* copyright (C) 1984-2019 merrill consultants dallas texas usa
Yüklə 28,67 Mb.
|
|
ods csvall body='procprintcsvall.csv'; ods markup tagset=chtml body='procprintchtml.html' (title= 'This Text Identifies Your Content.'); title 'Leading Grain-Producing Countries'; proc print data=grain_production; run;
ods csvall close; ods markup tagset=chtml close; CHANGE 36.155 Support for Z/VM LINUX LNXAPPL Process & summary APL data EOAPLLXP creates these new datasets: EXAPLLXF dddddd dataset description EXAPLLXP APLLXF VXAPLLXF LNXAPPL FILE SYSTEM DATA EXAPLLXS APLLXS VXAPLLXS LNXAPPL SUMMARY DATA IMACVMXA APLLXP VXAPLLXP LNXAPPL PROCESS DATA VMACVMXA -The VXAPLLXP process dataset only outputs a process that VMXGINIT had non-zero TOTAL_TIME or CTOTAL_TIME; if you want all Aug 25, 2018 process records to be output, edit member EOAPLLXP into Sep 13, 2018 your tailoring library and remove the conditional test. -SAMPTIME and UPTIME are not valid values, investigating. -Crypto Type 12:CEX6C is now recognized and output in the VXPRCAPM dataset. -Dataset VXPRCMFC is now populated for z14, CSVN=5. -Sep 13: RESID corrected to RESIDENT, comments deleted.
IMACAAAA CICS SMF 110 CICSTRAN segment and new NRXENTRY variable. IMACICWY UTILEXCL
VMAC110 Aug 18, 2018 Change 36.153 New parameter added that can be used to reverse the GRAFWLM order of data on the charts produced. GRAFWLM has GRAFWRKX always produced bar charts with the most important Aug 24, 2018 work at the top and the least important (discretionary) at the bottom. Now this parameter HIGHTOLOW (default is YES, original order) if set to NO reverses the order. -Old default YES has UNCAPTURED at top and DISCRETIONALY at bottom, NO reverses that order. -Some ODS logic was corrected - changed from NE to EQ. -Some statements reordered for logical ordering Preceding changes were only in GRAFWLM. -Only in GRAFWRKX, the default values for width and height were reduced to 7.5 and 10 inches to eliminate warning messages that the values were too large. Thanks to Daniel McKinzie, Zions Bank, USA.
FORMATS SMF6ACTP MG106CT. SMF6ACTY MG106CD. SMF6ATYP MG106SE. VMAC106 -TYPE1061 dataset SMF6ASET has character and numeric value Aug 17, 2018 that are not documented; MXG creates multiple SMF6ASETxx. -TYPE1062 dataset SMFCMDPM contains two binary fields that are decoded in SMF6A001/002 variables but seem too large to be look up tables for parameters. Thanks to Joe Faska, DTCC, USA. Change 36.151 Updates from SMF Manual Jul 30, 2018. FORMATS -Format MG022ET adds values VMAC42 5='5:COUPLING FACILITY CONTROL UNIT' VMAC62 6='6:LOGICAL PARTITION ENTRY' Aug 12, 2018 9='9:PCIE FUNCTION' -Variables added to TYPE42SR dataset: S42SCRRU='AVG*RANDREAD*CACHEHIT*RESPTM' S42SCRSU='AVG*RANDREAD*CACHEHIT*SERVICTM' -Variables added to TYPE62 dataset: SMF62IND_2='CATALOG*OR CRA*RECORD?' SMF62IND_3='VVDS*OR*ICF*RECORD?' SMF62IND_4='SMS*CLASS*INFO*INCLUDED?' SMF62IND_5='DATASET*IS*ENCRYPTED?'
BUILD005 adds these new variables to TYPE30_4 and PDB.STEPS: BUIL3005 SMF30CAS_OA54589_0='SMF30CAS*OA54589*BYTE 0' VMAC30 SMF30CAS_OA54589_1='SMF30CAS*OA54589*BYTE 1' Aug 12, 2018 SMF30CAS_OA54589_2='SMF30CAS*OA54589*BYTE 2' Aug 27, 2018 SMF30CAS_OA54589_3='SMF30CAS*OA54589*BYTE 3' These variables are created from the bit tests for the preceding Byte variables listed in the SMF Manual: SMF30CAS_OSPROTECT='OSPROTECT'; SMF30CAS_UNTRUSTED='UNTRUSTED?' SMF30CAS_TRUSTED='TRUSTED?'
VMACCIMS (STRTTIME,ENDTIME) but TRNCVTTZ (GMT Offset) is zero and VMXGINIT the site has not responded with their local/GMT times, so Aug 8, 2018 this may be a temporary circumvention, but macro variable IMFGMTOFF is created with value of zero and when it was set &LET IMFGMTOFF=7200; before the %INCLUDE, the times were correct. This change text will be updated when it is known why these times, previously always the same time zone as the ARRVTIME, are now different at this site.
ANALRMFR with a DELETE statement with a wildcard S: caused QA job Aug 8, 2018 to fail with WPS because it honored the MT=ALL default to delete both datasets starting with S, but unexpectedly also deleted the SASMACR CATALOG dataset, which caused QA to subsequently fail. But SAS only deleted datasets so QA did not fail. Since the actual intent was to only delete datasets, adding MT=DATA to the PROC DATASETS corrected for both SAS and WPS. However, using MT=ALL with SAS still only deleted datasets, printing this note: WORK.SASMACR cannot be deleted because it's in use.
FORMATS dataset, adds new variables: VMACDCOL Aug 8, 2018 Thanks to Luc Gielis, KBC, BELGIUM
ASUMUOW in the OBS count between CICSTRAN and ASUMUOW on the MXG NOTE order of 2:1 or better, it may indicate that the use of Aug 7, 2018 MRO is not significant and it may be that running ASUMUOW is a waste of resources. ASUMUOW has to sort all of the CICSTRAN and DB2ACCT and MQ records to be able to merge everything together and can be very resource intensive. -If you have DB2 ROLLUPS that data becomes very suspicious since one DB2 record can represent many transactions and it is best in that case to not use ASUMUOW with DB2 data as documented in change 36.107. -If you find that running ASUMUOW is expensive or that you are not achieving any serious reduction in OBS in ASUMUOW dataset using ASUMUOW compared to the OBS in CICSTRAN, then don't run ASUMUOW: instead, use ASUMCICS to create the PDB.CICS dataset. -All of the DB2 CPU time is captured in the CICS records; only the class 3 wait times/counts are lost, but those are primarily deep diagnostics, where it may be much better to run ANALDB2R against the PDB.DB2ACCT dataset selecting the desired transaction. Change 36.145 QCMPRES is an AUTOCALL macro and we try to avoid using ANALDB2R them since they have caused issues in the past. This MONTHASC change replaces QCMPRES by a %SYFUNC(COMPBL) call to MONTHBLD compress blanks for safety. MONTHBL3
MONTHDSK PRODTEST
PRODTESW Aug 7, 2018 Change 36.144 -Compare Interval CPU time Captured RMF/SMF/CICS/DB2 with COMPINTV totals and plots of each system. The ancient program was Aug 8, 2018 restructured as a macro that can either read SMF data or extract the data from an existing PDB. Parameters allow you to limit the scope of the data being analyzed. The output is a combination of PROC MEANS, TABULATEs, and SGPLOTs and will show you capture ratios for RMF and SMF for both CPUs and ZIPs as well as the CPU times from RMF, SMF, CICS, and DB2. -See the member for documentation on parameters and usage. -Some examples of usage Read SMF and create reports on half hour intervals %compintv(compinterval=halfhour); -Get data from pdb and report on half hour intervals %compintv(compinterval=halfhour,readsmf=no); Change 36.143 Support for RSD Folders Version 6.0 Audit Records with FORMATS INCOMPATIBLE updates. Only version 6.0 is supported, and VMACRSDA one inconsistent record format is under investigation Aug 5, 2018 and is not output. Thanks to Jan Tielemans, KBC, BELGIUM.
VMXGINIT MXG 36.04, but the default MXGKEEP=NO is restored, as Aug 5, 2018 YES can cause errors if there are INCODE variables.
MXG Note zHPF was introduced by IBM to reduce channel utilization Jul 31, 2018 and to improve data transfer performance. At the core, the difference between zHPF enabled and traditional FICON channels is the schema employed to transmit channel programs and data. For traditional CCWs, a channel program (comprised of multiple CCWs) is transmitted from the channel to the subsystem in very small chunks, CCWs and data blocks. Each turn around increases channel utilization and as the utilization increases, the acknowledgement coming back from the subsystem experience delays like Volkswagens driving in a sea of semis. What zHPF does is it packs the CCWs and data blocks into Transmission Control Words, i.e., TCWs. By reducing the number of turnarounds required to transmit CCWs and data blocks, the portion of channel utilization resulting from the turnarounds is reduced. Moreover, by reducing the number of turnarounds required for data transfer, the effective data transfer rate is increased. Finally, your mileage may vary. If your I/O stream primarily supports VSAM or DB2 datasets, you will see a lot of difference. If you have a large fraction of traditional access methods, your level of benefits may be reduced. The TYPE73 Channel Busy is the primary indicator of the benefit of zHPF. Thanks to Dr. H. Pat Artis, Performance Associates, USA. Change 36.140 New READDB2 parameter SORT102 with SORT102=YES default READDB2 can be changed to SORT102=NO to suppress the sorting of Jul 27, 2018 T102Sxxx trace datasets and to suppress VFMT102 execution that creates $MGDB2DB and $MGDB2OB formats when 105/107 IFCIDS are selected. They are used by ANALDB2R to match DBID/OBID in trace datasets. Thanks to Laifai Wong, Bank of America, USA.
UTILBLDP -If you asked for DB2 trace records and specified Aug 7, 2018 SORTOUT=NO the T102xxxx datasets were still sorted. -New parameter AUDITAFTER added with a default of NO. With BUILDPDB=NO, PDBAUDIT is not executed by UTILBLDP. With BUILDPDB=YES and AUDITAFTER=NO, there is no change: PDBAUDIT is executed after BUILDPDB but before MXGINCL and INCLAFTR programs are executed. With BUILDPDB=YES and AUDITAFTER=YES, PDBAUDIT is now executed after all of the specified MXGINCL and INCLAFTR members are executed. With BUILDPDB=YES and AUDITAFTER=NEVER, PDBAUDIT is not executed. -Now if you specify SUPPRESS=ID with BUILDPDB=YES, the creation of dataset ID and execution of ANALID are suppressed. Previously this caused an ABEND. Thanks to Laifai Wong, Bank of America, USA.
VMAC102 input lengths, for both DB2 V12 and V11. Jul 27, 2018 Thanks to Jan Tielemans, KBC, BELGIUM. Change 36.137 -INVALID DATA for variable MICROCODE that MXG input as a VMACRHEL numeric, is caused by character values '0x3c', so new Jul 30, 2018 MICROCODECH character variable is created and MICROCODE is set to a missing value to prevent VARIABLE NOT FOUND errors. -WARNING COUNT USER IN DROP/KEEP/RENAME list was printed if the MXG Default OPTIONS DKROCOND=NOWARN was changed, but they shouldn't have been in the KEEP list and are now removed, and PROGNAME has been added to RHELUARG dataset. -Datasets left in WORK are now deleted. Thanks to Harald Seifert, HUK-COBURG, GERMANY.
VMXGRMFI fall thru to OTHER, i.e., a Service or Reporting Class VMXGINIT was found that was not mapped in your WORK= definitions, Jul 23, 2018 you can force a user ABEND nnnn using non-zero NNNN in %LET MXGABNDRMFI=nnnn; in your sysin. The log will list the first undefined SRVCLASS value. Thanks to Wayne Bell, UNIGROUP, USA.
VMACPOEX some variables should not have been output in some of Jan 9, 2018 the datasets, and a number of new variables are kept. Jan 22, 2018 The four CPU times POEXCPUG, CPUC, CPUD, and POEXCPUL: Jul 20 ,2018 POEXLIST keeps only POEXCPUG and POEXCPUL POEXCLIE keeps only POEXCPUG and POEXCPUC; the variables from FILE and DB2 segments are removed as they were only from the last segment. POEXDB2 keeps only POEXCPUG and POEXCPUD POEXFILE has no CPU times. -New CPU time variables are added to POEXLIST POEXCLIE: POEXGSID='POWER*CENTER*SESSION*ID' POEXGMNM='MAP*NAME' POEXGTCP='CPU*TIME*ON*CP' POEXGTOT='CPU*TIME*ON*ZIIP' POEXGTOF='CPU OFFLOAD*ELIGIBLE*ON*CP' -New DB2 variable added to POEXDB2 dataset: POEX2QTY='DB2*CONNECTION*02X=CAF*12X=RRSAF' Thanks to Scott Wiig, US Bank, USA.
VMAC120 method per bean, but there can be many. The logic was Jul 20, 2018 corrected and all are now output in TYP120JC/TYP120JI. Thanks to Nick Varley, SYNCSORT, USA. Change 36.133 Variable LPARBUSY was not calculated for z14 processor ASUM113 in dataset PDB.ASUM1131. Jul 18, 2018 Thanks to Scott Barry, SBBWorks Inc., USA. Change 36.132 Support for EOS Version 160 (INCOMPATIBLE) Audit/Account VMACWSF records. Only the END records (subtype 4 and 20) have Aug 5, 2018 valid begin and end times and have their GMT times reset Sep 26, 2018 to local time zone. Change 36.131 Reading compressed DB2/CICS data with MXGREADSMF=LOGGER VMACSMF did not invoke the EXITCICS CICSIFUE exit; the &SMFEXIT Jul 12, 2018 macro variable was not in the LOGGER's INFILE statement. -If you have compressed CICS or DB2 records and want to read the LOGGER data, you must use the CICSIFUE INFILE exit (EXITCICS); the internal MXG decompression code is not supported for that combination. Thanks to John Compton, World Programming, ENGLAND. Change 36.130 Variable NDMUID is now populated from NDMZUID in NDMCT VMACNDM and NDMFI datasets with the full 64 byte user id. Jul 12, 2018 Thanks to Amlyn Parry, Barclays, ENGLAND Thanks to Heimir Hauksson, Barclays, ENGLAND
SAS NOTE experience poor performance in DATA steps, with CPU time Jul 12, 2018 increase of 40% reported. The issue is fixed in TS1M4 Aug 8, 2018 and later. See http://support.sas.com/kb/61/906.html A hot fix was planned for TS1M3, note created Feb 28 and modified Mar 8, 2018; no Hot Fix was created since M4/M5 corrected the poor performance. But Aug 8, 2018 SAS Tech Support suggested for SAS 9.4 at TS1M3 that using // EXEC SASPROC,OPTIONS='MSYMTABMAX=20000000' to increase the size of the macro symbol table from 1M to 19M would and did eliminate the CPU time increase. Subsequently, Tech Support said using 2M syntax would and did resolve the problem.
SAS NOTE can fail with ERROR: LIBRARY WEEK31 IS NOT A VALID FORMAT Jul 4, 2018 FOR ACCESS METHOD SASV7SEQ, but that error message is not correct. The error is not an invalid format, but is a memory allocation error. SAS 9.4 M2 added support for LBI (Large Block Interface), which allocates a buffer for each tape data library below the 16MB line, and a large number of SAS tape libraries can exhaust that memory area. Unfortunately, increasing the REGION size does NOT increase the below the line size. Reducing the number of tape data libraries can circumvent the error, which is corrected in SAS 9.4 M4, and there is a Hot Fix for M2 and M3.
EXT11912 DDDDDD Dataset Description EXT119C1 T11912 TYP11912SUM ZERT 12 SUMMARY EXT119C2 T119C1 TYP11912TLS ZERT 12 TLS EXT119C3 T119C2 TYP11912SSH ZERT 12 SSH EXT119C4 T119C3 TYP11912IPSEC ZERT 12 IPSEC FORMATS T119C4 TYP11912DN ZERT 12 DISTINGUISHED NAME IMAC119
VMAC119 VMXGINIT
Jul 3, 2018 Thanks to Rodger Foreman, Black Knight, USA Thanks to Luis Mendoza, Black Knight, USA
Jul 2, 2018 Change 36.125 Variable SRDGCOMM was replaced by SRDGNAME, but due to VMACSRDF use of SRDGCOMM in existing reports, both are kept now. Jun 29, 2018 Thanks to Joe Faska, DTCC, USA. Change 36.124 SMF 42 Subtype 5 ABEND, more invalid LENSR=560/640. VMAC42 As reported in Change 36.027, APAR OA54663 corrects. Jun 29, 2018 MXG circumvention extended to protect these values: IF LENSR IN(232,240,320,376,400,448,480,560,640) THEN LENSR=160; LENSR=376 added Oct 15 in Change 36.124. Thanks to Jan Tielemans, KBC, BELGIUM.
VMAC74 field is incorrect, and causes R748AEBC to be invalid. Jun 26, 2018 BMC1316 corrects. NO MXG CODE CHANGE.
SAGANAL are revised and enhanced to create HTML output files in Jun 22, 2018 either a PDSE or in a ZFS File System or ftp to MXG. Jul 6, 2018 The KEEP CPI: CPU: replaced by specific list of TYPE30_V Jul 22, 2018 variables, reducing kept from 1533 to the intended 64. Thanks to Tennie Olson, TIAA,USA. Change 36.121 If you specified READDB2(IFCIDS=ALL), dataset DB2STATS READDB2 wasn't created; circumvent with (IFCIDS=ALL STATISTICS). Jun 16, 2018 MXG 35.03-36.05. Thanks to Hans Coolen, Allianz Technology, THE NETHERLANDS. Change 36.120 Support for BVIR History HSM Compression Container V412 EXBVR303 creates new dataset: FORMATS DDDDDD Dataset Description VMACBVIR BVR303 BVIR303 HSM COMPRESSION CONTAINER Jun 16, 2018 and new format MGBVIME decodes Compression Method Thanks to Bradley Leis, TELUS, CANADA. ====== Changes thru 36.119 are in MXG 36.05 dated Jun 13, 2018========== Change 36.119 -Formats MG119CI and MG119MA did not decode new values for FORMATS variables SSH_CIPHER and SSH_MAC in TYP11994 & TYP11995, VMAC119 and protection for unknown values prints the $HEX4 value. Jun 13, 2018 -New variable T119RCID='RECORD*ID' added to all datasets. Change 36.118 Support for TANDEM TMF data creates new TANDTMF dataset. EXTANTMF DDDDDD DATASET DESCRIPTION IMACTAND TANTMF TANDTMF TMF TRANSACTION DATA VMACTAND
VMXGINIT Jun 11, 2018 Thanks to Kurt Gramling, TSYS, USA.
ANALDB2R report. If you happened to specify QWHSSSID it worked Jun 11, 2018 but that was a coincidence. Now produces a message to tell you and sets SORTBY to QWHSSTCK QWHSSSID. Change 36.116 STCVSM11 variables NIO and CUB in VSM6 are now bytes so VMACSTC new variables are created with B added as last character. Jun 13, 2018 The previous code did not initialize the six calculated Jun 15, 2018 variables, causing their sum to be greater than the CUB Jun 22, 2018 and NIO variables; that is corrected. Jul 2, 2018 -Jun 15: First 36.05. Two Debugging PUTLOGs removed. -Jun 22: Test to identify VSM6 vs earlier now tests for VSM6 or ELSE DO; for any other STC11VTS name value. -Jul 2: Test to identify VSM6 now tests STC11CSP for values of 1000 or 8000 based on this site's values for STC11INM to modify that test: PROC SORT DATA=STCVSM11; BY STC11CSP; PROC FREQ; TABLES STC11CSP*STC11TOL/NOROW NOCOL NOPERCENT; TITLE STCVSM11 TABULATIONS; RUN; PROC FREQ;BY STC11CSP; TABLES STC11INM*STC11TOL/NOROW NOCOL NOPERCENT; TITLE STCVSM11 TABULATIONS; RUN; PROC MEANS N MIN MAX SUM; BY STC11CSP;RUN; Thanks to Randy Hewitt, DXC, USA. Change 36.115 Unused Change Number. Change 36.114 DB2ACCTR dataset has been misaligned when QLACOFF1 is not VMACDB2 zero, i.e. if QLACLOCN field is longer than 16 bytes, and Jun 10, 2018 there is more than one QLAC segment, due to 2 undoc bytes in the second and subsequent segments, but it was only INVALID DATA FOR QLACCPUL/QLACDBWT messages that exposed the error - no user had reported the bad QLACxxxx values, suggesting DB2ACCTR has not been important nor used! And those QLAC variables are also kept in DB2ACCT, but only from the LAST QLAC segment, because originally there was only one QLAC segment. Option a: Leave the QLAC variables in DB2ACCT as-is with this documentation that they are only from the last segment. Option b: Set all QLAC variables in DB2ACCT missing but keep them; dropping existing variables by MXG is unsafe because it could cause an ABEND with VARIABLES NOT FOUND if you have an old report that references then. I have NOT chosen this option. Option c: Create macro _DRPQLAC listing all QLAC vars kept in DB2ACCT so you can add MACRO _KDB2ACR DROP=_DRPQLAC % in your IMACKEEP to always drop them from DB2ACCT, or use you can use %LET MACKEEP= MACRO _KDB2ACR DROP=_DRPQLAC % ; in the //SYSIN of your DB2ACCT create job. MACRO DRPQLAC is created and available, but it does not drop QLACLOCN,QLACCNVR due to references to those variables in others. Thanks to Scott Wiig, US Bank, USA.
VMACDB2 QPAC_PIPE_WAIT and QPAC_PIPE_COUNT to correct those two Jun 5, 2018 variable's values. Thanks to Giuseppe Giacomodonato, EPVTECH, ITALY. Change 36.112 Support for Dell EMC Mainframe Enablers 8.30, previously EXSRDF02 SRDF, Symmetric Remote Data Facility, creates separate EXSRDF03 datasets for each subtype: EXSRDF04 DDDDDD DATASET DESCRIPTION EXSRDF06 SRDF02 SRDFA02 TOLERANCE MODE CHG EXSRDF07 SRDF03 SRDFA03 ACT CHANGE EXSRDF08 SRDF04 SRDFA04 SECONDARY DELAY IMACSRDF SRDFAA SRDFAA REGULAR INTERVAL 05 VMACSRDF SRDF06 SRDFA06 RECULAR MSC INTERVAL VMXGINIT SRDF07 SRDFA07 WRITE PACING GROUP Apr 27, 2018 SRDF08 SRDFA08 WRITE PACING DEVICE Jun 5, 2018 Change 36.111 JCL examples to CPORT/CIMPORT data from WPS to SAS, Yüklə 28,67 Mb. Dostları ilə paylaş:
|