When using SDSF on the mainframe to view the output of my batch jobs, I've always used the action character s
to concatenate all the various members into one, since it's easier to see the whole job output together:
SDSF STATUS DISPLAY ALL CLASSES
COMMAND INPUT ===>
PREFIX=* DEST=(ALL) OWNER=PAX
NP JOBNAME JobID Owner Prt
PAX TSU39756 PAX 1
s_ PAXJOB01 JOB39757 PAX
PAXJOB02 JOB39759 PAX
I've done this manually but also with scripts running on Linux to screen-scrape job output for automated status updates.
However, just today, I started seeing binary information in that output and went investigating (using ?
instead of s
, to get a list of all related members). There are some members that I've never seen before:
SDSF JOB DATA SET DISPLAY - JOB PAXJOB01 (JOB39757)
COMMAND INPUT ===>
PREFIX=* DEST=(ALL) OWNER=PAX SYSNAME=*
NP DDNAME StepName ProcStep DSID Owner C Dest
JESJCLIN 1 PAX A
JESMSGLG JES2 2 PAX A LOCAL
JESJCL JES2 3 PAX A LOCAL
JESYSMSG JES2 4 PAX A LOCAL
$INTTEXT JES2 5 PAX A
$JOURNAL 6 PAX A
STDIN RUNLOG 101 PAX A
STDOUT RUNLOG 102 PAX A LOCAL
Specifically, JESJCLIN
, $INTTEXT
, $JOURNAL
and STDIN
are the members that are new, and the journal is the one containing the binary information.
I want to know what controls this behaviour since it's screwing up my automated processes. I'd rather not have to ?
in to the above screen and extract only the outputs I'm interested in since that will change per job and require some heavy configuration - it's far easier to get the concatenated job output, provided I can get back to the original behaviour.