|
|
| Author |
Message |
ksk
Senior Member
Joined: 08 Jun 2006 Posts: 315 Location: Pune, India
|
|
|
|
Hi,
One JOb in our shop is concatenating 5 files and sorting on the following creteria.
| Code: |
SORT FIELDS=(9,262,CH,A)
SUM FIELDS=NONE
|
Actual record size is 271 and sort was done on the whole record except 1st 8 fileds. What I feel is 9 to 37 characters are quite enough to SORT the record as 38 to 271 has amount fields.
This single step is taking 3 hours clock time to execute. Could this happen due to length 262 characters mentioned in the SORT card or it is only due to removing duplicate records? |
|
| Back to top |
|
 |
References
|
|
 |
gcicchet
Senior Member
Joined: 28 Jul 2006 Posts: 696
|
|
|
|
Hi,
can you provide the SYSOUT messages from the job and job log ?
Gerry |
|
| Back to top |
|
 |
ksk
Senior Member
Joined: 08 Jun 2006 Posts: 315 Location: Pune, India
|
|
|
|
Following is SYSOUT.
| Code: |
ICE000I 1 - CONTROL STATEMENTS FOR 5694-A01, Z/OS DFSORT V1R5 - 00:40 ON WED AU
SORT FIELDS=(9,262,CH,A)
SUM FIELDS=NONE
ICE201I E RECORD TYPE IS F - DATA STARTS IN POSITION 1
ICE751I 0 C5-K26318 C6-K90007 C7-K90000 C8-K23476 E9-K90007 C9-BASE E5-K24705
ICE193I 0 ICEAM1 ENVIRONMENT IN EFFECT - ICEAM1 INSTALLATION MODULE SELECTED
ICE088I 1 ZNDD803N.STEP10 . , INPUT LRECL = 271, BLKSIZE = 27913, TYPE
ICE092I 0 MAIN STORAGE = (67108864,67108864,67108864)
ICE156I 0 MAIN STORAGE ABOVE 16MB = (66827704,66811320)
ICE127I 0 OPTIONS: OVFLO=RC0 ,PAD=RC0 ,TRUNC=RC0 ,SPANINC=RC0 ,VLSCMP=N,SZERO=Y
ICE128I 0 OPTIONS: SIZE=67108864,MAXLIM=1048576,MINLIM=450560,EQUALS=Y,LIST=Y,E
ICE129I 0 OPTIONS: VIO=Y,RESDNT=ALL ,SMF=FULL ,WRKSEC=Y,OUTSEC=Y,VERIFY=N,CHALT
ICE130I 0 OPTIONS: RESALL=0,RESINV=0,SVC=109 ,CHECK=N,WRKREL=Y,OUTREL=Y,CKPT=N,
ICE131I 0 OPTIONS: TMAXLIM=4194304,ARESALL=16384,ARESINV=0,OVERRGN=65536,CINV=Y
ICE132I 0 OPTIONS: VLSHRT=N,ZDPRINT=Y,IEXIT=N,TEXIT=N,LISTX=N,EFS=NONE ,EXIT
ICE133I 0 OPTIONS: HIPRMAX=OPTIMAL,DSPSIZE=0 ,ODMAXBF=0,SOLRF=N,VLLONG=N,VSAM
ICE235I 0 OPTIONS: NULLOUT=RC0
ICE084I 0 EXCP ACCESS METHOD USED FOR SORTOUT
ICE084I 0 EXCP ACCESS METHOD USED FOR SORTIN
ICE750I 0 DC 29338265693 TC 0 CS DSVUU KSZ 266 VSZ 266
ICE752I 0 FSZ=108259283 RC IGN=0 E AVG=276 0 WSP=38808415 C DYN=705255 5321
ICE751I 1 DE-K24705 D5-K24705 D3-K24705 D7-K24705 E8-K24705
ICE090I 0 OUTPUT LRECL = 271, BLKSIZE = 27913, TYPE = FB
ICE055I 0 INSERT 0, DELETE 42507764
ICE054I 0 RECORDS - IN: 108258986, OUT: 65751222
ICE134I 0 NUMBER OF BYTES SORTED: 29338185206
ICE165I 0 TOTAL WORK DATA SET TRACKS ALLOCATED: 705375 , TRACKS USED: 557595
ICE199I 0 MEMORY OBJECT STORAGE USED = 0M BYTES
ICE180I 0 HIPERSPACE STORAGE USED = 2097020K BYTES
ICE188I 0 DATA SPACE STORAGE USED = 0K BYTES
ICE052I 0 END OF DFSORT
|
|
|
| Back to top |
|
 |
gcicchet
Senior Member
Joined: 28 Jul 2006 Posts: 696
|
|
|
|
Hi,
you are sorting some large files, one way to speed this up and I'm sure there will other suggestions would be to have 5 jobs sorting individual files followed by a merge job.
Gerry |
|
| Back to top |
|
 |
gcicchet
Senior Member
Joined: 28 Jul 2006 Posts: 696
|
|
| Back to top |
|
 |
Dave Betten
New User
Joined: 24 Jan 2006 Posts: 13
|
|
|
|
You are sorting 29GB of data and using 2GB of Hiperspace. So that means the remaining 27GB has to be written/read from DASD sortwork. So 3 hours is not entirely unreasonable but it's tough to say since I don't know what kind of I/O response times you are getting. This job is going to be heavily dependent on I/O response times for SORTIN, SORTWORK and SORTOUT.
Reducing the key field from 9,262 to 9,37 will help reduce CPU time but I don't think it will have a huge difference in elapsed.
I'd suggest you add //SORTDIAG DD DUMMY to the jcl for this sort and then send the sysout to dfsort@us.ibm.com. That's the email address for our DFSORT hotline and I can take a look and see if there's any tuning we can do for this sort. Be sure to send the entire sysout, not just the DFSORT messages. |
|
| Back to top |
|
 |
ksk
Senior Member
Joined: 08 Jun 2006 Posts: 315 Location: Pune, India
|
|
|
|
Hi Dave/Gerry,
Thanks for your immediate responses. I have increased work files and space and executed in test region with Production files, now it's taking 45 mins.
We will incorporate the same in production and if we face any problem I will let you know.
KSK |
|
| Back to top |
|
 |
ksk
Senior Member
Joined: 08 Jun 2006 Posts: 315 Location: Pune, India
|
|
|
|
Dave,
I forgot to ask one question. can you let me know what is the use of the following statement.
| Code: |
//SORTDIAG DD DUMMY
|
|
|
| Back to top |
|
 |
Dave Betten
New User
Joined: 24 Jan 2006 Posts: 13
|
|
|
|
| The SORTDIAG causes additional diagnostic messages to be printed and these help us when performing indepth analysis. It does not do anything to improve performance so it has nothing to do with your decreased run time. How much Hiperspace did the sort use when you ran it in your test environment? It could be that there was less resource contention in the test environment and that's why it ran so much faster. Increasing work space reduces the risk of failure but shouldn't have that drastic an impact on elapsed time. Why don't you send me the entire sysout from that run and I can take a look. |
|
| Back to top |
|
 |
ksk
Senior Member
Joined: 08 Jun 2006 Posts: 315 Location: Pune, India
|
|
|
|
Here is the SYSOUT of changed JOB.
| Code: |
ICE201I E RECORD TYPE IS F - DATA STARTS IN POSITION 1
ICE751I 0 C5-K26318 C6-K90007 C7-K90000 C8-K23476 E9-K90007 C9-BASE E5-K31980
ICE143I 0 BLOCKSET SORT TECHNIQUE SELECTED
ICE250I 0 VISIT http://www.ibm.com/storage/dfsort FOR DFSORT PAPERS, EXAMPLES AN
ICE000I 1 - CONTROL STATEMENTS FOR 5694-A01, Z/OS DFSORT V1R5 - 03:36 ON THU AUG
SORT FIELDS=(9,21,CH,A)
SUM FIELDS=NONE
ICE201I E RECORD TYPE IS F - DATA STARTS IN POSITION 1
ICE751I 0 C5-K26318 C6-K90007 C7-K90000 C8-K23476 E9-K90007 C9-BASE E5-K31980
ICE193I 0 ICEAM1 ENVIRONMENT IN EFFECT - ICEAM1 INSTALLATION MODULE SELECTED
ICE088I 1 SSY361AB.STEP10 . , INPUT LRECL = 271, BLKSIZE = 27913, TYPE =
ICE092I 0 MAIN STORAGE = (67108864,67108864,67108864)
ICE156I 0 MAIN STORAGE ABOVE 16MB = (66827704,66811320)
ICE127I 0 OPTIONS: OVFLO=RC0 ,PAD=RC0 ,TRUNC=RC0 ,SPANINC=RC0 ,VLSCMP=N,SZERO=Y,
ICE128I 0 OPTIONS: SIZE=67108864,MAXLIM=1048576,MINLIM=450560,EQUALS=Y,LIST=Y,ER
ICE129I 0 OPTIONS: VIO=Y,RESDNT=ALL ,SMF=FULL ,WRKSEC=Y,OUTSEC=Y,VERIFY=N,CHALT=
ICE130I 0 OPTIONS: RESALL=0,RESINV=0,SVC=109 ,CHECK=N,WRKREL=Y,OUTREL=Y,CKPT=N,S
ICE131I 0 OPTIONS: TMAXLIM=4194304,ARESALL=16384,ARESINV=0,OVERRGN=65536,CINV=Y,
ICE132I 0 OPTIONS: VLSHRT=N,ZDPRINT=Y,IEXIT=N,TEXIT=N,LISTX=N,EFS=NONE ,EXITC
ICE133I 0 OPTIONS: HIPRMAX=OPTIMAL,DSPSIZE=0 ,ODMAXBF=0,SOLRF=N,VLLONG=N,VSAMI
ICE235I 0 OPTIONS: NULLOUT=RC0
ICE084I 0 EXCP ACCESS METHOD USED FOR SORTOUT
ICE084I 0 EXCP ACCESS METHOD USED FOR SORTIN
ICE750I 0 DC 29356353317 TC 0 CS DSVUU KSZ 25 VSZ 25
ICE752I 0 FSZ=108326027 RC IGN=0 E AVG=276 0 WSP=38832342 C DYN=687315 53216
ICE915I 0 MOFSZ=0,MOSZ=0,MOSYS=2058(2),MOSTG=3072,MEML=2058(10)
ICE916I 0 MOFR=0304,MOVR=UU
ICE996I 0 ESM=786432,ESO=786432,ESR=437478,ESP=4096,ESS=16384,CES=4374784,HSZ=52
ICE997I 0 HWSP=18248281,HMAX=786432,HES=786432,ASV=786432,EQ=I2,HN=1
ICE898I 0 OMAX=373004,NMAX=2269666,TMAX=2269666,CMAX=524288,HU=99,BUN=53216,MD=M
ICE880I 0 QP=128 QA=128 HI=192 LI=65 MI=192 TZ=524288 N1=524288 N2=524288 SZ=129
ICE889I 0 CT=MAX , SB=8, L=0, D=0000
ICE901I 0 W 0CPS13 01PS13 09PS13 0FPS13 04PS13 06PS13 07PS13 08PS13
ICE901I 0 W 0EPS13 03PS13 05PS13 10PS13 0DPS13 0BPS13 0APS13 02PS13
ICE902I 0 O PS10 I PS10
ICE897I 0 AVAL=0,BVAL=0,CVAL=0,ASV=0,HSZ=0,HM=M0,HN=3
ICE751I 1 DE-K24705 D5-K24705 D3-K24705 D7-K24705 E8-K24705
ICE900I 0 CON=5,MUV=0,VOL=132,ENU=0,SBK=5,SRC=223,VEM=0
ICE999I 0 PWK=15 PSP=687315 SWK=0 SSP=0 TWK=0 TSP=0 RWK=0 RSP=0 AWK=15 AWP=68731
ICE090I 0 OUTPUT LRECL = 271, BLKSIZE = 27913, TYPE = FB
ICE055I 0 INSERT 0, DELETE 42669466
ICE054I 0 RECORDS - IN: 108325735, OUT: 65656269
ICE134I 0 NUMBER OF BYTES SORTED: 29356274185
ICE165I 0 TOTAL WORK DATA SET TRACKS ALLOCATED: 687375 , TRACKS USED: 534900
ICE199I 0 MEMORY OBJECT STORAGE USED = 0M BYTES
ICE180I 0 HIPERSPACE STORAGE USED = 3144852K BYTES
ICE188I 0 DATA SPACE STORAGE USED = 0K BYTES
ICE891I 1 67099672 WMAIN, 9192 CMAIN, MAX CALLOC, N SCN, A BA, 0 AZ, 2 BZ, NN QC
ICE892I 1 271 RIN 27913 BLI 27913 BLO 276 RUN 53216 BUN 2094 CPU 00 CVC
ICE893I 1 200 XIN 14 WIN 210864 GIN NDEY PFP 5 B00 CM000 CIX UPTH LMD VS RUX
ICE894I 0 186 STR 186 MOR 1076 IPB 36 OPB 35660 CYL O MN
ICE881I 0 EQ=I2 DX=0 D2=0 D3=127 D4=1 AS=0 SA=0 SB=0 SC=0 HN=1
ICE885I 0 DAT 00 DSR 0304 BINS 0 BSZ 0 RCP 0 AJC 0 RLC 0 DUNIT 0
ICE895I 0 443 MUNIT 42 SUNIT 31 OUNIT
ICE896I 0 37247 SET 10666138 DEXTOT 558290 BLK 108326027 CSZ 69482 WE
ICE804I 1 SORTWK10 EXCP COUNT: 5336
ICE804I 1 SORTWK0F EXCP COUNT: 5349
ICE804I 1 SORTWK0E EXCP COUNT: 5350
ICE804I 1 SORTWK0D EXCP COUNT: 5331
ICE804I 1 SORTWK0C EXCP COUNT: 5334
ICE804I 1 SORTWK0B EXCP COUNT: 0
ICE804I 1 SORTWK0A EXCP COUNT: 0
ICE804I 1 SORTWK09 EXCP COUNT: 5340
ICE804I 1 SORTWK08 EXCP COUNT: 5348
ICE804I 1 SORTWK07 EXCP COUNT: 5346
ICE804I 1 SORTWK06 EXCP COUNT: 5357
ICE804I 1 SORTWK05 EXCP COUNT: 5349
ICE804I 1 SORTWK04 EXCP COUNT: 5349
ICE804I 1 SORTWK03 EXCP COUNT: 5351
ICE804I 1 SORTWK02 EXCP COUNT: 0
ICE804I 1 SORTWK01 EXCP COUNT: 5342
ICE804I 1 SORTOUT EXCP COUNT: 17805
ICE804I 1 SORTIN EXCP COUNT: 6660
ICE899I 0 HSR=2729,HSW=2876,HRE=40189,HWE=4043,HRP=522457,HWP=522457,HWM=524242,
ICE052I 0 END OF DFSORT
******************************** BOTTOM OF DATA ********************************
|
|
|
| Back to top |
|
 |
Dave Betten
New User
Joined: 24 Jan 2006 Posts: 13
|
|
|
|
| It would be better if you could send the entire sysout, not just the dfsort messages, as a text file to dfsort@us.ibm.com. Also if you can run the original sort with only the SORTDIAG change (not the sort control card and work space changes) and send that entire sysout as well? Then I can compare both and provide some feedback. |
|
| Back to top |
|
 |
ksk
Senior Member
Joined: 08 Jun 2006 Posts: 315 Location: Pune, India
|
|
|
|
Dave,
Just I want to know an information from you on Hiperspace. Is hyperspace allocated by default by DFSORT or it would depend on other parameters like CLASS parameter, number of Work files, space allocated for work files and output files etc...
I would like to know because when I didn't use any Sort Work files it used 2GB hyperspace. When added Workfiles and more space, used 3GB. Today we changed CLASS parameter also apart from workfiles and space, it used Hyperspace 4GB.
With CLASS change, Work files and more space, above mentioned 2 steps are executed in 70 Mins in production.
I will send the entire Sysout of 2 cases in a day or 2.
KSK |
|
| Back to top |
|
 |
Dave Betten
New User
Joined: 24 Jan 2006 Posts: 13
|
|
|
|
| The amount of Hiperspace used depends on the amount of storage available to DFSORT at the time it executes. So it's quite possible there was just more available on the system at the time you ran the other sorts. Adding workspace does not effect how much Hiperspace we use. |
|
| Back to top |
|
 |
Guru Bob
New User
Joined: 31 Jan 2008 Posts: 21 Location: Malaysia
|
|
|
|
My estimate is 20 SORTWK files at 2450 cylinders each. I assumed 3390 device going by the BLKSizes in the SORT output.
2GB for memory is good if you can get it all the time.
Even your output file requires 21,248 cylinders on a 3390.
Is the input on DISK or CART? |
|
| Back to top |
|
 |
|
|
|