RE: SNDEMAIL - Volume testing

["Steve Richter" <srichter@xxxxxxxxxxxxx>]

-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx
[mailto:midrange-l-bounces@xxxxxxxxxxxx]On Behalf Of rob@xxxxxxxxx
Sent: Thursday, July 08, 2004 6:16 PM
To: Midrange Systems Technical Discussion
Subject: RE: SNDEMAIL - Volume testing


>Hey Steve,

>Was it you that had the PMR opened with IBM trying to deal with this? They
>keep telling me that they had another customer with this same issue that
>used some sort of a delay technique.

No, not me. Interesting that others have the same problem.

>They get real defensive about their MSF process and insist that I do not
>need to drop it down from 3 QMSF jobs and I could run 10 at a time to
>improve performance.  I am doing my best to not get defensive.

>I even
>offered to modify SNDEMAIL to copy the IFS file out of /tmp to another
>directory so that we can verify that the mime file is correct.

I built a history log into my mail code.  Every message sent is archived.
When we got the complaint that a message destined for one customer was sent
to another I checked the archive.  The address was correct.

>They are leaning towards a comm trace next.  I have a funny feeling that
>it will always fill the buffer before we notice an email ending up in the
>wrong box.

What if you write a program that sends a 100 messages, all with the same
subject line, to an email address outside of your company?  Run a 2nd job
that sends the same number of messages to another outside address, with a
different subject line.   If you run the two jobs at the same time, without
any delay between the sends, the bug should show up and the two recipients
should get messages intended for the other.

( If I could run that test myself I would!  Since the takeover ..., well
whatever. )

-Steve