|
View previous topic :: View next topic
|
| Author |
Message |
justjpr
New User
Joined: 03 Nov 2022
Posts: 12
Location: INDIA
|
|
|
|
I have a job (cancelled) in the spool with a huge number of lines in the job log. SYSOUT has 10 million lines because it has written 20 lines of error message for each input record (the input file also has a million lines) due to some issues. I need to purge this job to clean-up and restore the space in SMS/DASD. I would like to copy the first 500 lines and the last 500 lines of SYSOUT into a dataset, before I purge this job. How do I copy it?
We have SORT/ICETOOL to extract a few records if the source is in PDS or PS. The source data is in spool here. I can do screen copy if 20–100 lines are needed. coping 1000 lines using F7/F8 is a cumbersome work.
Any help, is highly appreciated. I need to finish it by the weekend. Monday morning (28-Nov-2022) I need to purge it.
| Code: |
PREFIX=MYJOB1* DEST=(ALL) OWNER=* SYSNAME=
NP DDNAME StepName ProcStep DSID Owner C Dest Rec-Cnt Page
JESMSGLG JES2 2 MYID001 V LOCAL 10M
JESJCL JES2 3 MYID001 V LOCAL 85
JESYSMSG JES2 4 MYID001 V LOCAL 7M
SYSOUT STEP001 103 MYID001 V LOCAL 10M
A000000J STEP001 106 MYID001 V LOCAL 15
VIOSTAT1 STEP001 107 MYID001 V LOCAL 423
|
|
|
| Back to top |
|
 |
sergeyken
Senior Member
Joined: 29 Apr 2008
Posts: 2024
Location: USA
|
|
|
|
| justjpr wrote: |
| I can do screen copy if 20–100 lines are needed. coping 1000 lines using F7/F8 is a cumbersome work. |
Options to copy from SYSOUT
1. Try to view it using SE inline command from SDSF menu. If it works, copy the first 500 lines using:
- C500 as inline command at the first output line, and
- CUT F500 as operation in the Command line
Then goto BOTTOM of your output, make a note the last line number, subtract 500, and use operation L(ocate) {line} at Command line. Next enter:
- C500 as inline command at the first row on the screen,
- CUT L500 as operation in the Command line
Now you can copy your lines to any other place using operations PASTE F500, and PASTE L500 into another dataset.
2. If SE doesn't work for your 10M lines, try to replace SYSOUT=* to DSN=YOUR.TEMP.OUTPUT with DSNTYPE=HUGE (or whatever is defined by your sysadmin) |
|
| Back to top |
|
|
justjpr
New User
Joined: 03 Nov 2022
Posts: 12
Location: INDIA
|
|
|
|
I appreciate sergeyken for immediate response with appropriate answers.
"EDIF ERROR RC= 20" is the error message for the SE command.
This problem has already been escalated because it took up a lot of space, and I can't run it again with a dataset in SYSOUT.
I'll try to run the same step again with 10 input records. I hope SE / XDC will work in that case.
| Code: |
SDSF JOB DATA SET DISPLAY - JOB MYJOB001 (JOB01199) EDIF ERROR RC= 20
COMMAND INPUT ===> SCROLL ===> CSR
|
|
|
| Back to top |
|
|
Rohit Umarjikar
Global Moderator
Joined: 21 Sep 2010
Posts: 3053
Location: NYC,USA
|
|
|
|
Do you not have a Sysout at your shop which keeps the job logs archived? This is dangerous if you don’t have it.
If you have it then simply move it there and purge this one and you don’t need anything else to do.
XDC should work. |
|
| Back to top |
|
|
Joerg.Findeisen
Senior Member
Joined: 15 Aug 2015
Posts: 1255
Location: Bamberg, Germany
|
|
|
|
| sergeyken wrote: |
| 2. If SE doesn't work for your 10M lines, try to replace SYSOUT=* to DSN=YOUR.TEMP.OUTPUT with DSNTYPE=HUGE (or whatever is defined by your sysadmin) |
It should be DSNTYPE=LARGE, or EXTPREF/EXTREQ. With that you can XDC the output as proposed by Rohit. |
|
| Back to top |
|
|
Willy Jensen
Active Member
Joined: 01 Sep 2015
Posts: 712
Location: Denmark
|
|
|
|
You can do it using the REXX/API interface. The following shows 50 lines from the front (head) and 50 lines from the back (tail).
Somewhat crude I'll admit and could use some rework, but it works. You can run it in batch, if you are still able to run jobs.
| Code: |
/* REXX */
numeric digits 16
rc=isfcalls('ON')
Address SDSF
isfprefix = 'RXGV074' /* jobname */
isffiltermode = 'AND'
isffilter = 'jobid EQ JOB04698'
sdsid = 102
recn = 50 /* number of lines in head and tail */
"ISFEXEC ST"
if rc<>0 then exit isferr('Initial ST')+8
jtoken=token.1
"ISFACT ST TOKEN('"jtoken"') PARM(NP ?) (prefix @)"
if rc<>0 then exit isferr('ISFACT ST NP ?')+8
"ISFACT ST TOKEN('"jtoken"') PARM(NP SA)"
if rc<>0 then exit isferr('ISFACT ST NP SA')+8
say 'Job' jname.1 jobid.1 'has' value(isfdsname".0") 'files'
do dn=1 to isfdsname.0 /* do over dsnames */
/* isfdsname i.e. XX.RXGV074.JOB04698.D0000105.? */
parse var isfdsname.dn .'.'.'.'.'.'dsid'.'.
dsid=substr(dsid,2)+0
if dsid<>sdsid then iterate
filerecn=@reccnt.dn
say 'DDname' isfddname.dn 'has' filerecn 'records'
/* now pull data from file */
address TSO "execio" recn "diskr" isfddname.dn "(stem line.)"
if rc<>0 then exit xmsg('read 1 failed rc' rc)
rn=line.0
/* show head */
say ''
do n=1 to line.0
say '(head)' strip(line.n,'t')
end
/* show trail */
say ''
address TSO "execio" filerecn-(recn*2)-1 "diskr" isfddname.dn "(skip)"
if rc<>0 then exit xmsg('read 2 failed rc' rc)
address TSO "execio * diskr" isfddname.dn "(stem line. finis)"
if rc<>0 then exit xmsg('read 3 failed rc' rc)
do n=1 to line.0
say '(tail)' strip(line.n,'t')
end
end
rc=isfcalls( OFF )
exit xmsg()+xmsg('Done...')
XMsg: say arg(1);return 0
IsfErr:
say 'IsfErr' arg(1)
if isfmsg<>"" then Say "isfmsg is:" isfmsg
do jn=1 to isfmsg2.0
Say "isfmsg2."jn "is:" isfmsg2.1
end |
return 0 |
|
| Back to top |
|
|
|
|
