On 28/10/2020 12:41 am, Paul Gilmartin wrote:
But since OMVS manages address spaces by highwater mark, calling BPXAS
only when the level rises and reaping after 30 seconds (is this configurable?)
I'd expect orders of magnitude fewer SMF 30 than you observed.
The SMF records are written for the work running in the address space,
not the address space itself (the same as batch jobs).
Fork-exec writes SMF records for 2 pieces of work - the initial fork,
then a new sub-step created by exec. You can see that in the report -
bash creates another bash process, which is then replaced by the command
in a new sub-step.
All these are reusing the same address space STC06734.
/bin/sh is executing the commands in a separate address space but
without the sub-step created by fork-exec (I guess this is the spawn
method that Kirk Wolf described).
As I said in the write up, this isn't a criticism of bash. It is very
useful to have bash available, and it makes sense to minimize the
changes for the z/OS port to make it easier to keep up with changes in
the mainstream. Whereas /bin/sh has probably been optimized for z/OS.
--
Andrew Rowley
Black Hill Software
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
