Paul Gilmartin wrote:
On Wed, 13 Jun 2007 13:48:19 -0400, Kirk Talman wrote:
More accurately, single ID, single LPAR, single application, multiple
session facility.
I stand corrected. Thanks.
That is an application issue. E.g., we have two TORs using a single
AOR/FOR pair. For some transactions we disallowed using both AORs at the
same time in the transaction's security signon part because the
transaction was not sophisticated enough to build temp storage queue names
to distinguish simultaneous signins (and related issues). And it was
deemed not worth the effort to correct that.
Theology. Dogma. Simply to start doing it right, you must
stop doing it wrong. Somehow I feel a major culprit is VTAM
because whenever this issue arises an expert starts spouting
VTAM jargon. Get rid of VTAM; let TCP/IP connect directly to
the TMP input/output data streams.
Has nothing to do with VTAM. We allow users to logon as many times as
they want to our TOR and execute the same transactions no matter what.
There is nothing in VTAM that prevents this.
IIRC even using TCP/IP CICS builds a LU name under the covers. The LU
name happens to be the hex value of the IP address. So if your IP
address is 10.10.10.10, your LU name is 0A0A0A0A.
My guess is that the program was building TS queues based on the full,
or part of the userid. One userid, one TS queue.
Whereas our programs build TS queue names based on the TERMID (the 4
character CICS TERMID). The same user logging on from 4 different
terminals can easily have 4 TS queues.
----------------------------------------------------------------------
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
