* copyright (C) 1984-2019 merrill consultants dallas texas usa
Yüklə 28,67 Mb.
|
|
%VMXGGOPT(DEVICE=xxxxxxxx); instead of using SAS's GOPTIONS DEVICE=xxxxxxxx command. All devices are now supported by modifying the background and color lists. All GRAFxxxx routines use the PCBATCH device driver (a modified IBMPCGX driver) with background color of black and a color list of: WHITE,RED,GREEN,BLUE,YELLOW,CYAN,PINK A new parameter HARDCOPY= has been added. If you specify HARDCOPY=Y, then BLACK and WHITE are reversed. See the SAS Technical Note in MXG Newsletter 25 about the problems moving graphs between mainframes and other SAS platforms; using VMXGGOPT to specify devices should help minimize those problems. Change 11.172 If you use WEEKBLD to create your weekly PDB on tape, it JCLWEEKT can cause multiple tape mounts on the WEEK DD name if you WEEKBLDT have more weekly data than will fit on one tape. The new Sep 23, 1993 member WEEKBLDT should be used instead, as it uses the logic described in MONTHBLD (it creates each dataset in a temporary DASD library (DDname=TAPETEMP) in tape format, and then writes each dataset from TAPETEMP to DDname=WEEK using DISP=MOD so as to eliminate rewinds and remounts, which also reduces the elapsed run time of the job. Member JCLWEEKT has been modified to use this new member. Note that the DDNAME of TAPETEMP must be allocated only as big as the largest weekly dataset (usually STEPS or CICSTRAN or DB2ACCT). Thanks to Eric Huning, Republic Insurance, USA Change 11.171 Zero observations in TYPE72MN for MVS/ESA 4.2 or earlier VMAC7072 occurs because the test added by changes 11.016 was only Sep 22, 1993 tested with MVS/ESA 4.3 data, and variable VALDSAMP only is non-missing for 4.3. The test to create observations in TYPE72MN only for valid intervals, located just before the %%INCLUDE SOURCLIB(EXTY72MN... statement: IF VALDSAMP GT 0 THEN DO; must be changed to IF VALDSAMP GT 0 OR VALDSAMP=. THEN DO; Thanks to Dennis Smith, Compuware Corporation, USA. Change 11.165 Reserved Change Number. ==Changes thru 11.164 were included in MXG PreRelease 11.06 dtd 1Sep93= (Note in 11.06 there were two change 11.162's - the second is now 163!) Change 11.164 Complete rewrite of GRAFRMFI. Will only run under 6.07 GRAFRMFI and later. Significantly more graphs are produced and ZGRAFRMD more flexibility is provided. New parameters have been Aug 31, 1993 added to allow overriding the first line of the headings with your company name and to allow you to "clean up" old graphics catalogs in your PDB. See the documentation in the member. If you are still running version 5, the old GRAFRMFI is kept in member ZGRAFRMF. Thanks to Chuck Hopf, Primerica, USA.
VMACTCP for TCP/IP API and for FTP and Telnet clients, and now Aug 30, 1993 uses SMF record type 118. In theory, your TCP/IP tech sets subtype values (using the LOGGING command statement in PROFILE TCPIP, see APAR PN40511), but both the API and FTP/TELNET server records have the same ID and subtype! Thus MXG does not use the subtype value, instead deducing from other fields to determine actual data: Dataset Contains TYPETCPF FTP Server - existed previously in MXG TYPETCPC FTP Client TYPETCPT TELNET Server- existed previously in MXG TYPETCPL TELNET Client TYPETCPA API Calls (Sockets) Thanks to Richard Warren, Salomon, Inc, USA. Thanks to Bob Muller, Grumman, USA. Change 11.162 Support for ASTEX Release 1.7 Interval Database SMF data VMACDMON record. Variables RIRCX and RNVCCNT were added to all 3 Aug 30, 1993 DMON datasets, and variables RVDISCNH,RVLATNCY,RVMSAMPL, RVSAMPWH and RVTLOST were added to DMONVOL. Additionally the label for RHUSTX is now MUST PROBLEM I/O COUNT. I now realize that I used an "H" in some variable names that were really spelled with an "M", but the Release 1.5 doc was a smudged fax! While having M for MUST/MAY would be better, changing now will cause your report code to have to be changed, and I don't intend to make extra work for you now, so we'll live with the MXG misspellings! And RHTCSW should be RHTCFW, but I've left it alone, too. Release 1.7 added these new fields compatibly, so that prior versions of MXG will not fail; they just won't have the new variables.
VMACDB2 all tests GE 2.3 were changed to GT 2.2 to avoid the VMACDB2H exposure. The specific occurrence affected only the DB2 VMAC102 correlation decoding in IMACDB2, but prudence dictated a IMACDB2 change to all occurrences. Aug 30, 1993 Thanks to John Shuck, SunTrust, USA.
BUILDPDB 1476 DD segments) did not correct JELAPSTM. In extensive BUILDPD3 notes, this contributor located the several changes to BUILD005 carry MULTIDD5 from the 30_5 into the final logic so that IMACPDB JELAPSTM will non-missing in this instance. Aug 30, 1993 Thanks to Andy Vick, Allied Dunbar Life Assurance, ENGLAND.
VMACNSPY again. Now, LEGENT says the field is at offset +34 and Aug 27, 1993 not +33 as previously documented, and so the MXG test Previously: IF NSUBNTCN GT 0 AND NSUBNTLN GT 0 THEN DO; Ch 11.130: IF LDLANLST GT 0 THEN DO; Now Is: IF LDLANMAX GT 0 THEN DO; Thanks to Gary Allen, LEGENT, ENGLAND.
VMAC90 it conflicted with variable ACTIVE in TYPEAPAF, which is Aug 27, 1993 numeric. ACTIVEMN is the identity of the Active SMF data set, and this should have minimal impact on TYPE90 users, but avoids a conflict of variable types. Thanks to Bryant Osborn, Department of Transportation, USA. Change 11.157 IMS Appended Log processing of type 36 log record did not ASMIMSLG recognize that a '04'X bit in QLDQFLGS indicates the true Aug 27, 1993 dequeue of a transaction. (The '04'X indicates the free of a queue-block as opposed to the '01'x which indicates a CNT (terminal) dequeue. Find AP TYPE36CT,PAKEDONE and insert after that line: TM QLDQFLGS,QLDQFFQB BZ DROPMAP Thanks to Siegfried Trantes, Gothaer Versicherungsbank VVAG, GERMANY. Change 11.156 Support for Release 4.5 of Software AG's COM-PLETE SMF VMACCOMP record (which was incompatibly changed; MXG 11.06+ is now Aug 24, 1993 required). New variables were added to the existing data sets. Still in test is verification of PIB4.2 formats. Thanks to Jill Hansen, South Dakota Regents Information Systems, USA.
IMACPDB in PDB.JOBS, PDB.STEPS, and PDB.ACCOUNT, even when you Aug 24, 1993 specified more in member IMACACCT. This was a holdover from earlier MXG versions, and is no longer correct nor needed. All occurrences of ACCOUNT1-ACCOUNT3 and of LENACCT1-LENACCT3 in member IMACPDB were changed to ACCOUNT1-ACCOUNT9 and LENACCT1-LENACCT9, and then the number of account variables will be controlled by the number you keep in member IMACACCT. Thanks to Esther Ader, TRW Redi Property Data, USA. Change 11.154 MXGTMNT abended at only one site, with a 0C4. An invalid ASMTAPEE value for the UCBASID in the UCB Common Extension caused Aug 24, 1993 this single error, which can be corrected by: a.Relocate the line MVC DEVASID,UCBASID from after the line LH R4,UCBASID to before that LH. b.Insert after line USING ASVT,R5 these two lines: CL R4,ASVTMAXU IS ASID LARGE? BNL NOJOB Thanks to David Froberg, Software AG, USA.
VMAC110 SMF records are invalid. If you must process those TMON Aug 24, 1993 records, you can change SUBTYPE=SMFPSSTY; to SUBTYPE=1; and as long as ONLY Landmark type 110 records are in your SMF file, MXG will successfully create its CICS datasets. This is a circumvention for you to apply if you need it, but there was no change to member VMAC110 in MXG. Thanks to Amy Stammerjohn, IBM WSC, USA. Change 11.152 TYPE70 dataset now supports CPUIDs of 0 thru 15, because VMAC7072 it is possible (in PR/SM,MLPF,or MDF) to assign CPUIDs RMFINTRV of 9 thru 15. Sites which had assigned CPUIDs above 8 Aug 23, 1993 found the wait time for those CPUIDs above 8 was lost, causing an inflated CPU utilization percentage during periods of low activity. Not only does this change protect current users of high CPUID values, but also it protects all users if/when more than 8 concurrent CPUs are supported. Variables CAIn, CPUSERn, CPUWAITn, CPUEDTMn, CPUPDTMn, IORATEn, PCTCPBYn, PCTTPIn, VFAFFTMn, and VF0Nn, now exist with n values of 0,1,2,3, 4,5,6,7,8,9,A,B,C,D,E and F. (RMFINTRV was changed only to keep CPUSER9,A-F variables). Thanks to Ron McAllister, Hitachi Data Systems, USA. Change 11.151 Variables DCUSYSID and DCUTMSTP were not kept in the new VMACDCOL SMS Construct datasets DCLODC, DCOLSC, DCOLMC, DCOLBC, Aug 23, 1993 DCOLSG, DCOLVL, DCOLAG, DCOLDR, and DCOLLB, but now are! Thanks to John Rosza, Depository Trust Company, USA. ==Changes thru 11.150 were included in MXG PreRelease 11.05 dtd 20Aug93= Change 11.150 MXG redesign for execution under any SAS platform began CONFIG with this major structural change, but fortunately, as CONFIG07 long as you use the CONFIG member from this library for CONFIG08 your CONFIG= value, the change is transparent under the VMXGINIT currently supported platforms. (Note that members DOC CONFIG, CONFIG07, and CONFIG08 are now identical.) Oct 4, 1993 Now with this change, all of MXG can be stress tested and validated for execution under the "ASCII" versions (WINDOWS, OS/2, UNIX). MXG's BUILDPDB has already been executed under SAS 6.08 under Windows, OS/2, and UNIX. The original text of this change was revised and is now included in the Technical Note "Executing MXG and Other SAS Applications Under Windows, OS/2, and UNIX" in MXG Newsletter TWENTY-FIVE.
TYPEDOS START and STOP occur on the STOP date, but there is no Aug 10, 1993 START date in the record, and most DOS event records also have no elapsed duration by which true START date can be known, so the date of STARTIME can be incorrect. For the DOSJOBS data, however, the elapsed time can be estimated (CALCTM) and can be used to correct the date of start. (Before this change, MXG wrongly recalculated STOPTIME instead of STARTIME, causing both to be wrong when STOP and START occurred on different dates.) After: CALCTM=SUM(CPUTM,OVERTM,WAITTM)/3; Replace the two lines recalculating STOPTIME with IF CALCTM GT TIMEPART(STOPTIME) THEN STARTIME=DHMS(-FLOOR(CALCTM/86400),0,0,STARTIME); Thanks to J. Geluk, ROBECO Group, THE NETHERLANDS Change 11.148 SAP journal data for release 5 is supported in MXG 11.03. IMACICSA For SAP release 4.3, the time part of STRTTIME is wrong, Aug 8, 1993 because SAP 4.3 used PDTIME4. while SAP 5.0 uses PIB4.3. Sites with release 4.3 need only to change the input for STCTIMTR from PIB4.3 to PDTIME4. and the value of STRTTIME will be correct; remember to back out the change when you convert to SAP release 5.0. Unfortunately, SAP does not store any indicator of release in the record! Thanks to Waldemar Schneider, SAS Institute Europe, GERMANY.
EXTYODSD User SMF records. Three datasets are created: EXTYODSR ODSDEVIC - Device Statistics. Physical activity between EXTYODSU the ODS and the ODI. FORMATS ODSREPOR - Report Termination. Control Unit information IMACODS about the report just recorded. TYPEODS ODSUSER - Report Statistics. Reflects all activity for VMACODS that report for that user. Elapsed times Aug 5, 1993 will be totals of all page activity while that report was open. Thanks to Cindy Ng, Charles Schwab, USA.
EXTYSAR by the SARSRQU3 exit, written when a report is printed FORMATS from the SAR database. One dataset is created: IMACSAR SARSRQU3 - SAR Print Report from SAR database. TYPESAR VMACSAR
Aug 4, 1993 Thanks to Bob Mattingly, ARCO, USA. Change 11.145 ASUM70PR was still wrong in MXG 11.03; the BY statement ASUM70PR BY SYSTEM STARTIME; (after the MERGE statement) Aug 4, 1993 that was supposed to have been added by Change 11.041 was lost in a later change to ASUM70PR. If you have PR/SM or MDF or MPLF data with some MVS systems partitioned and some native, the absence of this BY statement can corrupt the PDB.RMFINTRV dataset. Insert it after the MERGE! Thanks to Randy Shumate, Mead Data Central, USA. Change 11.144 MXG CIMSTRAN dataset (from Boole's IMF) did not contain a VMACCIMS flag variable that the transaction was a BMP, which some Jul 29, 1993 users needed to separate BMP from MPP transactions. New variable BMP now exists. Thanks to Vladimiro Viano, Assicurazioni Generali, ITALY.
ANALDB2R if reports PMSTA01 or PMSTA02 are requested, only under Jul 27, 1993 SAS 6.08, is caused by SAS Usage Note 6719. The INCODE arguments of PMSTA01/PMSTA02 have 330+ lines each,which exceeds a SAS internal limit for macro argument length as discussed in the Usage Note. Most of the 300+ lines are "faker" statements IF A=. THEN A=.; whose purpose was to eliminate "UNINITIALIZED VARIABLE" messages on the log. The simple circumvention is to simply delete all of those "faker" statements and tolerate a note on the log: After the second INCODE, delete the first 313 lines. After the third INCODE, delete the first 311 lines. The permanent MXG solution is to replace the "faker" statements with a RETAIN list-of-variables 0; statement which takes many fewer lines. There are three INCODE arguments, but only the last two INCODEs must be changed. Thanks to Ken Rowe, Resolution Trust Corp, USA. Thanks to Ivan Vasquez, Resolution Trust Corp, USA. Change 11.142 VM/ESA duration variables that were INPUT as PIB6.6 must VMACVMXA be stored in LENGTH 8 for the deaccumulation, otherwise Jul 27, 1993 the subtraction of these very large numbers can cause small truncation of values (6557.83 seconds became 6556.56 of VMDTTIME). The truncation seems to have been the cause of small negative values of VMDCTIME as well! These variables are now length 8: CALMPNTR CALMPTRV CALQDNTR CALUPNTR CALUPTRV PFXPRBTM PFXTMSYS PFXTOTWT PFXUTIME PLSVFOTM PLSVFVTM RDEVRTPD SRMRSCTM SRMSTRD SRMTSHOT SRMTSLIC VMDESLIC VMDTTIME VMDVFOTM VMDVFVTM VMDVTIME Thanks to Stan Laugher, NM Computer Services, AUSTRALIA. Change 11.141 POOL/DASD User SMF record caused INPUT STATEMENT EXCEEDED VMACPOOL error. The INPUT of PLD $CHAR3. must be changed to Jul 27, 1993 $CHAR4. Thanks to Mike Lambert, IHS, USA ==Changes thru 11.140 were printed in MXG Newsletter TWENTY-FOUR 2Aug93 ==Changes thru 11.140 were included in MXG PreRelease 11.03 dtd 26Jul93= Change 11.140 The Asynchronous Data Mover Facility APAR OY65142 adds to IMACPDB the type 30 SMF record variables ADMFREAD,ADMFWRIT with VMAC30 count of pages moved by ADMF. Both variables were also Jul 24, 1993 added to the PDB.JOBS and PDB.STEPS datasets. Change 11.139 A RACF segment with zero length RACFDLN was not expected VMAC80A nor protected for, causing INPUT STATEMENT EXCEEDED ... Jul 24, 1993 error. Change SELECT (RACFTYPE); to read IF RACFDLN GT 0 THEN SELECT (RACFTYPE); Thanks to Jeff Burnett, Anixter Information Services, USA.
VMAC110 accomplished by inserting INPUT @LOC @; after both of Jul 23, 1993 the occurrences of LOC=LOC+JCRLL; In particular, this will permit MXG to skip over SAP Journal records until you install the MXG 11.03 needed for SAS Release 5.0! Thanks to Thomas Lowin, GESIS, GERMANY. Change 11.137 An extraneous close-parenthesis on line 011800 (after the VMACZRB _LZRBUWP statement) must be deleted. Jul 23, 1993 Thanks to Ray Dickensheets, Yellow Freight System, USA. Change 11.136 Major revision to the subtype 102/103 Omegamon/CICS User VMACOMCI SMF record (Clocks and Counters for VSAM, DL/I, IDMS, Jul 22, 1993 ADABAS, SUPRA, and DATACOM) in Candle's SPE QOC0553. Now, two MXG datasets are created for each DB manager, one with total I/O activity, and one with detail activity at the "File/Database" level. Candle's misdocumentation of the VSAM record pre-SPE was also corrected, as were MXG errors in inputting as PIB4.4 what was PIB4.6! Datasets: OMCIVSAM,OMCIVSAT - VSAM Detail, Total OMCIADA ,OMCIATAT - ADABAS Detail, Total OMCIDLI ,OMCIDLIT - DL/I Detail, Total OMCIIDMS,OMCIIDMT - IDMS Detail, Total OMCISUPR,OMCISUPT - SUPRA Detail, Total OMCIDTCO,OMCIDTCT - DATACOM Detail, Total Thanks to Thomas Lowin, GESIS, GERMANY. Thanks to Ross Pover, Immigration & Ethnic Affairs, AUSTRALIA. Change 11.135 The _K macro name for the _LTY80CM dataset was misspelled VMAC80A as _KTY8007 instead of _KDY80CM. Jul 22, 1993 Thanks to Thomas Lowin, GESIS, GERMANY. Change 11.134 MXG 11.02 only. Changes 11.113 and 11.089 were not in TYPEVM the code, although both were documented as being there. VMAC26J2 Labels were missing for CHANIND,CHANINDY,CHANINDY vars in VMAC73 TYPE73 dataset. Jul 22, 1993 Thanks to Norbert Korsche, OMV-AG, AUSTRIA. Change 11.133 STOPX37 undocumented fields VOLSER and MSGCODE were found VMACX37 by this user. Add: @441+OFFSMF VOLSER $CHAR6. Jul 22, 1993 @497+OFFSMF MSGCODE $CHAR7. and add MSGCODE to KEEP= list. Thanks to Lou Sassani, National Australian Bank, AUSTRALIA Thanks to John Astle, National Australian Bank, AUSTRALIA Change 11.132 MXG 11.02 only. Change 11.119 was incomplete. Insert: VMAC37 BRFDATLN=LENDETT; Jul 22, 1993 after INPUT @OFFDETT @; Thanks to Rod Fernandes, Albert Heijn B.V., HOLLAND. Change 11.131 HSM BCDS dataset MCB records were incorrect (incomplete VMXGHSM and too few observations, plus INVALID ARGUMENT TO FUNCT Jul 22, 1993 DATEJUL message). Line 114200 should be P2=P2+64; vice P2=P2+(I*64); In addition, lines 113500 & 113800-114100 were relocated above the DO loop. Thanks to Willi Weinberger, Gothaer Versicherungsbank, GERMANY. Change 11.130 Legent LANSPY Problem # DGL249 acknowledges documentation VMACNSPY error for extended subtype D record that caused INPUT Jul 22, 1993 STATEMENT EXCEEDED RECORD LENGTH. The MXG test: IF NSUBNTCN GT 0 AND NSUBNTLN GT 0 THEN DO; must be replaced with IF LDLANLST GT 0 THEN DO; Thanks to Mike Hicks, Centrefile Ltd., ENGLAND. Change 11.129 SAS Error 76-322 occurred when unnumbered and numbered CONFIG07 statements were input, and CONFIG07 used.SAS Option S2=0 Jul 22, 1993 should have been S2=72 in CONFIG07. (CONFIG08 was okay.) See Change 10.109 for discussion. Thanks to Ann Wheeler, American President Lines, USA.
TYPETMS5 TMS.DSNBRECD so that the type of tape device is known. Jul 22, 1993 Thanks to Chuck Hopf, Primerica, USA. Change 11.127 In MXG 11.02. an extraneous period in column 1 of line ASUM70PR 031200 must be removed. Additionally, in some instances, Jul 21, 1993 the new LPnDUR variables were missing because of sort reordering. To correct, LPARNUM was added to the BY statement following INCODE= and FIRST.STARTIME was changed to FIRST.LPARNUM. ==Changes thru 11.126 were included in MXG PreRelease 11.02 dtd 6Jul93= Change 11.126 Type 30 APPC variables were accumulated across steps, and VMAC30 a second step (PGM=IEFBR14) that did not use APPC would Jul 2, 1993 contain non-zero APPC counts if the first step used APPC! APAR OY63634 now creates an additional segment in type 30 records which contains the "Delta" values for the step. MXG supports the new APAR, and will use the Delta segment if it exists; otherwise, the cumulative values will be kept (as occurs now) until you install the APAR. Change 11.125 RMF-like reports from MXG datasets were enhanced with the ANALRMFR EMIF channel reporting of both total Channel Busy and the Jul 2, 1993 Partitions part of total Channel Busy. All reports now match their RMF 4.3 counterparts. Change 11.124 Variable LSBACS (ACS ID) was added to the STCLMU dataset: VMACSTC Replace the line INPUT @21+OFFSMF LSBMON PIB2. Jul 1, 1993 With these two: INPUT @543+OFFSMF LSBACS PIB1. Oct 21, 1993 @21+OFFSMF LSBMON PIB2. (The above text was reworded for clarity from that that was published in Newsletter TWENTY-FOUR, Oct 21, 1993.) Also, Change 10.116 was not implemented correctly. After that change, the code should read: ELSE INPUT STC07CON PIB4. @; INPUT STC07LBL $CHAR1. ... Thanks to Rodney Marwick, Vereinte Versicherungen, GERMANY. Thanks to Tom Yarbrough, Rockwell Space Operations Company, USA. Change 11.123 The order of INPUT statements was modified to make all VMAC64 TYPE64 variables available to the TYPE64X (extent) data Jul 1, 1993 set, so that you could optionally add desired variables into TYPE64X (using _KTY64X macro in member IMAC64). Thanks to Rodney Marwick, Vereinte Versicherungen, GERMANY.
VMACOPC variables MTDRESTA,MTDDIRES,MTDRERUT, and MTDDIRER were Jul 1, 1993 added to OPC24D_C data sets. This change supplements Change 11.065. Thanks to Rodney Marwick, Vereinte Versicherungen, GERMANY.
VMACNSPY Insert after ARSPNET calculation (line 123700): Jul 1, 1993 AOUTSZT=COTPUTSZ/NETRSPNO; and then after ARSPNET=.; (line 126700), insert: AOUTSZT=.; Thanks to Joanne Turpie, Department of Labour, NEW ZEALAND. Yüklə 28,67 Mb. Dostları ilə paylaş:
|