On 28/08/2019 11:52 am, Paul Gilmartin wrote:
I wonder whether nowadays more address spaces are created for batch
job steps or for fork()? Of course, it's environment-sensitive. Which
path should be optimized? Whenever I see the message
BPXAS ON INTRDR
it appears the batch path is optimized. I'm astonished and dismayed to
think that fork() is realized (sometimes) by sending imaginary cards
through an imaginary card reader.
Creating an address space on z/OS is relatively expensive, BPXAS is a
reusable address space used for unix work. BPXAS ON INTRDR just means
that the system has run out of idle unix address spaces and is creating
more, triggered by some form of unix work. Fork may be a trigger, but
it's more a resource management function than part of fork.
In my test, you can see all the forks were reusing the same BPXAS
STC06734. Parent tasks used STC06731 and STC06733.
If the BPXAS is unused for a period of time (30 minutes?) it is shutdown
and a new one will need to be created next time.
Unix is sending everything through an imaginary teletypewriter, is there
a difference? :-)
--
Andrew Rowley
Black Hill Software
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
