On Fri, 7 Mar 2008 12:02:25 -0600, McKown, John
<[EMAIL PROTECTED]> wrote:
>Well, in the case of CICS, I was wrong. The manual states:
>
><quote>
>If the task is to be attached, DFHZCNA obtains a TIOA and moves the data
>from the CIB to the TIOA. DFHZATT is then called to attach the task. If
>the attach fails, the TIOA is freed. A QEDIT macro frees the CIB if the
>attach is successful, and the scan continues.
></quote>
>
>So the CIB (command input buffer?) is freed once the transaction is
>ATTACHed, not once it finishes. However, in the case of a single
>console, this is almost the same thing.
However, in the OP's case they're submitting commands via JCL, and I believe
JES will pump those commands through very fast as it reads the job. I could
easily see JES issuing a series of commands to a single CICS region much
faster than CICS could read each one and tell the system that it had done so.
You have, in essence, two processes running:
(1) Issue command
Issue another one
Issue another one
...
There are no delays between the commands.
And you have process 2,
(2) Read a command
Tell system you're ready for another one
It would be quite easy for process 1 to drive commands faster than process 2
can remove them.
Of course, this is all my guess about what the OP's job contained, and what
JES and CICS would do.
--
Walt
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
