At 14:24 -0500 on 10/24/2013, Paul Gilmartin wrote about Re:
Allocation problem with LMCOPY:
On Thu, 24 Oct 2013 15:12:39 -0400, Robert A. Rosenberg wrote:
This will also fix the INIT issue that if a JOB Step has a DISP=OLD
with further steps as DISP=SHR, the dataset remains exclusively held
until the last DISP=SHR step ends (since INIT can not downgrade the
EXC ENQ to the SHR ENQ used by DISP=SHR to allow other jobs access to
the dataset).
This has been announced as a long-awaited feature of z/OS 2.1.
I don't know how it plays with DYNALLOC.
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
If the INIT can release the DISP=OLD (ie: EXC) ENQ at the end of the
last step where there a need for the DISP=OLD while allowing
subsequent steps where the dataset is declared as DISP=SHR to have a
SHR ENQ, then this TO ME implies that ENQ support has been fixed to
allow an EXC ENQ to be converted to a SHR ENQ and allow the ENQ chain
to then allow subsequent ENQs in the chain to be satisfied so until
the next EXC in the chain is encountered or the chain end is reached
(IOW: If the first ENQ on the chain after the EXC being converted to
SHR is a SHR, then it and all the following waiting SHR ENQs will be
satisfied just like what occurs if the EXC ENQ is released via a DEQ).
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
