I am not aware of a time-out in JDE, however the AS/400 will time-out
inactive sessions after
a set time. Check system value QINACTITV for number of minutes before timing
out. You can either
increase the number of minutes or disable this feature by entering *NONE.
The wait time is on the format - the file timeout doesn't do anything when
there is a screen up. The timeout only works for "multiple-requester"
display files (I think that is the right name, I'm going from a 5-year-old
memory). There was no organized drive to insert the feature in every
display file or to minimize the number of display files.
Richard, I respectfully disagree. I spent considerable time at my last JDE client undoing this 'feature', because the video doesn't timeout until the user returns from an absence and presses <enter> or any Fkey. The F continuation spec on the video of many JDE online RPG programs, w/ KNUM 1 in conjunction with some parameter on the video (sorry, can't remember, maybe the WAITRCD?) forces the programs to bail out if the user has been away from his/her desk for the specified amount of time.
As soon as I sent the previous reply, my memory triggered: the other 2 pieces in the RPG programs are the 99 error indicator on the initial READ of the video file, and the statement a few lines down, *IN99 CABEQ *ON EOJ (sets on LR). By simply removing the F continuation w/ KNUM, it will prevent the video timeouts.
I was quite pessimistic about the scenario from below. So I went
into our test environment and took a video and changed WAITRCD
to be 30 seconds. No problem. Then it occurred to me, was the
program I tested, truly using KNUM in the RPG with WAITRCD?
I found an example. Using V0006 (single bus unit revisions), and in
YOUR TEST ENVIRONMENT (duplicating the video to your test
environment), change the WAITRCD parameter to 30 seconds.
Change it to 5 seconds if you want a real impact. Run it off the menu.
Let it sit. I was surprised! I have never seen this on JDE before. We
have our system inactivity set to 45 minutes. But I think most JDE
videos are set to WAITRCD=1800 (30 min). So it does appear
possible that JDE will timeout many of the videos after 30 minutes.
On the other hand, most installations should have their system inactivity
set to somewhere between 20 & 60 minutes, purely for security sake.
So, if a video times out after 30 minutes due to inactivity, that does not
seem to be an issue. Users should not leave a video displayed on their
screen that long anyway!
<P ID="edit"><FONT SIZE=-1>Edited by ekempter on 3/12/01 03:05 PM.</FONT></P>
Sorry to have been unclear. I was disagreeing w/ this statement: <<There was no organized drive to insert the feature in every
display file.....>> The client was on A3.1, and I think we had that 'feature' in every single video. The time I spent undoing the video timeout 'feature' was to remove the KNUM 1 and *IN99 CABEQ *ON EOJ LR records from the programs. I realize I could have simply changed the WAITRCD to *NOMAX. If IBM had implemented it differently, so that the videos timed out and the programs ended after the specified time, without further user intervention, it would not have been so annoying. The fact that the video doesn't time out until the user returns to some type of input produced some very frustrating users, particularly when it happened in the middle of a large G/L batch.
A3 was before my time. A2 was the F7 code recompiled for the system/38
environment on the AS/400. I recall that A3 was the first version that been
altered to work native on the AS/400 but had no application changes from
A2 - that recollection may be incorrect but A3 wasn't much beyond the A2
When I started at JDE, A4.2 and A4.3 were the GA releases and A5.1 as in
development. I worked on A5.1 A5.2 A6.1 A6.2 A7.1 A7.3 B7A B7B and B71 in
various technical development capacities. By A5.1, I seem to recall that
many programs had the timeout removed - the timeout had been present but it
had been removed. During the A5.1 development cycle, we wrote slimers to
remove the "multiply by 1.000001" data conversion because of negative
performance impact on some of the applications and there were plans to add
significant functionality in the M&D area that was going to require date
conversions. There were a couple more major A5 slimers but I don't recall
what they were - vocabulary overrides maybe. There was another big slimer
for insuring that double byte fields were properly terminated but I don't
recall when that one was run.
I think that I would recall if we did a slimer to add the timeout and that
is the reason that I said "no organized drive". When I came on board, the
timeout didn't work in many programs - I recall handling some response line
calls about the timeout in 1990 and 1991. I assumed that it didn't work in
many programs because it was being added in slowly. Based on your
experience, it sounds like, by A3, it had been added in for everything then,
in A4, removed from programs that suffered because of it. If that was the
case then there was an organized drive but it happened before A4.