* copyright (C) 1984-2019 merrill consultants dallas texas usa
Yüklə 28,67 Mb.
|
|
MXGERROR: YOU SPECIFIED BUILDPDB=NO BUT DID NOT; MXGERROR: SPECIFY A USERADD= SO UTILBLDP DOES; MXGERROR: NOT HAVE ANYTHING TO DO AND WILL SET; MXGERROR: THE OUTFILE TO A NULL DATASET; and since the OUTFILE is NULL a following include of that file will not fail since it is empty. Change 33.040 Support for SystemWare XPTR 5.2 subtype 140 record which VMACXPTR now contains what used to be in their subtype 40 record. Feb 18, 2015 Thanks to Phil Grasser, Norfolk Southern, USA. Change 33.039 Summarization of Mobile Work CSV files to combine hours MOBWRKSU that were split between two CSV files. All files from a Feb 19, 2015 family are concatenated to the //CSVINPUT DD and the summarized CSV data is written to //CSVSUMRY DD. -CSVSUMRY was originally DISP=MOD, to collect the input from four sources, but DISP=MOD can not be used for a PDSE member, so the MOD file is written to an internal catalog file, which is then copied to //CSVSUMRY. Change 33.038 Change 30.082 added protection for SMF type 60 record VMAC60 with no VVR segment, but the protection was insufficient, Feb 17, 2015 and an INPUT STATEMENT RECORD EXCEEDED error could still occur. Now, the remaining length is verified. Thanks to Tabbitha Flink, FirstData, USA.
ADOCMWLX -ADOCMWLX has the REPORT command to extract the fields MXG EXMWLXAP expects, so it must be used to create MXG's input data. EXMWLXCO -CPU Times in the MWLXGLOB dataset are presumed to be the EXMWLXDS per-CPU values in the record; they have been multiplied EXMWLXGL by the number of CPUs. Other times are as found. EXMWLXLA -ALL STORAGE variables contain BYTES and are formatted EXMWLXPR with MGBYTES for total bytes and MGBYTRT for byte rates. EXMWLXTT Fields in the record in KB have been converted to bytes. IHDRMWLX -IHDRMWLX or &MACMWLH can be used to select which TYPE IMACMWLX records are to be read. TYPEMWLX
TYPSMWLX VMACMWLX
VMXGINIT Feb 13, 2015 Feb 20, 2015 Thanks to Roman Gudz, Penske, USA. Change 33.036 -RACF SMF 80 record, SMF80TP2=301, Command Segment Data VMAC80A segments "CTOKENKY" and "CTOKENTM" are supported. Feb 12, 2015 -The variable length TOKxxxxx fields do not document their max length, and MXG can get INPUT RECORD EXCEEDED errors if an input field is longer than the defined length in its $VARYINGnn informat. These variables are increased to $VARYING64 to hopefully avoid the error condition: TOKCOMPANY TOKCOUNTRY TOKFNAME TOKMNAME TOKLNAME TOKMCARRIER TOKMCANSWR1 TOKMCANSWR2 TOKMCANSWR3 TOKMCBADCNT Thanks to Phil Grasser, Norfolk Southern, USA.
VMAC80A segment "AUTOGID" decodes 4 variables but their contents Feb 11, 2015 is not known at this time: TOKMAUTOGID01 TOKMAUTOGID02 TOKMAUTOGID03 TOKMAUTOGID04 Thanks to Harald Seifert, HUK-COBURG, GERMANY. Change 33.034 The variables SYSABEND and USRABEND were created but not VMACIMS decoded in the IMS56FA transaction dataset. Now, COMPCODE Feb 10, 2015 is decoded into these two useful IMS transaction abend code variables in all record that contain it. Thanks to Raymond J. Smith, OPTUM, USA.
WEEKBLD datasets to sequential format data libraries without the Feb 10, 2015 need to use DISP=MOD. Without MOD and 731, each write overwrote the prior, leaving only the last dataset. Change 33.032 -Variables SMF70WTI/SMF70WTS/SMF70WTU (new WTI-s) were ok VMAC7072 in dataset TYPE70EN, but were always zero in TYPE70PR. VMXG70PR TYPE70PR is now corrected, but those values only exist Feb 8, 2015 in the "THIS PART" TYPE70 records, so you must read ALL system's TYPE70s to properly populate these variables. They are always missing in the LPARNAME='PHYSICAL's -VMXG70PR is updated so the PDB.ASUM70LP and PDB.ASUMCELP LPAR Summary datasets contain those variables, but only if RMF type 70 records from ALL systems are input.
BLDSMPDB incorrectly calculate the start of week, depending on the VMXGALOC case of your typed code. Now when you specify FORCEDAY Feb 19, 2015 to rerun SMF data, ZDATE will be set to FORCEDAY+1. Feb 26, 2015 Error was introduced in MXG 32.11, Change 32.253. And all of the tests where casing can impact are now now protected with %UPCASE compiler functions, but the members remain lower case because of earlier problem in casing in Linux systems; it might not still be required, but not worth the exposure and time spent for no value. Change 33.030 Mainview for MQ 5.2 support is corrected; the wrong byte VMACBBMQ was used to detect compressed records, causing spurious Feb 5, 2015 messages that individual records were compressed. Thanks to Greg Tuben, BMC Software, Inc., USA. Change 33.029 The iSeries a/k/a AS400 QAPM "suffix" was replaced by the TYPEQACS "QACS" suffix years ago, but those old QAPM members are Feb 4, 2015 now removed; they were causing false positives in the MXG QA tests, and served no purpose. All of the xxxxQAPM are deleted, and these exit members that weren't carried into the QACS product were also deleted. dddddd exit dddddd exit token member token member _WQAPASY EXQAPASY _WQAPPOO EXQAPPOO _WQAPBSC EXQAPBSC _WQAPRWS EXQAPRWS _WQAPDDI EXQAPDDI _WQAPSTD EXQAPSTD _WQAPECL EXQAPECL _WQAPSTL EXQAPSTL _WQAPFRL EXQAPFRL _WQAPSTY EXQAPSTY _WQAPIDL EXQAPIDL _WQAPX25 EXQAPX25 _WQAPJOB EXQAPJOB _WQAPLAP EXQAPLAP While the SUFFIX QAPM is retired, the names of all of the datasets created by TYPEQACS are unchanged and all start with QAPMxxxx.
COMPALL creates new dataset: EXTYNTMU DDDDDD DATASET DESCRIPTION FORMATS TYNTMU NETMENU SMF NETMENU RECORD DATABASE IMACNTMU TYPENTMU
TYPSNTMU VMACNTMU
VMXGINIT Feb 3, 2015 Thanks to Joe Babcock, General Motors, USA.
VMXGDELV Variables, used only in the QA JOB when new code caused Feb 3, 2015 an unexplained error (the "TURN SPOOL ON" message) in Step 34, but there was no error when Step 34 ran first. By clearing all Global Macros before Step 34, the actual unresolved macro variable error was then unmasked, but maybe only accidentally! To clear and re-initialize %VMXGDELV;%INCLUDE SOURCLIB(VMXGINIT);%VMXGINIT; is needed so that the GLOBAL statement in VMXGINIT is re-executed. And because %VMXGDELV cleared the flag variable MXGINIT, %VMXGINIT prints the Welcome To MXG message, instead of the the re-initialization message. Change 33.026 Reserved Change Number. Change 33.025 -Variable QBSTBPIN='BUFFER*POOL*IO*INTENSITY' based on IBM ANALDB2R Tivoli calculation is added to DB2STATB dataset and also VMACDB2 ANALDB2R reports the value. VMXGDBSS QBSTBPIN=SUM(QBSTRIO,QBSTSPP,QBSTLPP,QBSTDPP,QBSTIMW, Feb 3, 2015 QBSTWIO)/QBSTVPL; Feb 26, 2015 -ASUMDBSB now reports both average and maximum QBSTBPIN. -Debugging test IF SYSTEM='SYSG' removed from DB2B1 Buffer Pool Statistics reports -Feb 26: Protected QBSTBPIN from divide by zero error when QBSTVPL=0 when no buffers were allocated in the pool. Thanks to Tim Hare, Hare Systems Support, USA. Thanks to Jonathan Bathmaker, Southern California Electric, USA. Thanks to Tom Buie, Southern California Electric, USA.
UTILRMFI The UTILRMFI utility is cited in error messages that your UTILWORK CPUTM does not match CPU72TM when building RMFINTRV: run VMXGRMFI UTILRMFI to examine your VMXGRMFI workload definitions to Feb 3, 2015 find what work was overlooked. However, if there is any work in the SYSOTHER Service Class (which itself is an error in your site's WLM Classification Rules), that may cause the mismatch, since that work probably also doesn't have a Reporting Class. Now you are alerted of the WLM definition error. Note that SYSOTHER is also undesired because that work is MTTW DISCRETIONARY, so whatever is falling thru is NOT going to get much service. Change 33.023 Support for new z/OS z13 hardware metrics (COMPATIBLE): FORMATS -APAR OA43493 RMF Support Cryptographic Express5S cards: VMAC30 New Crypto Processor types in R7023CT/24CT/25CT variables VMAC7072 decoded by $MGRMFCY/MGRMFCZ formats in datasets TYPE7002, VMAC74 TYPE70X2, and TYPE70Y3. VMAC113 New variables in TYPE70Y2 dataset: VMACVMXA R702AMGB='DATA BYTES*FOR*GENERATE*AES MACS' Jan 29, 2015 R702AMGC='CALLS TO*GENERATE*AES MACS' ASUM113 R702AMGI='INSTRUCTIONS*TO GENERATE*AES MACS' Feb 25, 2015 R702AMVB='DATA BYTES*FOR*VERIFIED*AES MACS' May 21, 2015 R702AMVC='CALLS TO*VERIFY*AES MACS' R702AMVI='INSTRUCTIONS*TO VERIFY*AES MACS' R702DEGC='CALLS*TO GENERATE*ECC*SIGNATURES' R702DEVC='CALLS*TO VERIFY*ECC*SIGNATURES' R702DRGC='CALLS*TO*GENERATE*RSA*SIGNATURES' R702DRVC='CALLS TO VERIFY*RSA*SIGNATURES' R702FPDB='DATA BYTES*DECIPHERED*USING FPE' R702FPDC='CALLS TO*DECIPHER*USING FPE' R702FPDI='INSTRUCTIONS*TO DECIPHER*USING FPE' R702FPEB='DATA BYTES*ENCIPHERED*USING FPE' R702FPEC='CALLS TO*ENCIPHER*USING*FPE' R702FPEI='INSTRUCTIONS*TO ENCIPHER*USING FPE' R702FPTB='DATA BYTES*TRANSLATED*USING FPE' R702FPTC='CALLS TO*TRANSLATE*DATA*USING FPE' R702FPTI='INSTRUCTIONS*TO TRANSLATE*USING FPE' -APAR OA44502 RMF Support Coupling Channel Path Type CS5: Format $MG074OM updated and applied to variable R744HOPM in dataset TYPE74HO. -APAR OA30563, Enhanced SMF 30 and 89 recording; metrics listed in the updated APAR were added in Change 28.175. The APAR adds new MAXEVENTINTRECS with zero default, so the new capacity changed interval 30/89 records will not be created with that zero value. -APAR OA35175, SMFPRMxx options DSPSIZMAX; the TYPE23 dataset updates were actually made in Change 29.177. -APAR OA45985, RMF support for zHYPERWRITE in 2107 CU. Adds new variable HYPERWRT='Y' to TYPE74 if requested. -SMF 113 has new z13 counters and revised calculations, including RNI and its component variables: RNI=2.6*(0.4*L3P+1.6*L4LP+3.5*L4RP+7.5*MEMP)/100; Corrections made March 5. However, the new counter descriptions do not yet exist. Note that variables L15P L2LP L2RP are always missing values on the z13. Without this change, negative RNI values for z13 even in non-SMT mode were created. -May 21: ZEDC added to the label for those eight metrics.
DOC //SYSPRINT DD RECFM=FBA,LRECL=133,BLKSIZE=13232, but with Jan 28, 2015 only the message AMA504I RETURN CODE: 12 and no clue. Removal of the DCB attributes successfully untersed. However, it was NOT the non-multiple of LRECL for BLKSIZE that was the error. A prior IDCAMS step had a //SYSPRINT DD that JES3 made into a temporary DSNAME, and the same DSNAME was then used in the TERSE step, but the different BLKSIZE caused the conflict and RETURN CODE. That's my story and I'm stikin' to it until I learn more about why JES3 passed a temporary DSNAME for //SYSPRINT. However, there is NO need for DCB attributes on SYSOUT= DDnames, so to avoid ANY future exposure, I've removed all DCB from those DDs in all MXG examples. -All //OUTPUT DDs for PGM=FTP were removed, as //SYSPRINT is used for all messages when there is no //OUTPUT DD. Thanks to Tom Adams, State Farm Mutual Automobile Insurance, USA.
ADOCRMFV RCD tables as invalid due to a faulty calculation, with a ASMRMFV RMFV035W **WARNING:159 RCD TABLES SKIPPED DUE TO ERRORS CLRMFV and ultimately set a Return Code 16. Should be very rare JCLRMFV since most sites DO define Reporting Classes. JCLCRMFV -Change 32.055 did not fully resolve the RED table issue: JCLdRMFV RMFV035E:1 RED TABLES SKIPPED DUE TO INVALID ID OR FLAG Jan 26, 2015 the Invalid Resource bit may be set for Processor entries in the RED table in a MINTIME interval due to normal processor change events, so these are not true RED table errors and ASMRMFV was incorrectly issuing an RMFV035W warning message and Return Code 16 for this transient and insignificant return code. -ASMRMFV now will NOT alter the Return Code, issuing a distinct RMFV035I information message for this condition regardless of the setting of the TABERR= option: RMFV035I -IGNORED-: nnnnn RED TABLES SKIPPED DUE TO INVALID PROCESSOR RESOURCE FLAG--- -After Change 32.289 the number of Policy Indexes Used in informational message RMFV028I as triggered by the INDEXES option could be incorrect. -The MXG table was shown in the Detail Report with the NOSHOWZERO option in effect even when zero records were output. This table is only output once for the first RMF III data set processed. -The RMFV006I FILTER message for ASI, DVT, and CFI table specific options is now always shown even if none of these tables are selected for a minor performance gain, as selection logic for what to display was revised. -Unnecessary checks for I/O errors on the SYSPRINT file are removed. An ABEND occurs with any I/O error on this file so the additional checks were unneeded. -Updates for revised messages RMFV006I and RMFV035I are made to Section 12 "Messages" in ASMRMFV and ADOCRMFV documentation members. -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. Thanks To John Wynn, IBM Global Technology Services, USA. Change 33.020 AS/400 iSeries fixed-length records change lengths with a VMACQACS new version, so you may have to change the LRECL in your Feb 2, 2015 FILENAME statement on ASCII or in your JCL on z/OS. The length change is usually detected with MXG messages on the log with "PAD RECORD FOUND AND DELETED". Now, you can specify %LET MXGABND=NNNN; in //SYSIN to force USER ABEND NNNN if more than 10 PAD records are found. Use 0001 thru 4096 for the NNNN value. New messages identify the filename and dsname being read. Thanks to Raymond J. Smith, OPTUM, USA. Change 33.019 Variables TOKQUEST1-TOKQUEST3, TOKCANSWR1-TOKCANSWR3 and VMAC80A CBADCNT are added to dataset TYPE80TK. Jan 21, 2015 Thanks to Phil Grasser, Norfolk Southern, USA. Change 33.018 -RACF TYPE8025 Record with DTP=30 segment with 'C0'X value VMAC80A for the status ('80'x=ACTIVE,'40'x=BACKUP) caused ERROR: Jan 21, 2015 INVALID RACF ID=80 RACFTYPE=30 SEGMENT SKIPPED, because I Feb 5, 2015 didn't realize a DSNAME could be both ACTIVE and its own Feb 23, 2015 BACKUP, so I had created two sets of variable names, with RACFDBxx for the ACTIVE segment and RACFDKxx for BACKUP. Test for '80'x first, now stores into RACFDBxx variables, or '40'x or 'C0'x store into RACFDKxx variable names. -For BACKUP-ONLY ('40'x) segments, both UNIT NAME/VOLSER are not populated, while both are in the ('C0'X) record. -Feb 23: TOKMPROCUSERMAX variable decoded. See Change 33.084. Thanks to Karl Lasecki, American Chemical Society, USA.
UTILBLDP DB2ACCTx datasets to a single //DB2ACCT LIBNAME, and uses Jan 21, 2015 MACKEEPX in UTILBLDP to control dataset destinations and sorts and could also be used to limit variables. The DB2KEEP= parameter in BLDSMPDB is used to tailor the number of generations retained to control the storage needed for these normally-large datasets (ASCII only with AUTOALOC=YES). Change 33.016 Support for z/VM 6.3 on z13 processor (INCOMPATIBLE). VMACVMXA -The PRCMFC 5.13 record's new CSVN=4 hardware counters Jan 20, 2015 caused MXG code to fail with "BROKEN CONTROL RECORD". Circumvented by skipping the extra 24 bytes, await the actual documentation of those new counters before decoding. IN PROGRESS.
VMACDCOL is added to dataset DCOLDSET; the time part wasn't in the Jan 20, 2015 original record, but was added by APAR OA30006 in 2009. DCDCREAT is only populated if both DCDCREDT and DCDTIMEC are non-zero. Thanks to Scott Barry, SBBWorks Inc., USA. Change 33.014 Support for z/13 processor increase to 85 LPARs was added VMAC7072 in MXG 32.06, Change 32.162, July, 2014, changing the Jul 27, 2015 statement to ARRAY S70LPCP {86} _TEMPORARY_; to support Jan 19, 2015 85 "real" LPARs plus the PHYSICAL "LPAR". Support for 255 engines was added silently in MXG 31.04 in 2013.
EX111WS -Variable added to CTGSE System Environment dataset: EX111WSX CTGC31MAX='LIMIT OF*USED*MEMORY*ELOAL' IMAC111 -Two new segments, WS and WSX create new datasets: VMAC111 DDDDDD DATASET DESCRIPTION VMXGINIT 111WS TY111WS WEB SERVICE ALL Jan 20, 2015 111WSX TY111WSX WEB SERVICE INSTANCE Thanks to David Marone, SGS, ITALY. Change 33.012 Unused Change Number. Change 33.011 -INPUT STATEMENT EXCEEDED, DFSORT SMF 16, OFFSTAT=640 and VMAC16 LENSTAT=64, but record is TRUNCATED with LENGTH only 636, Jan 14, 2015 so there is no STAT segment in the record. A PUTLOG ERROR message is written for the first three instances, the TYPE116 dataset is output, but variables RECIND01-RECIND16 will be missing values. This text will be revised when an APAR exists. DFSORT Release 2.01. -These variables added in DFSORT 2.01, but not listed in the manual's changes, nor marked as new with the expected vertical bars, are now added to dataset TYPE16: ICEDYINC='INITIAL*INCREMENT' ICEDYMAX='FINAL*EXPMAX*DYNAMIC*VALUE' ICEDYOLD='FINAL*EXPOLD*DYNAMIC*VALUE' ICEDYRES='FINAL*EXPRES*DYNAMIC*VALUE' ICETUNE ='TUNE*VALUE*IN*EFFECT' Thanks to Kerry Sommers, John Deere, USA. Change 33.010 No errors have been reported, but ARRAY WORDS statements VMAC112 raised conflicts if TYPE112/TYPEEZAM/TYPETMNT are read VMACEZSM together; only one ARRAY NAME can be used in a DATA step. VMACTMNT These are temporary arrays for parsing with different Jan 12, 2015 dimensions and variable lengths, so the ARRAY NAMEs are made unique to eliminate the exposure. Change 33.009 IBM Updates to SMF 14/15 for zEDC Compression Indicator FORMATS SMF30XF1 adds values and is now decoded by $MG014ED: VMAC1415 '0'='0:SIZE VALUES INVALID' Jan 12, 2015 '1'='1:COMPRESSION REJECTED' '2'='2:ZEDC WRITTEN UNCOMPRESSED' '3'='3:ZEDC SOFTWARE DECOMPRESSED' and new SMF14CMPTYPE='Compression Type' variable is added and decoded by MG014CT format: 0='0:NOT COMPRESSED/UNKNOWN' 1='1:GENERIC' 2='2:TAILORED' 3='3:ZEDC' Thanks to Scott Barry, SBBWorks Inc., USA.
ONLYCIDS Statistics dataset, that contains interval CPU time for Jan 12, 2015 each CICS Interval, and can be used for CPU Time Metric in IBM Mobile Work Discount Thanks to Scottie Long, Navy Federal Credit Union, USA.
TYPERMF not updated when the 2005 "SPLIT70 logic" created the Jan 11, 2015 TYPE70EN dataset, written to PDB by default. Like the two existing %LETS for PTY70 and PTY70PR in TYPERMF, a %LET PTY70EN=WORK; is needed so //PDB is not used. Thanks to George Baranoff, Safeway, USA. Change 33.006 -SAS Error "No MKLEs found" and "ERROR: VM 1319:" is a DOCUMENT virtual storage issue, usually too small a REGION= size, Jan 11, 2015 but also a known SAS Version 9.2 defect, see below. -The MXG QA "BUILDPDB" JOB (with a few added SMF types) increased from 130MB in 31.31 to 140MB in 32.32. -Each new MXG version adds new variables, datasets, and code which cause the REGION size to increase a little. Unfortunately, the z/OS REGION is a LIMIT not a BENEFIT, so you get to periodically increase the REGION= value, and/or check with your site's "REGION SIZE POLICE" to find their actual limit for REGION=0M, and/or what JOB CLASS/RACFUSER will let you allocate the REGION needed. -There is NO ACTUAL RESOURCE "consumed" by REGION size. -The actual step region size is in the IEF032I message on joblog for each step: EXT: 130,852K SYS: 12,088K EXT is the region size "Above the 16MB Line" and SYS is the size of the Private Area on this system. Their sum is the region size allocated to this step, or 130,852+12,088=142,940K=140MB REGION size. -The SAS log shows "And 130852K Above the Line" matching the EXT value exactly. SAS also reports a value for the "Below the Line", but it is (small) amount used, not the actual size of the Private Area, which must be included in the REGION limit. -One site moving from an ancient MXG version still had REGION=6M on their JOB Card from the prior millennium. You get the system default of 32MB Above the Line plus the Private Area, or about 42MB, but because their daily job was split into parallel pieces, it actually had worked (accidentally?) until 32.32. -SAS needs REGION size for referenced FORMATS, for dataset buffers, for arrays, for SAS VIEWs, and for the CICSIFUE SMFEXIT=CICS INFILE exit to load. -SAS Version 9.2 is Class C support from SAS and it had fatal errors in virtual storage on z/OS that were ONLY corrected by SAS 9.3 or 9.4. -One 9.2 site's tailored BUILDPDB ran in 130MB with 30.01, needed 159MB with 32.32, but when the CICSIFUE exit was enabled, the job failed and IEF032I reported only 97MB had been allocated. Rerun with 32.32 and SAS 9.4 and WITH the exit needed 179 MB, and ran with no error. -MXG 33.03 QA BUILDPDB on z/OS 2.1, SAS 9.4(TS1M2) needed: 142M with View Disabled and no CICS EXIT Enabled. 147M with the View Enabled (default), and no Exit. Yüklə 28,67 Mb. Dostları ilə paylaş:
|