change 16.064 In dataset TYPE72GO, variable RESPAVG is kept and EXETTM VMAC7072 is created and kept. RESPAVG is equal to AVGELPTM, but Apr 28, 1998 keeping RESPAVG, which was the name in TYPE72, will help Aug 14, 2001 sites migrating to Goal Mode. The Total Transact Execution time, R723CXET, is now kept in new variable EXETTM. Previously, only the average value AVGXETTM was kept in TYPE72GO, and it was calculated only if TRANSEXC was non-zero (because I thought EXETTTM was updated only when TRANSEXC - when a transaction in execution phase had ended). But Brenda found non-zero EXETTM with zero count in TRANSEXC, and Don Deese's current paper confirms that EXETTM is non-zero for batch service classes, even though TRANSEXC is always zero for batch. So now, not only is EXETTM a kept variable, but also AVGXETTM is calculated as EXETTM/TRANSEXC if there were subsystem transactions terminating, or else AVGXETTM=ELAPSTM/TRANS if there were no TRANSEXC but TRANS was non-zero.
and TSO), variable EXETTM now measures in Goal Mode what was measured in ELAPSTM in Compatibility Mode. In debugging my change, (examining the missing value notes on the log), I found a line with PCTUSNDI that did no harm but is now deleted, and I discovered that replacing: SUMPTRAN=SUMPTRAN+RTSTRN(M); /*ADDING*/ with SUMPTRAN=SUM(SUMPTRAN,RTSTRN(M)); /*SUM FUNCTION*/ not only eliminated the missing value notes on the log, but also saved CPU time! While SUMPTRAN (the accumulated transaction count as I loop across the thirteen response count buckets) was always set to zero before the loop, if a service class had no response buckets, RTSTRN(M) was missing, and the loop ran thru all thirteen buckets, because the ADDING zero to missing is still missing. Instead, using the SUM function, zero plus missing is zero, so the test that follows (IF SUMPTRAN GE NRPTRANS) is now satisfied on the first iteration, so we bail out of the loop, instead of DOing it thirteen times! Thanks to Brenda Rabinowitz, Prudential Securities, USA. Thanks to Carole Storby, LMCO, USA. Change 16.063 Support for OS/400 Version 4.2.0 (COMPATIBLE). VMACQAPM -New QAPMCONF configuration GKEY values of 'CI' and 'SJ'. Apr 27, 1998 -New QAPMJOBS variables JBTCPU,JBTHDF,JBTHID,JBTHAC, JBTHCT,JBMTXT, and JBSTSF were added. The QAPMJOBS file is now LRECL=812 vice 617. -The +52 after JBSJFG was changed to +58. This change corrects 4.1.0 variables after JBSJFG. -New QAPMDISK variables DSBGDR,DSBGDW,DSBGS,DSCOMP, DSFGDR,DSFGDW,DSFGRE,DSFGS,DSFGWE,DSLBA,DSLBW,DSPBA, DSPBCO and DSPBU were added. The QAPMDISK file is now LRECL=346 vice 267. The variable PCTIOPBY seems to be invalid, as the DSIDLC field contains very large values. This needs to be examined further, and will be updated if I learn more. -New QAPMECL variable EMDUP was added. -New QAPMETH variable ETMDUP was added. -New variable XIADRN was added to QAOMIOP1-4 datasets. Except for possible LRECL changes in your JCL, the 4.2 records were compatibly changed. Thanks to Clark Jennings, Reynolds Metal, USA. Thanks to Stuart Burnett, Reynolds Metal, USA.
ASMRMFV support for the CSR, the Common Storage Remaining data May 10, 1998 segments. This level 002 of ASMRMFV has been tested, but I have not yet updated the TYPERMFV code for the CSR. Change 16.061 The UTILCICS output listed UN98309 for any region, but UTILCICS that APAR only applied to CICS 4.1; the fields added by Apr 27, 1998 that APAR are in CICS TS by default, so the test to set PTF='UN97309' now also tests SMFPSRVR=41.0. This CICS dictionary-reading utility also now knows about the new CANMQ and CANWLMSC sections and sets the ERRMQ=64 error if CANMQ segments of only 64 bytes are found. The early Candle documentation showed 64 bytes, but the correct length is 76; if you get ERRMQ=64, then your CICS guru needs to correct the MCT length for CANMQ to be 76 bytes. Thanks to Melanie Floyd, UPS, USA. Thanks to Jon W. Taff, UPS, USA. Change 16.060 Sterling's Connect Direct (formerly NDM) subtype "PT" VMACNDM process termination data is trashed; timestamp variables Apr 27, 1998 NDMSUBTM and NDMSKDTM get INVALID VALUES messages, and other PT variables are also wrong. Sterling has an open problem report, item PI-754129, but no fix yet, but now called Sterling CASE 7916. In 1994 there was a similar error in "PT" in Change 11.326, but this time I found Sterling Tech Support and received same-day service!. But there was also an MXG error, as the INPUT for the PT segment was off by two bytes, and there was a new variable, NDMSEQ, added at the end that was overlooked. Also, to prevent those INVALID VALUE messages before you get a fix from Sterling, I inserted the ?? modifier for each packed PT field. The MXG fix to the PT input: change LOC=LENGTH-21; to LOC=LENGTH-23;. Thanks to Brian Crow, BT, ENGLAND. Change 16.059 -Variable NWMSINRT was not kept in dataset MSEXCHIS due to VMACDECS mispelling as NWMSINRJ in the KEEP=list. VMACNTSM -Repeats of six variables in KEEP= list in VMACDECS were Apr 27, 1998 removed, though they caused no problem. Thanks to Freddie Arie, Lone Star Gas, USA.
ASMIMSL5 assembly were corrected and the revised member was tested Apr 27, 1998 with IMS 5.1 log records without error. Thanks to George Ellard, FEDEX, USA. Change 16.057 -The label for variable OMVSOPR in dataset TYPE30OM built VMAC74 by BUILDPDB was wrong, because dataset TYPE74OM also has VMAC78 a variable OMVSOPR, and the 74's label replced the 30's. Apr 27, 1998 To resolve the conflict, the type 74 variable OMVSOPR was renamed to OMVSOPRP to protect the TYPE30OM variable, which could cause VARIABLE OMVSOPR NOT FOUND, but only if you use the TYPE74OM data set AND explicitly referred to OMVSOPR in your SAS report code. -The label for variable SYSNAME in member VMAC78 was used for all BUILDPDB datasets, but it had a typo. The correct spelling is IEASYSXX and not IESAYSXX. Thanks to Randy Shumate, LEXIS-NEXIS, USA. Change 16.056 Comments and INVOKEBY= were changed to ASUMCICX in this ASUMCICX optional replacement for ASUMCICS (see Change 15.205). Apr 25, 1998 Thanks to Jim Ray, Branch Banking and Trust, USA. Change 16.055 Archaic FMXGUCBL failed under SAS 6.09, but re-assembly FMXGUCBL of the member with the SETSSI AF010000 (as mentioned in Apr 25, 1998 Change 12.015), eliminated the 0C4, and by changing TXTUC15 DC X'0015',X'0001',X'0003' CODE 15 - UNIT ADDRESS to increase its length from three to five bytes: TXTUC15 DC X'0015',X'0001',X'0005' CODE 15 - UNIT ADDRESS eliminated the DYNAMIC ALLOCATION FAILURE FOR VOLUME xxxx with REASON-021C error on each volume. The SETSSI statement now sets AF010000 vice AF000000. This member is archaic in that DCOLLECT is now the MXG recommendation, and before that, ASMVTOCs had replaced the very old FMXGUCBL. However, it still works and fixing it with a re-assembly was easier than re-writing the existing reports to use ASMVTOC or DCOLLECT! Thanks to Karl Schlichting, CSC New Zealand, NEW ZEALAND.
VMACTMO2 records that were converted to V 8.1 format, member Apr 25, 1998 TYPETMO2 (which is for native V 2.0 unconverted format) will tell you that you should be using member TYPETMO8. Thanks to Diane DePasquale, CSC TMG Warren, USA.
ASUMPRTR ASUMPRTR now %INCLUDEs IMAC6 and IMAC7072 to protect the TRND72 _LTY6 and _LTY72 macro names when ASUMPRTR is run as part VMACIPAC of the MXG QAJOB. TRND72 had RESPSTD located after the VMACTRSN the ending parenthesis, causing syntax error. Apr 9, 1998 VMACIPAC syntax error. IF CC=. THEN IF .... changed to IF CC=. AND IF .... in four lines. VMACTRNS was deleted again; it was a misspelling that was removed from 15.15 but crept back into the first 16.01.
VMACIPAC record inserted two bytes to change YY to YYYY in four of Apr 8, 1998 their five subtypes. This change tests variable VERSION to be NE 'V-1', but I have not been able to verify if the record version is actually going to be useable to detect the changed record format. At worst, you might have to use a different SMF record number for the RDS 6.1 record, and change the MXG test in VMACIPAC that reads IF VERSION NE 'V-1' in four places to IF ID=nnn. V-1 is in current version records, but is also shown in the Mobius copybook for 6.1, so I don't know if they actually changed the VERSION field or not. Look for an update to this change text on our homepage once 6.1 records have been created. View Direct 6.1 is the new name of RDS / IPAC product. See revisions in Change 16.085. Thanks to Clark Jennings, Reynolds Metal, USA. Change 16.051 Variable RESPSTD, the standard deviation of the average TRND72 response of ended SRM transactions, is now created in the TRND72GO TRND72 and TRND72GO trending. The default MXG summary Apr 8, 1998 level of PERFGRP or SRVCLASS can be changed to break out each PERIOD, by simply adding variable PERIOD at the end of the SUMBY= statement in those members, if you want to see individual periods' data (so you can trend TSO period one, for example). I cannot safely change my default to a lower level without potentially destroying somebodys reports, but you can change it for your site and then add PERIOD to your reports. Thanks to Chuck Hopf, MBNA, USA. Change 16.050 PROC PRINT's that have multiple pages per BY group are ANALCNCR messed up due to a SAS problem, a conflict between the Apr 8, 1998 NOBYLINE parameter (which permits us to make reports prettier) and the existence of an ID statement with the same list of variables as the BY statement (and we used BY &SUMBY; ID &SUMBY; to make the reports even prettier). SAS Usage note V6-PRINT-6788 discusses the error, found in SAS 6.07, but there will be no fix sooner than SAS Version 7. So, while the reports are prettier still with both the NOBYLINE parameter and the ID statement, I have commented the two ID &SUMBY; statements, so you still get pretty reports, but now they will print All report pages! Thanks to Glenn Bowman, Wakefern Food Corporation, USA
TYPEMON8 Version 8.1 format by TMON utility program TMONCNVT are Apr 7, 1998 now Year 2000 compliant, but the change was INCOMPATIBLE so you will need MXG 16.01 or later is you still convert your new TMON records to the old format. Landmark also replaced the TMMDCCLK field which was a series of packed fields (0962821314185700 was 96282 13:14:18.57) with a new value (B043AA3E81B0F701 now for April 6, 1998), which causes INVALID DATA FOR D12 messages because D12 is no longer a valid packed field, and since I have no clue from Landmark documenting this change (I have only a dump of the record causing that message), and since it is late at night and I want to wrap up MXG 16.01 before I sleep, variable TMMDCCLK will be a missing value in these new Y2K-compliant-8.1-format records, and the INVALID DATA message is eliminated. Fortunately, TMMDCCLK is an internal datetimestamp that was not even kept by MXG. Thanks to Tim Downs, WCI/CSC Warren Ohio, USA. Thanks to Mike Cezarro, City of Rochester, USA. Thanks to Kumar Thavakumar, City of Rochester, USA. Change 16.048 -Support for NTSMF new Object SNA Logical Unit Sessions EXNTMSES creates new SNALUSES dataset with the same variables as EXNTSNAL in the SNA Connection Object (dataset SNACONN). EXNTSNAR -Support for NTSMF new Object SNA 3270 Response Times EXNTWIDE creates new SNARSPTM dataset with five response buckets IMACNTSM R3270TH1-R3270TH5 with percentage of responses (whatever VMACNTSM that is) in each bucket. However, the values are the Apr 8, 1998 same for each SNACONNM value, so I believe these data are still wrong in NTSMF 2.1.0 (which had to deal with the people writing this object not following the rules to write data correctly!). -Support for new MS Exchange Event Services object creates new MSEXCHES dataset. -Support for the SNA Adapter SnaDlC1 (yep, that's a lower case l, a upper case C and the numeral one!) outputs to the existing SNAADAPT, but there is no longer an SNAADAPT name field. I reused the same dataset because there never were any NTSMF records created with the original name field + counters. -Support for the WideToMBErr object creates new dataset WIDE2MBE. -The _BNT macros needed these variables added to the BY list for the dataset sorts: _BNTPROC has PID added, _BNTNETI has INTRNAME added, and _BNTMSMC has CONNNAME added. Thanks to Jim Quigley, Con Ed, USA. Change 16.047 Variable JCSPTOD in CICS Journal Segment was validated VMAC110 for Shared Medical System's journal data, but only with Apr 7, 1998 CICS Version 2 data. The CICS/ESA logic in VMAC110 still had JCSPTOD ?? PDTIME4., which produced wrong values. The logic from V2 journal was copied into the CICS/ESA journal processing logic. This error impacts you only if you were decoding your own journal segments. Thanks to John Croasdale, Barclays Bank Group, UK. Change 16.046 New ASUM member to summarize TYPE42DS (SMS Interval ASUM42DS I/O activity) by dataset to report total I/O to each Apr 6, 1998 dataset from all users, counting the number of users using that dataset during each interval. Thanks to Simone Niemczura, FISERV, USA.
EXNTSMTP Version 4.0 in INCOMPATIBLE with Version 3.0; the two EXNTWEBS counter objects Cache Size and Cache Used, MXG variables IMACNTSM CACHSIZE and CACHUSED in dataset IISG, were removed by VMACNTSM the new 4.0 version of IIS. Apr 3, 1998 -Support for Active Server Pages object's nearly complete restructure, with 14 fields kept from prior version, with 19 new fields added, and these 10 fields deleted: ASTHPOCU ASRQCURR ASRQERRT ASRQTOEX ASRQTONQ ASCOMFAI ASMEMUSE ASMEMFRE ASRQBREX ASRQTOTQ -Support for new FTP Service Object replaced the previous FTP Server Object, but MXG outputs into the same FTPSERV dataset that was built from the prior FTP Server Object. The only difference between the old and new objects is that the new FTP Service Object has an instance field, new MXG variable FTPSERVR (name of the FTP Server), that did not exist in the old FTP Server object. -Support for new Web Service Object creates new WEBSERV dataset; this new Web Service Object logically replaces the previous HTTP Service object that created MXG HTTP dataset, but as many HTTP fields no longer exist in the new Web Service Object, and as there are many new fields, the new WEBSERV dataset must be used in place of HTTP. -Support for new SMTP Server Object creates new SMTPSERV dataset. Thanks to Leigh Ann Payne, Wachovia Operational Services, USA. Change 16.044 The JCL example had //DB2ACCT DD pointing to the DB2ACCT ANALDB2C dataset, but the default in IMACDB2 creates PDB.DB2ACCT, Apr 2, 1998 so the example now has //PDB instead of //DB2ACCT. The newer combination of ASUMUOW and ANALUOW are really the better tools for matching CICS and DB2 transactions. Thanks to Alfred Holtz, Merck Medco, USA. Change 16.043 -ASUMIDMS was touched up. Variable TASTTYPE and TASKTYPE ASUMIDMS now have their length set as $1 and @26 respectively, in TESTUSER LENGTH statements instead of FORMAT statements. The TRNDIDMS -TESTUSER was touched up. The hardcoded IDMSxxx dataset Apr 2, 1998 names were replaced by the _LIDMxxx macro names. Jan 19, 1999 -TRNDIDMS is a new contribution from Alan, based on the TRNDCICS member of MXG, and provides IDMS trending. -Updated Jan 19, 1999: In ASUMIDMS, the line that read TSKTIMUS=TASTIMWT; was changed to TSKTIMUS=TASTIMUS; Thanks to Alan Deepe, Perot Systems, ENGLAND. Thanks to Richard S. Ralston, Whirlpool, USA.
VMACRACF created by IBM utility IRRDBU00, variable OMVSPROG was Apr 2, 1998 not created in dataset RACF0270, but now it is, being INPUT @1050 as a $CHAR200 variable. Thanks to Randy Shumate, LEXIS-NEXIS, USA.
UTILCONT two places and lost an &, so this new utility to print Apr 2, 1998 the contents (in mega-bytes and obs) of SAS datasets in a SAS Library fails. Change both PDBMXG to PDB, and add the missing & to %LET PDB=%UPCASE(&PDB); . Yüklə 28,67 Mb. Dostları ilə paylaş:
|