> We are trying to start a GTF trace for DB2 started tasks that are 
> already running.
> They end with AHL007I... RC=8.
> Parms are TRACE=USRP,USR=(FB9),END.
> When we add JOBNAMEP and the started task names - they fail the same 
way.
> Except - we can start GTF for any jobname that's not running on a 
> particular LPAR.
> Back to believing the GTF has to start before the target of the trace.

  If you are specifying DEBUG=YES in the PARM=  on the 
//   EXEC   statement in your GTF proc, then any abend  in the
data collection routine (even a 0C4 because the data area 
specified on the GTRACE macro to write a USR record happens
to be paged out) will result in GTF taking a dump and terminating.
But the dump could get suppressed by DAE.  This should be visible
in logrec.   I don't remember offhand if the termination message in 
that case would be AHL007I vs. some other AHLxxxI message. 

  Normally, you would want to specify DEBUG=YES only when you
are trying to debug some kind of "my GTF data is missing" problem. 
 

Jim Mulder   z/OS System Test   IBM Corp.  Poughkeepsie,  NY

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to