I wonder if anyone may be able to provide some pointers on my latest
investigation...

Our developers log on to the system via a PuTTY window which provides a
tunnelled connection to SSH - they very rarely log on to TSO.

Having logged on, they issue UNIX commands that initiate OMVS task(s) which
in turn may spawn further OMVS tasks and/or submit batch jobs. Those
spawned tasks and batch jobs may themselves cause the creation of further
tasks and jobs, etc. The name of the batch job(s) is formed by appending a
single digit to user's logon id.
Having issued his initial command, our erstwhile developer is more-or-less
forced to wait until the entire chain of tasks - a 'unit of work', for want
of a better term - is complete, before he can study the results.
Alternatively, he may elect to log on again through another PuTTY window
and set off yet another such 'unit of work'.

I want to be able to summarise resource consumption along with elapsed time
for each of these units-of-work.


My latest efforts are centred around using the data in the OMVS segments of
SMF type-30 subtype 4 records, and trying to chase down the OMVS PID and
parent PID values. The recursive nature of such 'chains' of records is
proving to be a difficult nut to crack, and I'm wondering if I'm
approaching the problem from the right direction.

It may well be that such an analysis tool already exists and I'm
re-inventing the wheel.

The whole job must be done in-house with using the current software
portfolio. The idea of buying in new software will not be sanctioned by
TPTB - even using some bit of freeware will be viewed with suspicion.

If anyone can offer guidance or further information, I'd really like to
hear it.

TIA
Sean

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

Reply via email to