|
View previous topic :: View next topic
|
| Author |
Message |
balakrishna reddy.bala
New User
Joined: 15 Sep 2010
Posts: 17
Location: india
|
|
|
|
Hi All,
I have a requirement where in which i need to capture the joblog in sdsf
I am trying to do it using rexx
The following are the scenarios which i need to takeup
Scenarios:
1) If the same job name was submitted by two different users, I have to take the backup of joblog which i have submitted
2) If there are two or three jobs in sdsf, I need to take backup of the latest job ran
Please find below the code which i used to get these details
| Code: |
ADDRESS TSO
"ALLOC F(ISFIN) TRACKS SPACE(1) REU" /* USED BY SDSF */
"ALLOC F(ISFOUT) NEW DELETE REU " , /* USED BY SDSF */
"CYLINDER SPACE(2,1) LRECL(133) RECFM(F,B,A) DSORG(PS)"
"ALLOC F(TEMPPRT) DA('"PDSNAME"("JOBNAME")')
MOD SPACE (80 80) CYL DIR(10)"
JOBID = JOBNAME
STR = SUBSTR(JOBNAME,1,7)
QUEUE "ST"
QUEUE "PRE *" /* SDSF COMMANDS IN BATCH*/
QUEUE "OWNER " || VUSRID
QUEUE "FIND " JOBID
QUEUE "SORT JOBID D"
QUEUE "++S" /* BROWSE MSGUSR DATASET */
QUEUE "PRINT FILE TEMPPRT " /* PRINT TO TEMP DATASET */
QUEUE "PRINT 1 999999999"
QUEUE "PRINT CLOSE"
QUEUE "END"
QUEUE "EXIT"
"EXECIO" QUEUED()" DISKW ISFIN (FINIS" /* INPUT TO SDSF BATCH */
ADDRESS ISPEXEC "SELECT PGM(ISFAFD) PARM('++25,80')"/* INVOKE SDSF */
"FREE F(TEMPPRT)"
RETURN
|
When i run these commands using rexx , The logon procedure joblog ( TSO user id job) is getting backedup
Can anyone help me out in resolving it.
Thanks in advance
Balakrishna |
|
| Back to top |
|
 |
Garry Carroll
Senior Member
Joined: 08 May 2006
Posts: 1205
Location: Dublin, Ireland
|
|
|
|
| Code: |
QUEUE "++S" /* BROWSE MSGUSR DATASET */
|
The comment is not consistent with the command (unless your system puts the MSGUSR output at the start of the job). You should ++? to expand the job and then FIND the MSGUSR output.
Garry. |
|
| Back to top |
|
|
balakrishna reddy.bala
New User
Joined: 15 Sep 2010
Posts: 17
Location: india
|
|
|
|
Hi Garry,
Thanks for your responce
Please ignore the comments, I want to take the backup of whole job.
If we give '++?' it will expand the job as you said, But if we give directly '++S' it will open all the messages then i will print them. |
|
| Back to top |
|
|
nevilh
Active User
Joined: 01 Sep 2006
Posts: 262
|
|
|
|
| The method you are using is not reliable. Jes has an installation defined number of job numbers available when these are used up jes starts at the lowest value again. If this happens your rexx will never select the correct job. To cater for this you will have to build extra logic into your rexx. Secondly if the job you are trying to backup has the same owner id as a tso user , the tso user will always be backed up as you do a "sort jobid d" and Txxxx is greater than Jxxxx |
|
| Back to top |
|
|
balakrishna reddy.bala
New User
Joined: 15 Sep 2010
Posts: 17
Location: india
|
|
|
|
Hi,
Is there anyway to get the latest job first and filter it based on user id
As you said 'Txxxxxxx' is greater than 'Jxxxxxxx', This is what exactly happening, Can you suggest me something to get the right job's joblog |
|
| Back to top |
|
|
nevilh
Active User
Joined: 01 Sep 2006
Posts: 262
|
|
|
|
Sorry I can't be much help a lot will depend on how your system is set up.
For example if you logoff does the output from your tso session stay in the spool or is it automatically deleted, if it stays in the spool to which queue does it go.These are questions only you can answer. You may however have more success if instead of using the ST command you go directly to the Output or Held queues as at least there you will only get a list of completed jobs and no active tasks will be shown. Good Luck |
|
| Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007
Posts: 10886
Location: italy
|
|
|
|
You will have to resort to a different approach
using the full programmable REXX/SDSF interface
as described in the manuals for Your level of zOS starting here
www-03.ibm.com/systems/z/os/zos/bkserv/index.html
and the redbook here
www.redbooks.ibm.com/abstracts/sg247419.html?Open
something along the lines of
| Code: |
****** ***************************** Top of Data ******************************
000001 /*REXX - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - */
000002 /* */
000003 /* - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - */
000004 Trace "O"
000005 Parse Source _sys _how _cmd .
000006 parse arg args
000007
000008 IsfRC = isfcalls("ON")
000009 if IsfRC ¬= 0 then do
000010 say "isfcalls RC" IsfRC
000011 exit
000012 end
000013
000014 isfprefix = "ENRICO*"
000015 isffilter = "queue = print"
000016 isfcols = "jname jobid "
000017
000018 Address SDSF "isfexec st"
000019 if RC ¬= 0 then do
000020 say "isfexec RC" RC
000021 exit
000022 end
000023
000024 do ij = 1 to jname.0
000025 say right(ij,2) jname.ij jobid.ij
000026 end
000027 call isfcalls "OFF"
000028 exit
****** **************************** Bottom of Data ****************************
|
applying some sliding window to the jobid You will be able to determine the right job to process
the next snippet will build on the previous one and
display the all the ddnames for all the jobs
| Code: |
****** ***************************** Top of Data ******************************
000001 /*REXX - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - */
000002 /* */
000003 /* - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - */
000004 Trace "O"
000005 Parse Source _sys _how _cmd .
000006 parse arg args
000007
000008 IsfRC = isfcalls("ON")
000009 if IsfRC ¬= 0 then do
000010 say "isfcalls RC" IsfRC
000011 exit
000012 end
000013
000014 isfprefix = "ENRICO*"
000015 isffilter = "queue = print"
000016 isfcols = "jname jobid "
000017
000018 Address SDSF "isfexec st"
000019 if RC ¬= 0 then do
000020 say "isfexec RC" RC
000021 exit
000022 end
000023
000024 do jnum = 1 to jname.0
000025 say right(jnum,2) jname.jnum jobid.jnum
000026 jopts = "(PREFIX J)"
000027 Address SDSF "ISFACT ST TOKEN('"token.jnum"') PARM(NP ?) " jopts
000028 if RC ¬= 0 then do
000029 say "isfact RC" RC
000030 exit
000031 end
000032 do idd = 1 to jddname.0
000033 say right(jnum,2) jname.1 jobid.1 right(idd,2) jddname.idd
000034 end
000035
000036 end
000037 call isfcalls "OFF"
000038 exit
****** **************************** Bottom of Data ****************************
|
as a further expansion the next snippet will do the above
plus it will read and display the first three datasets for each job
| Code: |
****** ***************************** Top of Data ******************************
000001 /*REXX - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - */
000002 /* */
000003 /* - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - */
000004 Trace "O"
000005 Parse Source _sys _how _cmd .
000006 parse arg args
000007
000008 IsfRC = isfcalls("ON")
000009 if IsfRC ¬= 0 then do
000010 say "isfcalls RC" IsfRC
000011 exit
000012 end
000013
000014 isfprefix = "ENRICO*"
000015 isffilter = "queue = print"
000016 isfcols = "jname jobid "
000017
000018 Address SDSF "isfexec st"
000019 if RC ¬= 0 then do
000020 say "isfexec RC" RC
000021 exit
000022 end
000023
000024 do jnum = 1 to jname.0
000025 say right(jnum,2) jname.jnum jobid.jnum
000026 jopts = "(PREFIX J)"
000027 Address SDSF "ISFACT ST TOKEN('"token.jnum"') PARM(NP ?) " jopts
000028 if RC ¬= 0 then do
000029 say "isfact RC" RC
000030 exit
000031 end
==CHG> do jdd = 1 to jddname.0
==CHG> say right(jnum,2) jname.1 jobid.1 right(jdd,2) jddname.jdd
000034 end
000035
==CHG> do jdd = 1 to 3
000037 Address SDSF "ISFACT ST TOKEN('"jtoken.jdd"') PARM(NP SA) "
000038 "EXECIO * DISKR " isfddname.1 " ( OPEN FINIS STEM SPOOL. "
000039 do i = 1 to spool.0
000040 say spool.i
000041 end
000042 end
000043 end
000044
000045 call isfcalls "OFF"
000046
000047 exit
****** **************************** Bottom of Data ****************************
|
why use obsolete methods when better ones are available ? |
|
| Back to top |
|
|
Bill Woodger
Moderator Emeritus
Joined: 09 Mar 2011
Posts: 7309
Location: Inside the Matrix
|
|
|
|
| Have you got any date/time that you can use? Much better for getting the "latest" of something. What if I submit two jobs, the first one held, the second one not, same name. One runs. A week later I release the other, which has the lower Jobid. Then I run another, but the Jobids have "cycled" and this has the lowest Jobid of all three. Etc. |
|
| Back to top |
|
|
balakrishna reddy.bala
New User
Joined: 15 Sep 2010
Posts: 17
Location: india
|
|
|
|
Hi,
I have to filter as per the date of run which ever is recently submitted and then if submitted on the same date then sort based on jobid and take backup of the least one of it in a pds. |
|
| Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007
Posts: 10886
Location: italy
|
|
|
|
| Quote: |
| I have to filter as per the date of run which ever is recently submitted and then if submitted on the same date then sort based on jobid and take backup of the least one of it in a pds. |
clear as mud 
if You want good answers You will have to ask clear questions
any reason to sort on jobid ?
what difference does it make if You just store something somewhere?
it is not that You browse it! |
|
| Back to top |
|
|
dbzTHEdinosauer
Global Moderator
Joined: 20 Oct 2006
Posts: 6966
Location: porcelain throne
|
|
|
|
our jobids roll about every two days.
that means one day the second job to run could be of a higher jobid
or it could be lower.
basing anything on jobid without some type of date/time qualification,
is asking for a problem. |
|
| Back to top |
|
|
Bill Woodger
Moderator Emeritus
Joined: 09 Mar 2011
Posts: 7309
Location: Inside the Matrix
|
|
|
|
With date and time available, all your problems go away, as far as the data is concerned anyway.
Without date and time you've got nothing but tortuous idiotic code to look forward to, and then it won't reliably give you the correct answer.
It's taken you two weeks to move to seeing that you need the date, pick it up a bit now and get the time.
Then you can get your data, simply, in the exact order that it was executed, irrespective of anything else (holds, jobid cycling, whatever).
At the same time, try to get everything else you need in one shot. Then you have file with all your data, whicih you can usefully sequence/select to your heart's content. |
|
| Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007
Posts: 10886
Location: italy
|
|
| Back to top |
|
|
|
|
