Re: quit files

---In [email protected], <winkelmann@...> wrote :

>> So I run startmvs2 and wait for the scripts to
>> complete, and then type "quit". The first time
>> I did this with the untouched TK4- it actually
>> worked, ie all the verbiage was displayed on
>> the screen and the log, like this:

> As I said: If used as designed, there is no race
> condition. However, I don't consider entering
> quit after IPL completion being "use as designed".

I think Hercules should be deterministic and
close its files, including the log file, cleanly
on response to "quit", as it *randomly* does
correctly already.

> How about shutting down MVS first, using the
> TK4- automated shutdown process?

I tried "/s shutdown" 3 times, and all 3 times it
produced a clean log file.

> This thing is there for a reason... 

And there's a reason why I use "quit" instead
of "/s shutdown". I want sub-second termination,
not a one-minute termination.

It has been working fine on a modified tk3upd
for many years. I suspect the extra devices
added to TK4 are causing the issue. What I
first noticed was data that I had populated
was not sticking, until I added a pause
before the quit. It's not a clean log file I want
so much as the DASD to be flushed before
program exit.

BTW, there are delays on IPLing TK4-, like
this:

05:03:43 HHC01603I hao cmd script scripts/SCR00010A
05:03:43 HHC00077I The command was placed at index 3
05:03:47 HHC00008I /d t
05:03:47 HHC02264I Script 1: file scripts/ipl.rc processing ended
05:03:47  IEE136I  TIME=20.03.47 DATE=19.172
05:04:10  +BSPPILOT - Running script STARTSTD
05:04:10  +BSPRS22I - Dataset processed: SYS1.PARMLIB

You can see a 23 second delay there.

I changed IEAIPS00 in my modified tk3upd to avoid
a similar problem, as outlined here:

https://groups.yahoo.com/neo/groups/hercules-os380/conversations/messages/11700

BFN. Paul.