* copyright (C) 1984-2019 merrill consultants dallas texas usa
Yüklə 28,67 Mb.
|
|
JCLCPORT and vice versa on z/OS and ASCII. Jun 4, 2018
ADOCRMFV SHOWSP subroutine after Change 36.068 (MXG 36.04 only). ASMRMFV Extended Function VSAM datasets support striping and Jun 2, 2018 compression and extended addressability. It is an attribute of the Data Class. LISTC ENT('dataset') ALL command will display EXTENDED attribute if file is EF. -Common RMFV030I and RMFV031I messages now issued for either EF or non-EF VSAM data sets. -Documentation Section updated to support the above changes: Section 12 "Messages" Thanks to Randy Shumate, RELX Group, USA
TYPENMON the maximum length of the input records is small. Adding TYPERHEL LENGTH X $1926; reduced 111 CPU seconds to only 40, so a TYPSNMON _NULL_ data step is added to TYPENMON to find the maximum TYPSRHEL record length, and the INPUT NMONTEXT $VARYING32000.; is VMACNMON replaced with this logic, using the _INFILE_ variable: VMACRHEL LENGTH NMONTEXT $ &NMONLENGTH ; Jun 5, 2018 NMONTEXT=TRANWRD(_INFILE_,',,',', ,'); It is the actual LENGTH of NMONTEXT that is the major impact on CPU and Elapsed times, but on z/OS, the LRECL has some impact, so you need to use an LRECL that is greater than the MAXNMONLENGTH, printed on the SAS log. The calculation of the NRWORDSIN that was needed for SAS V8 and early WPS was revised with added CPU reduction. The same changes are made for the RHEL/NMON processing. Steve Bagshaw gets credit for this discovery! Thanks to Steve Bagshaw, ITMetrics, ENGLAND. Thanks to Steve McCulloch, TMX, CANADA. Change 36.108 Support for RACF TOKENs REQTCRE and ADMINCII creates VMAC80A TOKMADMINCII='TOKEN*ADMINCII' Jun 1, 2018 TOKMREQTCRE='TOKEN*REQTCRE' variables in TYPE80TK dataset. Thanks to Bruce Hewson, Citibank N.A., SINGAPORE.
ASUMUOW data then it becomes unlikely that you will get a good VMXGUOW match between CICSTRAN and DB2ACCT. Further with some May 31, 2018 of the more recent changes in VMAC110 other than the class 3 wait times and counts from DB2ACCT there is not a lot of information added to ASUMUOW from the DB2ACCT data. To suppress the use of DB2ACCT in your ASUMUOW invocation all you need to do is to modify the _LDB2ACC substitution macro to point to _NULL_ as shown in this code: %LET MACKEEP=%QUOTE( MACRO _YESOBS % MACRO _NOOBS % MACRO _LDB2ACC _NULL_ % ); OPTIONS SOURCE SOURCE2; %INCLUDE SOURCLIB(VMXGUOW); _NOOBS
OPTIONS NODSNFERR NOVNFERR; _SUOWCIC /* SORT CICSTRAN DATA */ _SUOWDB2 /* SORT DB2 DATA */ _SUOWMQ /* SORT MQ SERIES DATA */ _SUOWSPN /* CREATE ASUMUOW DATASET */ %VMXGUOW; _YESOBS OPTIONS DSNFERR VNFERR; The new VMXGUOW drops the DB2ACCT variables with the above suppression, keeping only 99 in the new PDB.ASUMUOW, previously there were 144. Only comments were added in ASUMUOW with this example. Change 36.106 TYPE42DS Encryption variables were INPUT but not KEPT nor VMAC42 labeled nor formatted: May 31, 2018 S42AMRIB='S42AMRIB*BYTES*READ' S42AMWIB='S42AMWIB*BYTES*WRITTEN' S42AMRBD='READ BYTES*DECRYPTED*OR ELIGIBLE' S42AMWBE='WRITE BYTES*ENCRYPTED*OR ELIGIBLE' S42AMRCI='VSAM*CI-S READ OR*PHYSICAL*BLOCKS' S42AMWCI='VSAM*CI-S WRITTEN*PHYSICAL*BLOCKS' With the large number of TYPE42DS observations, if you want to only select datasets with encryption counts: %LET MACFILE= %QUOTE(IF ID=42 THEN DO; IF SUBTYPE=6; END; ); %LET MACKEEP= %QUOTE(
MACRO _ETY42DS IF S42AMRBD GT 0 OR S42AMWBE GT 0 THEN DO; OUTPUT _WTY42DS; END;
% ); %INCLUDE SOURCLIB(TYPE42);RUN; which will only populate TYPE42DS when bytes GT zero. Syntax note: The original MACFILE syntax suggested %LET MACFILE= %QUOTE( IF ID=42 AND SUBTYPE=6; ); is fine for ONLY the TYPE42 program, but if that was used with BUILDPDB, only the 42.6 would be read, hence the above, safer selection will pass all other SMF records in case you want to process other records. Thanks to David Cogar, Wells Fargo, USA. Thanks to Scott Barry, SBBWorks Inc., USA.
TYPESYSL was added in 34.04 renamed TYPESYSX/TYPSSYSX/VMACSYSX due May 24, 2018 to conflict with TYPESYSL dataset created by TYPETMNT.
May 23, 2018 Change 36.103 Format MGSTCCS for variable STC11CSP in STCVSM11 dataset FORMATS has new value '8000'x='8000X:VSM6 FICON CHANNEL'. May 23, 2018 Thanks to Randy Hewitt, DXC, USA. Change 36.102 DB2 V11 APARS PI71903,PI84045,PI82755 added offsets _SC, VMAC102 _PR, _INC, and _SQL to populate those fields that were May 22, 2018 previously only in DB2 V12. MXG test changed to GE 11.1. The timestamp variable QW0376TS in old data was invalid (e.g. '1A6CE0BD12FCB083'x, a date in 1914!) and was set to a missing value; now whatever is there is input so it may still be incorrect. Thanks to Joachim Sarkoschits, DATEV, GERMANY. Change 36.101 Support for NDM-CDI OP (Operator Clist Record) creates EXNDMOP NDMOP dataset. VMACNDM VMXGINIT
May 21, 2018 Thanks to Michael Oujesky, DTCC, USA. Change 36.100 ACF2 Version 6.2 circumvention in Change 36.075 exposed VMACACF2 another STOPOVER as LENLEFT was not correctly calculated. May 23, 2018 Thanks to Jim Horne, Lowe's, USA. Thanks to Mohammed Naseer, Lowe's USA.
VMAC80A variable in TYPE80TK dataset. May 23, 2018 Thanks to Coen Wessels, IBM, GERMANY Change 36.098 Format MGMOCTY, used for Information Builder's FOCUS, FORMATS has two new values for BEGIN and END. Only FORMAT was VMACFOCU changed, no change was made to VMACFOUU. May 17, 2018 Thanks to Tim Hare, Hare Systems, USA.
VMAC110 report there are excluded fields is corrected to 365/3260 May 14, 2018 from 373/3356 (which had included optional fields). Thanks to Paul Maradin, DXC, USA Thanks to Larry McCulley, DXC, USA
ANALACF2
May 14, 2018 Thanks to Scott Barry, SBBWorks Inc., USA. Change 36.095 The %LET MXGABND=nnnn; option to abend instead of error VMACBBMQ is added to the BBMQ processing. See Change 21.384. May 11, 2018
IMACSMFF WPS Version 4, under investigation, but add the comments May 9, 2018 as shown here to circumvent: /* COMMENT OUT - DEFAULT BREAKS WPS V4 '2047.000'='2047.000:MAX VALUE AND LABEL NO SUBTY' '2047.001'='2047.001:MAX VALUE LABEL WITH SUBTYPE' END COMMENT */ The default IMACSMFF is always executed when SMF is read. Change 36.094 MXG 35.12-36.04. If you use IMACFMTS to add your site's FORMATS own FORMATs, the RUN; statement in member FORMATS after Apr 10, 2018 the VALUE $MGRMVOS statement should NOT have been added. Thanks to Robert Debartolo, Cognizant, USA. Change 36.093 CICS Dispatcher Statistics CICDS dataset, DSGTMADQ was VMAC110 too large; field is now input &PIB.4.2 with two decimals. Apr 9, 2018 Thanks to Scott Barry, SBBWorks Inc., USA. Change 36.092 ACCTSORT=NO was not working as advertised. The datasets READDB2 ended up in WORK rather than PDBOUT. Apr 9, 2018
VMXGUOW resulting in many UNITIALIZED variable messages. VMXGUOW May 7, 2018 now checks only for the existence of the dataset and if it does not exist sets it to _NULL_. Change 36.090 If you tailored IMACDB2 to redefine MACRO _Lxxxxxx's and ANALDB2R you specified only a single level name, so those datasets May 3, 2018 are written to WORK, and did not specify a PDBOUT, you could get this syntax error ERROR: THE FUNCTION COMPBL REFERENCED BY THE %SYSFUNC OF %QSYSFUNC MACRO FUNCTION HAS TOO FEW ARGUMENTS. due to incorrect logic, now corrected, in ANALDB2R. Originally posted to MXG-L as possible issue with SAS V9.3 to V9.4 migration, the thread was updated/corrected. Change 31.104, MXG 31.03, May 2013, created the exposure; the user's good run was with MXG 31.01. Thanks to Dennis Longnecker, State of Washington Courts, USA. Change 36.089 APAR OA54884 for z/OS 2.3 ONLY reports very high I/O EXCP DOCUMENT counts in EXCPTOTL (Address Space Total, SMF30TEX) that May 1, 2018 was observed in the MASTER address space, but could occur in any address space. Change 36.088 SAS Note 51008 Java versions 1.6/1.7/1.8 can cause errors DOCUMENT ERROR: The Java proxy could not create a new xxxxxxxx. May 1, 2018 ERROR: shmag() failed in Java extension rc -1 errno 124 ERROR: Unable to attach current thread. on z/OS. That Note the circumvention is to add this line JREOPTIONS=( -Djava.lang.ClassLoader.lazyInitialization=false) to your SASHLQ.CONFIG(SITE) configuration PDS member. SAS Support reported SAS does not support Java 8 yet; see also SAS Note 51195. Change 36.087 Unused Change Number. Change 36.086 DCOLLECT Encryption Variables are now kept in DCOLDSET: VMACDCOL indicates if the LCU contains at least one FICON channel. Apr 24, 2018 DCDTYPE ='ENCRYPTION*TYPE' DCDKLBL ='ENCRYPTION*KEY*LABEL' The IBM Documentation does not provide DCDTYPE values to decode. These fields were added by z/OS 2.3. Thanks to Mike Creech, Black Knight, USA.
VMAC78 indicates if the LCU contains at least one FICON channel. Apr 24, 2018 Thanks to Lane Thorne, Honda of America Manufacturing, USA. Change 36.084 Dataset STCVSM11 variables added by Change 34.237 were VMACSTC incorrectly labeled and inconsistent, now corrected: Apr 23, 2018 STC11NHR='HOST*INTERFACE*I/OS' STC11NHW='HOST*INTERFACE*CUBUSY*DURATION' STC11NRR='REMOTE*INTERFACE*I/OS' STC11NRW='REMOTE*INTERFACE*CUBUSY*DURATION' STC11NIR='IP*INTERFACE*I/OS' STC11NIW='IP*INTERFACE*CUBUSY*DURATION' Thanks to Giuseppe Giacomodonato, EPVTECH, ITALY.
UTILBLDP and BUILDPDB=NO so you can create your own custom dataset Apr 27, 2018 with control of variables, etc. This example creates the PDB.SMFHEADER dataset with four variables kept from every SMF header: %UTILBLDP(USERADD=ID 118,BUILDPDB=NO, EXPDBVAR=PDB.SMFHEADER (KEEP=SYSTEM SMFTIME ID SUBTYPE), EXPDBCDE=OUTPUT PDB.SMFHEADER;, OUTFILE=INSTREAM ); %INCLUDE INSTREAM; -Unrelated, unprintable '08'x character introduced 35.09 is removed. Thanks to Randy Hewitt, DXC, USA.
VMACDB2 of RIO/SPP/DPP/LPP with DIO/LIO/RIO/SIO. Apr 19, 2018
ANALID the ANALID report. FORMATS -Format $MGSMFID text shifted one byte to the right; a few IMACSMFF record descriptions lost 1 character to keep 37 maximum. VMACID -VMACID,VMACSMF formats are now SMFIDSUB $8. SMFIDCH $4. VMACSMF increasing the LENGTH of those variables by one byte, Apr 23, 2018 format 7.3 references are changed to 8.4 for IDANDSUB. Apr 27, 2018 -Unfortunately, if you have used IMACSMFF to label your user SMF Record Descriptions, you will need to replicate all and insert a blank at the beginning of each existing 3-character record type, to match the new example in that IMACSMFF member: '2047.001'='2047.001:MAX POSSIBLE VALUE AND LABEL' -One line summary report with total records and bytes and the time range of the input SMF file is added.
VMXGGETM where nnnn is the SMF Record Type (max is now 2047) and Apr 18, 2018 where mmm is the subtype. VMXGGETM creates an output SMF file with example records of each selected type. Change 36.079 -Support for new SMF 119 subtypes 24, 38, 39, 40, and 45. EXT11924 creates these new datasets: EXT11938 dddddd Dataset Description EXT11939 T11924 TYP11924 TNPROFILE EXT11940 T11938 TYP11938 SmcdLnkStats EXT11945 T11939 TYP11939 SmcdLnkStart IMAC119 T11940 TYP11940 SmcdLnkEnd VMAC119 T11945 TYP11945 IsmStats VMXGINIT Untested with data. Apr 18, 2018 -New BitRate variables added to TYP11906 dataset: Apr 26, 2018 IFINBITRT='INBOUND*BITS PER*SECOND' IFOUBITRT='OUTBOUND*BITS PER*SECOND' IFBITRATE='TOTAL*BITS PER*SECOND' -CO:Z subtypes 192 and 193 are validated with data. Change 36.078 z/OS, SAS 9.4 M3 with IBM DFSORT, ABEND 0C4 in SASVZSR1, CONFIGxx when sorting a large dataset. SAS notes 57676 and 58629 Apr 17, 2018 circumvent the error with these options // EXEC MXGSAS94,OPTIONS='SORTBLKMODE SORTBLKREC=5000' which could alternately be specified in your CONFIGxx, but SORTBLKMODE has been the SAS Default for years. The SORTBLKREC option is not yet documented by SAS. http://support.sas.com/kb/57676 http://support.sas.com/kb/58629 This is documentation only, no code was changed. Change 36.077 With a BY statement in your VMXGSUM INCODE, there is no VMXGSUM guarantee that the data order will be correct, and if you Apr 14, 2018 also %LET MXGSUMCLASS=YES or CLASSNWAY to YES, the data May 8, 2018 step may fail. VMXGSUM now looks at the first word in the INCODE= and if it is BY sets CLASSNWAY to NO.
VMAC110 DURATM when DSGTWT was greater (Change 35.264), but that Apr 14, 2018 should only have been done for SMFSTRQT='INT' as DURATM doesn't exist in the 'REQ', 'USS', nor 'EOD' records. Thanks to Paul Volpi, UHC, USA.
VMACACF2 MXG tests for the last release, 6.2, but new ACF2 record Apr 14, 2018 has '00'x instead of '62'x in byte 119, causing MXG test for 6.2 to fail. This change forces ACSMFREL=6.2 if it is zero for this INCOMPATIBLE CHANGE to the ACF2 record. The CA fix is PI24126 and a reassembly of DMGSMF exit. Thanks to Michael K Yuan, Navy Federal Credit Union, USA.
VMACBETA reversed, OFF was ON and ON was OFF, bit test corrected. VMACBE97 -TYPEBE97 subtype 31 revised: like TYPEBETA, there can be Apr 17, 2018 an R1 and R2 value for each FIELDNAME, but TYPEBETA PUT May 15, 2018 the text value into a character variable for R1 and R2, but those values were then difficult to test. TYPEBE97 instead creates nine pair of variables with true values (like Dates, Times, HEX, etc). Thanks to Andreas Menne, Finanz Informatik, GERMANY.
VMAC7072 SMF70MAXPU='CORES*PHYSICALLY*AVAILABLE*THIS MODEL' Apr 12, 2018
VMAC99 PSERV ='SERVICE*DURING*INTERVAL' Apr 12, 2018 PISERV ='ZAAP*SERVICE*DURING*INTERVAL' PSSERV ='ZIIP*SERVICE*DURING*INTERVAL' TIME_AT_PDP_USING='TIME AT*PDP USING*SAMPLES' TIME_AT_PDP ='TIME AT*PDP*ACCUMULATOR' PCT_USING_PDP ='PCT*TIME*USING*SAMPLES' SMF996_FLAGS ='SMF996_FLAGS' EWLM_LOCAL_PI ='EWLM*LOCAL*PI' EWLM_GLOBAL_PI ='EWLM*GLOBAL*PI' SMF996EWLM ='EWLM*MANAGED?' SMF996IOPR ='I/O*PRIORITY?' SMF996INEL ='ZIP*INELIGIBLE?' SMF99_NUM_EXT_SC='EXTERNAL*SERVICE*CLASSES' Thanks to Randall Schlueter, First Data, USA.
VMACIAM short segment lengths IAMIAINL=148 (MXG expected 204) and Apr 8, 2018 IAMIASTL=148 (MXG Expected 204) were encountered, and now protected for these IAM 9.2 records. Thanks to Paul Naddeo, FISERV, USA.
VMXGDUR message said that was correct. It was looking for THREEHR Apr 8, 2018 but will now accept THREEHOUR, EIGHTHOUR, or TWELVEHOUR.
VMXGCICI values in CICS Statistics SMFSTREQ='USS','REQ',or 'EOD' Apr 6, 2018 records as the DURATM only exists in the 'INT' records. But using the DIF(COLLTIME) a pseudo DURATM is created and used to populate/correct DSGTWT and DURATM.
ADOCRMFV -A new RMFBSAM record with an MXG01 id is now output ASMRMFV for every successfully processed RMF III VSAM data set. VMACRMFV VSAM attributes and statistics are included as well as Apr 6, 2018 many ASMRMFV statistics and counters, and the record is May 5, 2018 output in new dataset ZRBAS1. May 9, 2018 -Between the existing MXG00 record and the new MXG01 record nearly all information on an ASMRMFV log is captured. The MXG01 data becomes the ZRBASMDS (?) data set in the result PDB. -MXG01 records are only created for RMF III VSAM data sets that open and close successfully. There are no MXG01 records generated for: Empty VSAM data sets (VSAM considers this an error) VSAM data sets that are not an RRDS type VSAM data with an invalid CISIZE for RMF III data VSAM data with an invalid LRECL for RMF III data Non-VSAM data sets The above conditions have been flagged in the ASMRMFV Log for a long time. -Two new parameters UPCASE/NOUPCASE control the handling of values assigned in keyword=value usage. -UPCASE (alias UC) is the default and provides the same behavior as in prior ASMRMFV versions which force all PARM and SYSIN (or alternative) input data to upper case internally. -NOUPCASE (alias NOUC) is the default and does not alter any values assigned to a keyword. And thus lower case values can be assigned to a keyword. -However, for most (if not all) data filters currently supported by ASMRMFV only upper case values are accepted. For example, Sysplex Ids, System Ids, Job Names, Job Classes, and so on are all required by IBM syntax rules to be in upper case. Lower case values are flagged as errors by ASMRMFV validation routines. -NOUPCASE is a feature primarily intended for future filtering enhancements where lower case values could be accepted. -NOINDEXES and/or NOSPACE parameters might not work correctly in all situations and this has been corrected. Messages were not always suppressed when they should have been. -Message RMFV105I produced for ASMRMFV Detail and Summary reports now shows the full 5 character RMF III table id instead of just the first 3 characters. This change was needed to distinguish MXG00 and MXG01 output record statistics. -Minor changes to messages RMFV037I, RMFV041I, RMFV051* (*= S,E,W,I), and RMFV106W. -Several documentation Sections are updated to support the above changes: Section 5 "Input Data Selection Parameters" Section 12 "Messages" Section 31 "Summary" Section 32 "Bibliography"
CONFIGxx to be upper cased, but MXG recommends NOCAPSOUT for ODS, Apr 5, 2018 and the SAS default on z/OS is NOCAPSOUT. I don't know why I changed the SAS default years ago, but "CAPSOUT" is removed from all of the example MXG CONFIGxx members, so your site's default value will be used. Change 36.066 Support for "IBM Developer for z Systems IDZ" SMF 122 EXTY122A subtype 1 record creates new dataset TYPE122A. TYPE122A IMAC122A is used because there is a TYPE122 record (that is/was?) TYPE122A previously written by Tivoli Allocation. The Product Name TYPS122A The Product Name field contains 'C2AE'x before and after VMAC122A the name, where 'AE'x is the ASCII registered copyright VMXGINIT symbol, but 'C2'x is a Danish A with a ring above! Both Apr 4, 2018 are printed on ASCII SAS, but both are blank on z/OS, and on z/OS lower case characters are converted to upper case by the $ASCIIn. INFORMAT. Thanks to Tory Lepak, Aetna, USA.
VMACQACS and are added to QAPMDISK dataset, transparently. Apr 3, 2018 MXG created the PCTCLEAN and DSFSMAPBY variables: DSFSMAPSZ ='FREE SPACE*MAP 4K*PAGES*COUNT' DSFSCLEAN ='CLEAN*4K PAGES*FREE SPACE*COUNT' DSFSCLEAN0='LEVEL 0*CLEAN BLOCKS*PAGES 1-7' DSFSCLEAN1='LEVEL 1*CLEAN BLOCKS*PAGES 8' DSFSCLEAN2='LEVEL 2*CLEAN BLOCKS*PAGES 16' DSFSCLEAN3='LEVEL 3*CLEAN BLOCKS*PAGES 32' DSFSCLEAN4='LEVEL 4*CLEAN BLOCKS*PAGES 64' DSFSCLEAN5='LEVEL 5*CLEAN BLOCKS*PAGES 128' DSFSCLEAN6='LEVEL 6*CLEAN BLOCKS*PAGES 156' DSFSFRAGIX='FREE SPACE FRAGMENTATION INDEX' DSFSDIRTY ='DIRTY*4K PAGES*FREE SPACE*COUNT' DSFSDIRTY0='LEVEL 0*DIRTY BLOCKS*PAGES 1-7' DSFSDIRTY1='LEVEL 1*DIRTY BLOCKS*PAGES 8' DSFSDIRTY2='LEVEL 2*DIRTY BLOCKS*PAGES 16' DSFSDIRTY3='LEVEL 3*DIRTY BLOCKS*PAGES 32' DSFSDIRTY4='LEVEL 4*DIRTY BLOCKS*PAGES 64' DSFSDIRTY5='LEVEL 5*DIRTY BLOCKS*PAGES 128' DSFSDIRTY6='LEVEL 6*DIRTY BLOCKS*PAGES 256' PCTCLEAN='PERCENT*CLEAN*PAGES IN*FREE SPACE' DSFSMAPBY='FREE*SPACE*SIZE*MGBYTES'
SMF MANUAL MXG Version 36.03, except new SMF 122, which is not in Mar 30, 2018 that SMF Manual.
VMACDB2 datasets listed in Change 35.299, which revised deaccum Mar 29, 2018 logic and expected one minute statistics intervals, but that IBM Change to force the DB2 Statistics Interval to one minute wasn't introduced until DB2 Version 10! Thanks to Don Blaszka, Wipro Limited, USA. Change 36.062 Further VXBYUSR logic revised to use only 2 decimals for VMACVMXA all _MT1 DIF() functions; these data have only two digit Yüklə 28,67 Mb. Dostları ilə paylaş:
|