Re: Re: quit files

    if ( sysblk.daemon_mode
#if defined(OPTION_DYNAMIC_LOAD)
         && !daemon_task
#endif /*defined(OPTION_DYNAMIC_LOAD)*/
       )
    {
#ifdef _MSVC_
        socket_deinit();
#endif
        fflush( stdout );
        exit(0);

Joe

On Sat, Jun 22, 2019 at 4:26 PM Joe Monk <[email protected]> wrote:

> If you. bothered, you'd see:
>
> " //                     PROGRAMMING NOTE
>
>     // If we're NOT in "daemon_mode" (i.e. panel_display in control),
>
>     // -OR- if a daemon_task DOES exist, then THEY are in control of
>
>     // shutdown; THEY are responsible for exiting the system whenever
>
>     // THEY feel it's proper to do so (by simply returning back to the
>
>     // caller thereby allowing 'main' to return back to the operating
>
>     // system).
>
>     // OTHEWRWISE we ARE in "daemon_mode", but a daemon_task does NOT
>
>     // exist, which means the main thread (tail end of 'impl.c') is
>
>     // stuck in a loop reading log messages and writing them to the
>
>     // logfile, so we need to do the exiting here since it obviously
>
>     // cannot.
>
> Joe
>
> On Sat, Jun 22, 2019 at 4:07 PM kerravon86-/[email protected] [hercules-390] <
> [email protected]> wrote:
>
>>
>>
>> ---In [email protected], <joemonk64@...> wrote :
>>
>> > You're wrong. Plainly.
>>
>> No. You are. Plainly.
>>
>> > When you type quit (in ANY hypervisor) with a running
>> > guest, the guest ( in this case MVS) hangs. There is no
>> > way for the log file to be flushed because the write(s)
>> > from the guest cant be completed. Hercules is flushing
>> > what it has in its buffers.
>>
>> No, the log file in question belongs to Hercules,
>> not MVS, and Hercules should be cleanly closing
>> all its disk files and reporting that verbiage in
>> the log file (as it randomly does already), and
>> ending up by reporting that the logger thread
>> is terminating (again, as it randomly does
>> already).
>>
>> BFN. Paul.
>> 
>>
>