* copyright (C) 1984-2019 merrill consultants dallas texas usa
Yüklə 28,67 Mb.
|
|
as DAYS= between the FROMDATE= and TODATE= values and also the setting of the WINDOW/NOWINDOW option. -ASMRMFV now uses 64-bit Grande z/Architecture instructions in all 64 bit binary TOD date and time calculations for better performance. Prior versions used an older method with two 32 bit registers and carry/borrow logic that was cumbersome with longer instruction paths. -Filter message RMFV006* (* = I or E) is now multi-severity. When the FROMDATE= value exceeds the TODATE= value then message RMFV006E will be issued. A U0998 abend will then occur as in prior versions. -The FINDGREG subroutine that determines the day of the week (dow) for TOD values will retain the results of the prior dow calculation and as a performance improvement not repeat it if the date has not changed between consecutive calls. -Relative date selection will now allow up to FROMDATE=*-999 or TODATE=*-999 for date backup values. Previously 99 was the limit. However, the need for this larger limit should be extremely rare as 999 days is about 2.7 years. -New option SHOWSAMP (alias SS) will display the begin and end timestamps for each RMF Monitor III Sample Set selected. SHOWSAMP is intended for diagnostic use and possible user testing with the new WINDOW or DOW= parameters to verify date and time selection is as desired.
use due to TOD conversion overhead for each Sample Set and ASMRMFV log output volume. -When SHOWSAMP is in effect a pair of new messages RMFV039I are issued showing Sample Set begin and end timestamps and the Sample Set number (1-1110). -New option NOSHOWSAMP (alias NOSS) is the default and provides the same ASMRMFV behavior as before with no RMFV039I messages issued. Normally there should be no need to code this option. -Options display message RMFV037I will show the setting of SHOWSAMP/NOSHOWSAMP. -Documentation in ASMRMFV and ADOCRMFV members has been updated to include details on WINDOW/NOWINDOW, SHOWSAMP/NOSHOWSAMP parameters as well as for messages RMFV006*, RMFV037I, and RMFV039I. -A new DOW= parameter allows optional RMF III data selection based on the day of the week. DOW= will be mainly useful for ad hoc studies where only activity on a certain set of days is of interest. -DOW= is coded as DOW=value where value uses one of these formats: dow dow:dow WEEKDAYS WEEKENDS ALL dow is a 3 character day of the week: SUN MON TUE WED THU FRI SAT : is a range operator indicating selection of a group of consecutive days including the first and last dow in the range and all days in between.
of a week on Sunday so that DOW=FRI:MON is perfectly valid selecting FRI, SAT, SUN, and MON.
coding DOW=MON:FRI . Value Alias(es) for WEEKDAYS are: WD, WEEKD, WEEKDA, WEEKDAY
coding DOW=SAT:SUN . Value Alias(es) for WEEKENDS are: WE, WEEKE, WEEKEN, WEEKEND
filtering occurs. DOW=ALL is the default and so should not needed to be coded.
conversion of the 64 bit binary TOD microseconds value to a day of the week occurs for both the RMF III Data Set Header (DSH) and Sample Set Header (SSH) tables. -However, dow determination is not repeated for consecutive table TOD values for the same date. -DOW filtering is applied AFTER all FROMDATE=, TODATE=, FROMTIME=, TOTIME=, and WINDOW data/time filters are applied. -DOW= may be combined with one or more of the these existing date/time selection parameters or may be used exclusively by itself. -DOW= usage examples: Parameter Selects RMF III Sample Sets for --------- ------------------------------- DOW=MON:FRI MON, TUE, WED, THU, FRI DOW=WD MON, TUE, WED, THU, FRI DOW=SAT:SUN SAT, SUN DOW=WE SAT, SUN DOW=ALL SUN, MON, TUE, WED, THU, FRI, SAT DOW=SUN:SAT SUN, MON, TUE, WED, THU, FRI, SAT DOW=FRI:MON FRI, SAT, SUN, MON (dow wraps) DOW=WED WED only DOW=MON DOW=WED DOW=FRI MON, WED, FRI only DOW=MON:TUE DOW=THU:FRI MON, TUE, THU, FRI only -Filter display message RMFV006I will show the setting of the DOW= filter. -Six more examples have been added to each of the JCLRMFV, JCLCRMFV, and JCLDRMFV sample JCL members showing use of the new DOW= parameter. -Documentation in ASMRMFV and ADOCRMFV members has been added to discuss the DOW= parameter in detail. -Null (blank) values for keyword=value type ASMRMFV parameters are now allowed. The parameter is ignored. For example, FROMDATE= is accepted with no change in the FROMDATE occurring. Previously zero length keyword values were flagged as an error. -This feature allows users to provide these keywords in templates for a possible future addition of an actual value. Also existing keywords can now be nullified by just setting the value to blanks rather than removing the entire keyword. -Multi-severity messages RMFV051* through RMFV055* (* = I, W, or E) have been updated for consistency and alignment with message RMFV008I for the DSN value. -A new subroutine WRITESEP will now handle the generation of separator lines used to delineate sections of the ASMRMFV log for visibility. Prior to this every message id was checked in the WRITEPRT subroutine for separator eligibility before the message was issued. Now WRITESEP will be called only as needed for specific messages. -A separator line will now be written prior to the final RMFV999I message for better visibility of important ASMRMFV final results. -Entries in Section 2 "Terminology" in ASMRMFV documentation have been alphabetized and a few new entries added. -All examples in the JCLRMFV, JCLCRMFV, and JCLDRMFV sample JCL members that use date/time parameters now include a small diagram that visually shows what time frames are selected. -The RMFV014I RMF III data set bypass message has been reformatted for better clarity and will now indicate whether the data set is excluded due to date/time or DOW= filters. -A new RMFV014W warning message is issued at the end of processing if ALL RMF III data sets have been bypassed due to date/time and/or DOW= filtering. -An invalid value for the DEVTYPE= or any *ERR= parameter could allow ASMRMFV processing to continue instead of abending as intended because the parameter error counter was not being updated for these keywords. -When multiple RMFV004E messages were issued due to parameter errors it was possible some of the descriptive problem text could be garbled. -Julian years before 2000 in FROMDATE= or TODATE= were not flagged as errors when they should have been. -Message RMFV008I Descriptor Code overlaid when issued as a WTO by the WTO option. -Messages RMFV017I and RMFV038I should not be issued as WTOs when the WTO option is in effect. -DOW table entries were not blanked if DOW=ALL coded. -NOTE: When the WTO option is coded it should be the FIRST parameter in the JCL PARM= field or in SYSIN (if no JCL PARM= field). Otherwise any errors for other parameters will not be shown in JESMSGLG and JESYSMSG until the WTO option is encountered. -Documentation in ASMRMFV and ADOCRMFV members has been updated as appropriate for the above changes. -REQUIREMENT: In order to implement these features the ASMRMFV utility program from this MXG change must be installed. See MXG SOURCLIB member JCLASM3 for sample JCL for the assembly and link-edit install steps. -Nov 25: Discovered the TODAY selection did not work correctly, made the correction, and redated ASMRMFV. This error was ONLY in the original 32.225 dated 9/22. Change 32.224 The _IDaaaa=512 protection added in Change 32.198/180/149 VMACXXXX costs 10 CPU secs per 20,000,000 records, caused by the Sep 16, 2014 statement ELSE IF LENGTH(COMPRESS(_IDaaaa)) that tests Sep 29, 2014 the value of the old-style macro _IDaaaa for 512 and also supports the "512 OR ID=513" multi-IDs syntax. That statement causes a NUMERIC to CHARACTER conversion, for each record, EVEN for the EXPECTED case where _IDaaaa has been defined (when the ELSE clause is NOT even executed). Since the "512" test only benefits "newbies" who didn't find the doc to define their _IDaaaa macro, it certainly should NOT cost "oldbies" who knew what to do. That 512 default value is only defined internally in MXG VMACaaaa members, so by changing the MXG default ID value for all USER SMF records from 512 to a MISSING VALUE, i.e., using MACRO _IDaaaa . % as the internal default definition, the test now uses the MISSING(_IDaaaa) function to detect and tell this new user why all the datasets have zero obs, and how to correct. And since no conversion is involved, old users see no increase: the MISSING function cost 0.08 CPU secs per 20,000,000 records. SEE CHANGE 32.234. Thanks to MP Welch, Bank of America, USA. Change 32.223 Update for MainView for MQ Version 5.2. VMACBBMQ Changes are COMPATIBLE, MANY new variables. Sep 16, 2014 -New variables added to BBMQQMGR (RTIN='E1'x) dataset: Dec 11, 2014 PARM1 PARM2 QMCERTLBL QMCHLAUTH QMCONNAUTH Dec 18, 2014 QMCRTVALPOL QMDEFCLXQ QMGRPUR QMICBC QMICBR Jan 2, 2014 QMICTLC QMICTLR QMIPUBSC QMIPUBSR QMISTUSC QMISTUSR QMISUBC QMISUBR QMISUBRC QMISUBRR QMMAXUMSG QMPSCLUS QMQSGCRTLBL QMRCBC QMRCBR QMRCTLC QMRCTLR QMREVDNS QMRPUBSC QMRPUBSR QMRSTUSC QMRSTUSR QMRSUBC QMRSUBR QMRSUBRC QMRSUBRR QMSCBC QMSCBR QMSCTLC QMSCTLR QMSPLCAP QMSPUBSC QMSPUBSR QMSSTUSC QMSSTUSR QMSSUBC QMSSUBR QMSSUBRC QMSSUBRR QMSTRDATE QMSTRTIME QMSUITEB -New variables added to BBMQBUFF (RTIN='E2'x) dataset: SBPLOC SBPPGCLS -New variables added to BBMQPAGE (RTIN='E3'x) dataset: SPSEXPNDT SPSPADC1 SPSBPID SPSPSIDX SPSRBAV8 SPSRBA2V8 -New variables added to BBMQLMGR (RTIN='E4'x) dataset: LMDATASETNM1 LMDATASETNM2 LMFULLLOGS LMICMPB LMICMPBR LMICMPF LMICMPFR LMICMPR LMICMPRR LMIDCMP LMIDCMPR LMIDECF LMIDECFR LMIDECR LMIDECRR LMIDUCMP LMIDUCMPR LMIUCMP LMIUCMPR LMLGSUSPND LMLOGCOPYNM1 LMLOGCOPYNM2 LMMRCRBA LMOFLOSTA LMQMSRBA LMRCMPB LMRCMPBR LMRCMPF LMRCMPFR LMRCMPR LMRCMPRR LMRDCMP LMRDCMPR LMRDECF LMRDECFR LMRDECR LMRDECRR LMRDUCMP LMRDUCMPR LMRUCMP LMRUCMPR LMSCMPB LMSCMPBR LMSCMPF LMSCMPFR LMSCMPR LMSCMPRR LMSDCMP LMSDCMPR LMSDECF LMSDECFR LMSDECR LMSDECRR LMSDUCMP LMSDUCMPR LMSUCMP LMSUCMPR LMTOTALLOGS -New variables added to BBMQCHAN (RTIN='E5'x) dataset: CHLRVER CHLCRTLBL -New variables added to BBMQQUES (RTIN='E6'x) dataset: QSBACKD QSBACKT QSCBD QSCBT QSCMITD QSCMITT QSCTLD QSCTLT QSFBACKD QSFBACKT QSFCBD QSFCBT QSFCMITD QSFCMITT QSFCTLD QSFCTLT QSFINQD QSFINQT QSFPUT1D QSFPUT1T QSFSETD QSFSETT QSFSTATD QSFSTATT QSFSUBD QSFSUBT QSIBOCPU QSIBOELA QSICCPU QSICELA QSICMCPU QSICMELA QSIGCPU QSIGELA QSIICPU QSIIELA QSIMAXOI QSIMAXOO QSIMAXOT QSINQD QSINQT QSIOCPU QSIOELA QSIP1CPU QSIP1ELA QSIPCPU QSIPELA QSISBACK QSISBCKR QSISBCPU QSISBELA QSISCB QSISCMIT QSISCMTR QSISCPU QSISCTL QSISELA QSISINQ QSISSET QSISSTAT QSISSUB QSITFGET QSIUBACK QSIUBCKR QSIUCB QSIUCMIT QSIUCMTR QSIUCTL QSIUINQ QSIUSET QSIUSTAT QSIUSUB QSPUT1D QSPUT1T QSRBOCPU QSRBOELA QSRCCPU QSRCELA QSRCMCPU QSRCMELA QSRGCPU QSRGELA QSRICPU QSRIELA QSROCPU QSROELA QSRP1CPU QSRP1ELA QSRPCPU QSRPELA QSRSBACK QSRSBCPU QSRSBELA QSRSCB QSRSCMIT QSRSCPU QSRSCTL QSRSELA QSRSINQ QSRSSET QSRSSTAT QSRSSUB QSRSTCKG QSRSTCKP QSRTFGET QSRUBACK QSRUCB QSRUCMIT QSRUCTL QSRUINQ QSRUSET QSRUSTAT QSRUSUB QSSBOCPU QSSBOELA QSSCCPU QSSCELA QSSCMCPU QSSCMELA QSSETD QSSETT QSSGCPU QSSGELA QSSICPU QSSIELA QSSMAXOI QSSMAXOO QSSMAXOT QSSOCPU QSSOELA QSSP1CPU QSSP1ELA QSSPCPU QSSPELA QSSSBACK QSSSBCKR QSSSBCPU QSSSBELA QSSSCB QSSSCMIT QSSSCMTR QSSSCPU QSSSCTL QSSSELA QSSSINQ QSSSSET QSSSSTAT QSSSSUB QSSTATD QSSTATT QSSTFGET QSSUBACK QSSUBCKR QSSUBD QSSUBT QSSUCB QSSUCMIT QSSUCMTR QSSUCTL QSSUINQ QSSUSET QSSUSTAT QSSUSUB -New variables added to BBMQCFAC (RTIN='E8'x) dataset: CFBERBA2 CFBSRBA2 CFCONLOS CFDSBLK CFDSBUFS CFDSEXPN CFDSGRP CFLOGS CFOFFLDU CFOFL1S CFOFL1T CFOFL2S CFOFL2T CFOFL3S CFOFL3T CFOFLD CFPARM CFQSYN CFRECAUT -New variables added to BBMQAPPL (RTIN='E9'x) dataset: APSSTCKB -Occasional compressed records were found in uncompressed data files, so the previous STOP on compressed record is replaced by a RETURN so the non-compressed records after a compressed record will be processed. -All records have VERSREL=6.1, but their VERSION IS 5.2. -Dec 10 Revisions: Records E1x/E5x/E6x don't have all of the fields in the DSECT, causing INPUT STATEMENT EXCEED; logic was inserted to read only what's there, but record E4x's eyecatcher is @89 while all other record's is found @85, and E4x record has ENTL=1192 but data visibly shows the repeated segment length is only 1136. Hardcoded circumvention for these two inconsistencies. Interspersed Compressed records in an uncompressed file are still skipped, but new message prints the first 25 so you can see the pattern of these records. -BMC found that PTF BPL2290 caused problems with history record decompression; BPL 2430 fixes for MVMQ 5.1 and BPL2429 fixes for MVMQ 5.2 -Jan 2, 2015: After Installing BMC Mainview PTF's BPL2427 and BPL2430, 753 (out of 1003) 'E6' records were missing the last four bytes of the last segment, with 2596 vice 2600 bytes visible in that last segment. These short records are now detected and reported but that last segment will be lost until a subsequent PTF corrects. Thanks to James S. Swinarski, Credit Suisse, USA. Change 32.222 IMS56FA log record is not output if NMSGPROC is missing, TYPEIMST but that MXG decision is now rescinded, and all IMS 56FA Sep 15,2014 records are output, having found pairs of records for some transactions, with the second record having a small amount of CPU time (1st 1265 microsec, 2nd 238 microsec) and elapsed/service time, but CALLMSGU=0 so NMSGPROC=. All of these second records have TCPCFLGS='01'x which is a "TERM THREAD SYNCPOINT" reason for the record, so I presume the second record is the CPU time associated with the SYNCPOINT and is output so that time is not lost. Now, you must use NMSGPROC to count transactions, as the second syncpoint records are not a new transaction. Note: the test for NMSGPROC GT 0 is in the MACRO _EIMS56G overridden in comments in TYPEIMST, so you will need to delete that macro in the SYSIN in your TYPE56FA job.
VMACSYNC record processing when MACRO _SYNCID nnn % was specified. Sep 14, 2014 That new message only tested _IDSYNC, but the original _ID macro for SYNCSORT was _SYNCID, which is still valid. Some early USER SMF _IDaaaa macros syntax was _aaaaID but the "standard" is now _IDaaaa so those with the old _aaaaID were enhanced to support either _ID macro name, but the new "_ID=512" protection didn't support both of the SYNCSort macro names. Now supports either syntax. SEE CHANGE 32.234. Thanks to Betty Wong, Bank of America, USA.
BLDSMPDB valid interval value and has been removed from comments. VMXGALOC -You can now suppress the creation of the DB2 and CICS Sep 10, 2014 directories by specifying DB2KEEP=0 and/or CICSKEEP=0. Sep 25, 2014 -New parameters BASECICS and BASEDB2 will allow you to route DB2 and/or CICS datasets to a different drive. Change 32.219 TMON/CICS variables TAUSRWCT and TAUSRWTM were reversed. VMACTMO2
Sep 9, 2014 Thanks to Michael Oujesky, DTCC, USA. ====== Changes thru 32.218 were in MXG 32.09 dated Sep 9, 2014========= Change 32.218 -BVIRVERS=04, dataset BVIR301 values were wrong due to two VMACBVIR fields not INPUT for that Hydra version. Sep 9, 2014 -DO Loop variable _I_ is replaced by kept variables Sep 25, 2014 CACHEPARTNR created in BVIR301 TDUCONTNR created in BVIR321 CSPNR created in BVIR322 to identify the segment number -DVAVDTMB/WRMB/RDMB now multiplied by 1048576 vs 1058576. Thanks to Giuseppe Giacomodonato, EPVTECH, ITALY.
VMXGODSO Some thoughts on ODS: Initially, VMXGODSO was thought Sep 7, 2014 to be a tool to make use of ODS features easy, but you may do better learning the ODS command set. VMXGODSO does work, but it does not touch all of the ODS issues and features, and many ODS statements are not supported. But it can be used for simple cases, like wrapping around a PROC to send its output to a CSV or EXCEL file, or if you want to send many reports to a single output, a PDF for example. So you could use VMXGODSO, externally, to run GRAFWRKX and GRAFCEC and send both ODS plots to a single PDF: %VMXGODSO(ODSTYPE=PDF,FILE=c:\mxg\daily.pdf); %GRAFWRKX: %GRAFCEC; RUN;
%VMXGODSC; RUN;
and not use the parameters for ODS built into those GRAFXXXX macros. Change 32.216 Cosmetic. Some unneeded MXGNOTES are suppressed, unless VMXGOPTR %LET MXGEXIMSG=YES; enables them for diagnostics. Sep 7, 2014
VMACZOSA for ZOSATYPE=32 should have been ZOSATYPE=50. Sep 6, 2014 The test for invalid ZOSAXDLN is again bypassed. Thanks to Jerome Vitner, EXPERIAN, ENGLAND. Change 32.214 DB2 V10 variable QW0225DMH is now INPUT and QW0225AR is VMACDB2 now INPUT for DB2 V11. See Change 32.231. Sep 3, 2014 Thanks to Kerry Sommers, Deere & Company, USA. Change 32.213 -WARNING: MULTIPLE LENGTHS FOR VARIABLE X1 in ANALDB2R ANALDB2R is not new; the X1 that shouldn't have been there isn't. Sep 1, 2014
ASMTAPEE System-level information each SMF interval, writing new EXTMNAS1 subtypes of the existing TMNT SMF record, synchronized. EXTMNAS2 The subtask does not "monitor" the system, but sleeps EXTMNSYQ until awoken at each interval end, reads EXISTING data IMACTMNT fields and writes them to SMF, and goes back to sleep. VMACTMNT This FIRST use of the ML-54 "SYSTEM" subtask captures two VMXGINIT sets of data: Sep 9, 2014 - ASM Page Data Set SLOT usage, with total, and by ASID - MSU used by each IMPORTANCE level, IWMWSYSQ data. Two new subtypes create these three new datasets:
DATASET DATASET DATASET SUFFIX NAME LABEL SUBTYPE
High Virtual counts: Frames in Use, Shared AUX for Non-VIO, for AUX, and for AUX for SCM, Common Aux, Common AUX SCM, Common in USE. Shared Non-VIO, Shared AUX, SHARED AUX SCM, Frames in use, Shared Non-VIO/High Virt Shared High Virt AUX SCM, Shared Non-VIO/High Virtual Aux, HV AUX SDM TMNAS2 TYPEASMA MXG SYSTEM ASM PAGE DATA SET SLOT ASID 10 Frames in Use, VIO/non-VIO slots, High Virtual AUX Slots for each ASID. TMNSYQ TYPESYSQ MXG SYSTEM WLM PERF/CAPACITY IWMWSYSQ 11 MSU consumed last 60, 180, 600 seconds by each Importance Level and Unused MSU, Free CSA & free ECSA, and SU_SEC value. -ASM subtask is a per-system monitor and must be enabled on ALL systems, creating records on all systems. -SYSQ subtask is a sysplex-wide monitor, so it should only be enabled on one SYSTEM in each SYSPLEX; duplicate data would be created if enabled on multiple systems within a SYSPLEX. However, the number of SYSQ SMF records is very small, one per system per SMF interval, so you may want to enable SYSQ to write records on each system, so that you have the identical ASMTAPEE/MXGTMNT on each SYSTEM, and remove the duplicate SYSQ using EXTYSYSQ exit member when you read the SYSQ SMF records. -The ASM and SYSQ subtasks are enabled by default in ML-54 ASMTAPEE, by adding them to OPTIONS2, but they can be disabled by removing them from OPTIONS2, or thru PARMs on the // EXEC PGM=MXGTMNT statement, or thru MODIFY commands issued from the console. Full documentation is in the comments in ASMTAPEE. Change 32.211 Cosmetic. Variable PAGETYPE='SCM' is set when SCMPGTYP VMAC75 is true, replacing a blank value in PAGETYPE. Aug 28, 2014 Thanks to Douglas C. Walter, CitiCorp, USA. Change 32.210 -Support for SMF 50 INCOMPATIBLE change in z/OS 2.1 that EXTY501 inserted four bytes that corrupted all values after that EXTY50R insertion in the TYPE50 dataset. EXTY50W -Redesign of TYPE 50 support to create separate datasets EXTY50I with only the relevant variables for that event, now that EXTY50O I realize it is no longer a single purpose VTAM record! EXTY503 These seven new datasets are created: EXTY504 DDDDDD DATASET Description IMAC50 TY501 TYPE501 01 CHANNEL TO CHANNEL Table 52 Subtype 1 VMAC50 TY50R TYPE502R 02-02 READ MPCNAME Table 53 Subtype 2 VMXGINIT TY50W TYPE502W 02-02 WRITE MPCNAME Table 53 Subtype 2 AUG 28, 2014 TY50I TYPE504R 02-04 READ MPCNAME Table 54 Subtype 2 TY50O TYPE504W 02-04 WRITE MPCNAME Table 54 Subtype 2 TY503 TYPE503 03 TCP CONNECTION Table 55 Subtype 3 TY504 TYPE504 04 SNA CONTROLLER Table 51 Subtype 4 -Dataset TYPE50 will continue to be created for backwards compatibility, but it is very messy with all possible variables in every observation, so there are always lots of variables with missing value; using these new datasets will provide much simpler analysis. Thanks to Jim Sherpey, Bank of America, USA. Change 32.209 -Support for SMF 103 HTTP Apache Server 8.5.5.0 Subtype 13 EXTY103D and 14 create new datasets: EXTY103D DDDDDD DATASET Description Yüklə 28,67 Mb. Dostları ilə paylaş:
|