I agree with your hypothesis ... the PAX fails because it cannot get an
Exclusive ENQ on MY.PAX
With respect to the spawned (PAX) Address Space, though, it is possible
that the spawn itself could take longer than the Batch Job switching
steps and defeat the entire purpose.
I think coding a SLEEP (or equivalent) in the TSO input could prevent this.
Regards,
David
On 2021-10-13 10:07, Paul Gilmartin wrote:
On Wed, 13 Oct 2021 09:36:51 -0400, David Spiegel wrote:
How does changing the allocation for SYSUT1 from JCL to DYNALLOC change
anything ... they're both non-Exclusive ENQs?
If a DD statement exists in a job the Initiator issues the needed ENQ at the
beginning of the job and it's in effect through all earlier steps.
The initiator never sees the DYNALLOC.
My guess is that pax runs in a different (forked) address space so its attempt
to allocate OLD conflicts with the SHR in the job's address space.
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
.
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
