* copyright (C) 1984-2019 merrill consultants dallas texas usa
Yüklə 28,67 Mb.
|
|
h or hh value of zero means scale timestamps to GMT time. Unsigned h or hh values imply a positive GMT offset. A '-' sign is required to specify a negative offset.
to make the unit measure clear if desired. -Positive GMT offsets are for time zones east of GMT up to the International Date Line including most (if not all) of Europe, Africa, Asia, Australia, and many island groups.
-Negative GMT offsets are for time zones west of GMT up to the International Date Line including North and South America and some island groups. -A few time zones have GMT offsets that are not integer hour values such as India, some Australian zones, and some island groups. For example, India is GMT+5:30. -For the support of these non-integer offset time zones any of the following formats are supported for GMTOFFSET= (and aliases GMTOFF=, GMT=) in minutes:
mmM -mmM +mmM mmmM -mmmM +mmmM
from 000 to 720 (12 hours). Values over 720 are flagged as errors and will abend ASMRMFV. An m, mm, or mmm value of zero means scale timestamps to GMT time. Unsigned m, mm, or mmm values imply a positive GMT offset. A '-' sign is required to specify a negative offset. Any user can still specify this form even for integer hour offsets by converting the hours x 60 to get minutes. For example, GMT=-4 and GMT=-240M are equivalent.
omitted the value will be handled as an hour value instead. -NOTE: When using GMTOFFSET= the FROMDATE=/TODATE= and FROMTIME=/TOTIME= filter values MUST be coded based on the REQUESTED offset time zone. This is NOT necessarily the Local Time for the time zone where ASMRMFV is executing. -GMTOFFSET= processing follows these steps for each RMF Monitor III MINTIME Sample Set: 1) The Sample Set begin (SSHTIBEG) and end (SSHTIEND) timestamps are first converted to GMT time using the SSHSTDIF GMT offset field from the Sample Set Header (SSH) present for each Sample Set. Then they are set into temporary timestamp fields for filtering.
are applied to the temporary timestamps and filtering continues with the timestamps in GMT Time. They are compared to the FROMDATE=/TODATE= and FROMTIME=/TOTIME= option settings.
fields are further altered with the negative or positive offset value. These altered timestamps are referred to as Adjusted Time in ASMRMFV documentation and messages. They are compared to the FROMDATE=/TODATE= and FROMTIME=/TOTIME= option settings. 4) As noted earlier any selected Sample Set tables are output to the RMFBSAM file with their original timestamps unchanged.
installation responsibility to transfer the multi time zone RMF Monitor III data sets to the ASMRMFV execution site prior to processing. The RMF provided ERBV2S and ERBS2V Clists are one method to create and retrieve a sequential copy of an RMF Monitor III VSAM data set. -Example 1: A London based company wants build an RMF III PDB for yesterday with RMF III VSAM data sets input from several different time zones in Europe and Asia for their peak hours of 09:00 to 15:00. They want to see what other activity is occurring elsewhere during this time.
FROMDATE=YESTERDAY TODATE=YESTERDAY FROMTIME=0900 TOTIME=1500 GMT
PDB for two days ago in June with RMF III VSAM data sets input from several different time zones in the United States for the prime time hours of 08:00 to 17:00. They need to see if some moving some workloads might result in fewer delays. They are using Daylight Saving Time and their time zone is at GMT=-4.
FROMDATE=*-2 TODATE=*-2 FROMTIME=0800 TOTIME=1700 GMTOFF=-4
Saving Time the GMT offset changes during the fall transition to Standard Time and the GMT offset increases by 1 hour. In the example above it becomes GMTOFF=-5. One advantage of using pure GMT offsets is that time changes such as this are not an issue because RMF III keeps the GMT offset for each Sample Set. RMF III does not have any awareness of Daylight Saving Time and so it is a user responsibility to code GMTOFFSET= correctly before and after a time change. -Example 3: An India enterprise wants build an RMF III PDB for the last five days with RMF III VSAM data sets input from several different time zones in Asia and Europe for their early morning hours of midnight to 07:00. India Time is at GMT+05:30 hours. They want to see if some workload balancing might be possible across multiple data centers to reducing processing delays or take advantage of available CPU cycles. ASMRMFV statements: FROMDATE=*-5 TODATE=*-1 WINDOW FROMTIME=0000 TOTIME=0700 GMT=+330M -Most ASMRMFV timestamp messages are revised or added to now display the GMT time when GMTOFFSET is used. These include:
RMFV001I Current Time and Last IPL Time RMFV008I Input data set Last Open Time RMFV012I Sample Set Found Begin and End Times RMFV013I Sample Set Selected Begin and End Times RMFV017I RMF and z/OS Version Found Time RMFV023W Sample Set Date/Time - Service Policy missing RMFV032E Sample Set Date/Time - Program service failure RMFV039I Sample Set Date/Time - SHOWSAMP option info RMFV070* Sample Set Date/Time - Service Class Find error RMFV071* Sample Set Date/Time - Report Class Find error RMFV072* Sample Set Date/Time - Workload Name Find error RMFV073* Sample Set Date/Time - Resource Group Find error RMFV076I Sample Set Date/Time - SHOWASI option info RMFV078I Sample Set Date/Time - Prior Service Policy use (* = E, W, or I depending on error settings) -In addition all above messages (except RMFV001I and RMFV008I) display the Adjusted Time if GMTOFFSET is non-zero. So it is possible to get up to three messages for each timestamp display when GMTOFFSET= is in effect: Local Time, GMT Time, and Adjusted Time. These provide an audit trail and verify program operation is correct. -New parameter SHOWGMT (aliases SHGMT, SG) will display GMT versions of timestamp messages even if GMTOFFSET is not in effect. SHOWGMT is forced if GMTOFFSET is in effect. Updated message RMFV037I displays SHOWGMT setting.
-New parameter NOSHOWGMT (aliases NOSHGMT, NOSG) will suppress GMT versions of timestamp messages. NOSHOWGMT is the default so there should be little need to code this option. -New parameter SHOWASI (alias SHASI) displays some ASI entry data when selected. This is intended primarily for debugging as it will produce voluminous output in the ASMRMFV log. This function formerly required a re-assembly and re-link to be enabled. Updated RMFV037I message shows status of this setting. -New parameter NOSHOWASI (alias NOSHASI) suppresses ASI data display. This is the default and should not need to be coded. -New RMFV006I message shows GMTOFFSET status and offset value. -New RMFV014I message displays when all tables have been excluded due to filtering that were not originally excluded by an entire data set bypass condition. For example, this can occur when using range and/or pattern filters to select specific jobs that did not run in the selected time range. -New aliases added SHSAMP for SHOWSAMP and NOSHSAMP for NOSHOWSAMP (default) -New aliases added SHMATCH for SHOWMATCH and NOSHMATCH for NOSHOWMATCH (default). -New aliases added SHZERO for SHOWZERO (default) and NOSHZERO for NOSHOWZERO. -New aliases added SHALL for SHOWALL and NOSHALL for NOSHOWALL (default). -Add sections to MXG00 ASMRMFV Initialization table for table capacities, table sizes, buffer/workarea settings, multiple filter logic options, filter options, GMT offset settings, report options, output options, error options, and table selection options. Updated BUILD00 subroutine. Nearly all ASMRMFV parameter options are now saved except contents of range and pattern tables. These fields become variables in the ZRBASM data set in the MXG PDB. -Raise MXG00 record version to X'04' from X'03'. -Code path improvements for SHOWTS and STCKCONV subroutines. -Some data set names and volume serial numbers were incorrect in the MXG00 ASMRMFV Initialization table. -The CAT and CPC tables were not included in example discussions in the JCLRMFV, JCLCRMFV, and JCLCRMFV members. -Documentation updates to: Section 2 Terminology (Timestamps) Section 5 Input Data Selection Parameters Section 6 Report Control Parameters Section 12 Messages Section 13 Filtered Records Section 15 Program and IBM Limitations Section 26 ASMRMFV and MXG PDB Data Relationships Section 27 Summary Thanks to MP Welch, Bank of America, USA.
BLDSMPDB caused by a typo, the lack of an underscore on the old Jun 1, 2016 style _TODAY macro, causing LASTWEEK to be miscalculated. Impacts MONTHLY job, but FORCEDAY=01JUN16 circumvents. Thanks to Jim Hayes, Huntington Bank, USA.
VMACRMFV now input and kept in RMF III dataset ZRBGEI, formatted May 31, 2016 with MGBYTES. Thanks to Dave Cogar, Wells Fargo, USA. Change 34.129 Variable R723MCPG, the number of periods in this service VMAC7072 class, is now kept in dataset TYPE72PD. May 30, 2016 Thanks to Jim Horne, Lowe's Companies, Inc., USA. Change 34.128 Zen CSM records ZOSAPOOL dataset new ZOSA_POOL_TYPE is VMACZOSA 'E' for ECSA Pool or 'D' for DataSpace Pool. May 27, 2016 Thanks to Jerome Vitner, Experian, USA. Change 34.127 If only one variable was being examined, a NOT SORTED ANALRANK error could occur due to an insufficiently specified May 24, 2016 BY list. Thanks to Tom MacCabe, Dominion Resources Services, Inc., USA. Change 34.126 TYPETMS5 observations for tapes created by old versions VMACTMS5 of DFDSS contained zeros for BLKSIZE but PGM='ADRDSSU' May 23, 2016 tapes are always BLKSIZE=65520, so MXG sets that value for these old tapes. Thanks to Jim Agrippe, Cleveland Clinic, USA.
VMACCIMS PUT 1502 PTFs BQI2154, BPK2892 were supposed to correct May 23, 2016 corrects zIIP CPU times where zIIP Eligible time TRXZIOCP was greater than the CPU time on CP, TRXZONCP, but does not appear to correct the problem as of this date.
FORMATS with new values. May 18, 2016 Thanks to Giuseppe Giacomodonato, EPVTECH, ITALY. Change 34.123 -Support for IFCID 365 populates T102S365 dataset. VMAC102 -Support for IFCID 376 corrects QW0376VN so only QW0376VL May 18, 2016 bytes are input. Variable QW0376TS is an invalid 8-byte Jun 1, 2016 value: '1982A5641F29CA5A'x and '0E5F1F1D09F14040'x are not valid TODSTAMP nor 10-byte DB2 time fields. Thanks to Lai Fai Wong, Bank of America, USA. Change 34.122 ANALCSQC counts concurrent MQ Applications from SYSLOG in ANALCSQX this tailored use of TYPSSYSL that selects only logon May 17, 2016 +CSQX500I and logoff +CSQX501I MSGID to create a session event observation, which ANALCNCR then processes to count and plot concurrent sessions for each quarter hour. Thanks to Tom M. Kane, AT&T, USA. Change 34.121 Formal support for SYSLOG (including multi-line messages) EXSYSLOG with all MXG dataset tokens, to replace SYSLOG member, in FORMATS particular, so EXSYSLOG/_ESYSLOG dataset exit exists so IMACSYSL only desired MSGID are output. TYPESYSL exists but only TYPESYSL creates raw data in WORK.SYSLOG; TYPSSYSL must be used as TYPSSYSL it invokes the _SSYSLOG sort macro that combines multi VMACSYSL messages into one observation and writes out PDB.SYSLOG. VMXGINIT
May 17, 2016 Change 34.120 ERROR: SPIN.SPINPDBAUDIT.DATA HAS TOO LONG A MEMBER NAME PDBAUDIT occurs if your //SPIN DD was created with SAS Version 6, May 17, 2016 which allowed only 8-character SAS dataset/member names. You need to create a new VERSION 9 format data library by copying the current //SPIN DD data to a NEW V9 SPIN DSN: // EXEC MXGSASV9 //SPIN DD DSN=YOUR.OLD.SPIN,DISP=SHR //SPINNEW DD DSN=YOUR.NEW.SPIN,DISP=(,CATLG),SPACE... //SYSIN DD * PROC COPY IN=SPIN OUT=SPINNEW MT=DATA; and then delete OLD and then rename NEW to OLD. You probably also need to examine all of your re-used SAS data libraries (PDB,MON,TUE,...WEEK,MONTH, i.e., those with DISP=OLD that are re-written each time), to see if any were also created with SAS V6, with the output of PROC CONTENTS DATA=PDB._ALL_ NODS DETAILS to see what ENGINE created each of those data libraries. While MXG has had long dataset names for some time, this is the first instance in the "mainline" SMF processing code members used in BUILDPDB, and was introduced in MXG 33.07 in the new PDBAUDIT report of your PDB libraries. (SAS 6.08 dates back to 1992, so this site's SPIN dataset has stood the test of time!!) Thanks to Jeanne Vetter, Dell Services, USA. Change 34.119 There is a known SAS exposure that can cause a CPU loop VMXGCNFG after a program has finished, in SAS termination, if you May 17, 2016 try to dynamically allocate a DD that was already in JCL. If you use the CONFIG= CONFIMXG option as your MXG JCL //MXGSTEP EXEC SAS,CONFIG=MXG.SOURCLIB(CONFIMXG) //MXGNAMES DD DSN=MXG.USERID.SOURCLIB(MXGNAMES), it does dynamically allocate the SOURCLIB and LIBRARY DD, but the restriction was only documented. This change prevents the loop by testing for their allocation, and causing the job to USER ABEND 777 before the SAS program actually starts.
VMAC30 usually negative because the three component variables May 15, 2016 should have been added to get the IFA/ZIP Instructions. Thanks to Paul Volpi, UHC, USA. Change 34.117 CPU and SU_SEC values for z13 processors added to the GRAFWRKX formats so that you can model these newer systems with May 12, 2016 your existing data. This is a very simplistic model that will only convert the CPU time from the current model to whatever model you specify with NEWMODEL=.
VMACEDGR SYSTEM and EDGRTIME to all datasets, retained from the May 13, 2016 Header record. Thanks to Linda Berkeley, USPS, USA. Change 34.115 Variable DCDTIMEC, Data Set Create Time in DCOLDSET is VMACDCOL only populated if May 11, 2016 - the dataset is on an EAV volume (more than 65K CYL) - the volume is the first volume for the dataset (DCDTIMEC is zero in the DSCB for other volumes) - for non-VSAM, EATTR=OPT must be specified (JCL or Data Class), because EATTR=NO is the default for non-VSAM, EATTR=OPT is the default for VSAM. The DCDTIMEC comes from the FORMAT 9 DSCB control block (in the VTOC), created for EAS-eligible datasets on EAV that have EATTR=OPT, except that these datasets do not have FORMAT 8/9 DSCBs. Thanks to Donna Roff, FISA NYC GOV, USA.
ANALTCP selection by user name remote IP address for FTP, API, or ANAL119 Telnet datasets from SMF 118 (TCP) or 119. Jun 8, 2016 Thanks to Dave Ireland, USDA, USA.
EXSIEM01 Also called Hiperstation Application Audit. EXSIEM02 New datasets are created from these subtypes: EXSIEM03 dddddd dataset description IMACSIEM SIEM01 SIEM3270 SIEM 3270 Session 01 TYPESIEM SIEMA1 SIME3270A SIEM 3270 Screen Lines 01 TYPSSIEM SIEM02 SIEMTCP SIEM TCP 02 VMACSIEM SIEM03 SIEMMQ SIEM MQ 03 May 9, 2016 Currently, only subtype 1 is created and supported. May 19, 2016 -May 19. Variable SIEMLPAR NOT FOUND corrected. Change 34.112 MXG 34.01,34.02. DCOLLECT dataset DCOLBKUP variables VMACDCOL UBDSIZE and UBRECSP were incorrectly multiplied by the May 9, 2016 original *1024 that should have been removed when the tests for UBFLAG4 was added in Change 34.042. Thanks to Thomas Peiper, TIETO SWEDEN AB, SWEDEN.
EXTY72PD created for every service and reporting class. IMAC7072 VMAC7072
VMXGINIT May 6, 2016 Thanks to Scott Barry, SBBWorks, Inc, USA.
ANALUOW code on all variables and not just time and transaction May 3, 2016 name. Logic added to detect that no data was found and terminate ANALUOW Thanks to Dave Ireland, USDA, USA.
VMACDB2 variables. QX variables exist only in DB2ACCT and the May 4, 2016 DB2STATB datasets (initially from DB2STAT1). Thanks to Jane S. Stock, USPS, USA. Change 34.108 DB2 Simulated Buffer Pool DB2STSBP/DB2STATS variables VMACDB2 QBSPIUS, QBSPSUS (current) and QBSPHSU, QBSDPHUS (hwmark May 3, 2016 pages) should not have been deaccumulated. Variable QBSPREADS is now correctly deaccumulated. And variable QBSTRHS in DB2STATB is now deaccumulated. Thanks to Lai Fai Wong, Bank of America, USA. Change 34.107 -A typo, SYSTYPE instead of &SYSTYPE caused unresolved ANAL9914 macro because no observations were created. The correct May 3, 2016 syntax for the report for SYSTEM=SYS1 and SYSTYPE=Z13 is May 4, 2016 %ANAL9914(SYSTEM=SYS1,SYSTYPE=Z13); May 5, 2016 -No longer restricted to a single system unless you use the SYSTEM= parameter; by default reports on all SYSTEMS. Cleans up after itself and produces NOTES to tell you when there is a problem. Thanks to Luis A. Mendoza, TRANSUNION, USA.
VMAC7072 of zIIP engines in the CEC, can be wrong in datasets VMXG70PR PDB.TYPE70, PDB.ASUMCELP, and PDB.ASUMCEC, but is correct Apr 30, 2016 in PDB.TYPE70PR, and variable ZIPCPUS, zIIPs online to an LPAR, was also correct in those datasets. -Note that IBM's CPC Report counts ONLINE ZIPs per LPAR, but MXG's ZIPCPUS='ONLINE*AND*NOT*PARKED' -Do NOT use the PDB.ASUM70PR nor PDB.ASUM70LP datasets; they are by SYSTEM and thus selection is required, and they don't have correct data on system's whose SMF data was not read; those LPARs are in PDB.ASUMCELP. Thanks to Elie Sawaya, Royal Bank of Canada, CANADA. Change 34.105 Support for SMF 123 Liberty z/OS Connect EE Audit Record: EXTY123A DDDDDD DATASET DESCRIPTION IMAC123A TY123A TYPE123A z/OS CONNECT EE AUDIT TYPE123A -In 2009, IBM used SMF 123 for S/390 Parallel Query Server TYPS123A which is still in TYPE123, although I presume that record VMAC123A is no longer created. VMXGINIT Apr 30, 2016 May 19, 2016 Thanks to Victoria Lepak, Aetna, USA. Thanks to Don Bagwell, Aetna, USA.
VMACOMCI was not listed in the test for valid versions. Apr 28, 2016 Thanks to Bob Duchesneau, Northwestern Mutual, USA. Change 34.103 Support for IBM Integration Bus, Version 9.0.0.5 SMF 117 VMAC117 INCOMPATIBLE changes to the FLOW record. Apr 28, 2016 Thanks to Ben Thompson, Northern Territory Government, AUSTRALIA. Change 34.102 -For IFCIDs that create more than one T102Sxxx dataset, READDB2 READDB2 needs IFCID-specific logic, but the new T102SA58 VMAC102 dataset for IFCID=58 was overlooked. Apr 28, 2016 -Support for T102SA58 dataset in Change 34.072 was correct only for DB2 V12 with a longer new segment; this change supports and validates the shorter DB2 10.1 record. -No "Truncated" name fields existed in test records so the support for those longer 0058 names awaits test data. Thanks to Phil Grasser, Norfolk Southern, USA.
GRAFWRKC indicating where the group cap lies on both the percent Apr 28, 2016 CPU and the MSU charts. This required summarization of the data so that there was only a single OBS per by group and uses the VLINE parameter of SGPLOT. The graphs only work if you are running SAS 9.3 or higher. If not a message will be on the log and a PROC TABULATE will be run instead. Change 34.100 -ZRBASI dataset variable ASILPGSZ was incorrect. VMACRMFV -ZRBGEI dataset variables below are now correctly divided Apr 28, 2016 by SSHSMPNR: May 5, 2016 GEISASL GEIRSTRF GEILCPR GEILCMO GEILF4K GEILP4K May 16, 2016 GEILPFRI GEILPFCI GEILCMU GEILCPU GEILFPF GEILSMO GEIRFREM GEISUSE GEILPAG GEILFUSE GEILPUSE GEIRSTRF, and variable GEIRSHR is now kept in dataset ZRBGEI. Thanks to Matthew Chappell, QLD Dept Transport Main Roads, AUSTRALIA Change 34.099 zVM 6.3 z13 SMT-Mode MONWRITE support and correction. VMACVMXA -PRCMFM (5.20) HIS SMT-Mode-Only was not as presumed. I Apr 23, 2016 had thought the new HIS record would be like the new z/OS Apr 26, 2016 SMF 113 Subtype 1, with interval data instead of the May 9, 2016 accumulated fields. But the new PRCMFM is unrelated to existing HIS counters in PRCMFC, and reports only the new MT-diagnostic counters, with only two (MTDIA448-MTDIA449) documented, and three others (MTDIA452,MTDIA453,MTDIA456) are populated, but IBM has claimed them proprietary so their content is not documented. But all five are kept in VXPRCMFM dataset. -PRCAPM (5.10) CRYPTO record has undefined Crypto Type 11 Yüklə 28,67 Mb. Dostları ilə paylaş:
|