On Wednesday, 03/11/2009 at 06:14 EDT, Kris Buelens 
<[email protected]> wrote:
> I you have VMUTIL issue the SET TIMEZONE command, it might be OK, 
because, when 
> the SET TIMEZONE has ended, VMUTIL will see the new time when it returns 
to 
> WAKRUP.  Other servers using WAKEUP will not notice the change, until 
they get 
> waken up by some other event.  As our servers didn't have real time 
critical 
> things on Sundays, I didn't mind.

To clarify, a guest only receives notification of a SET TIMEZONE if it 
uses diagnose 0x274 to register to receive it.  CMS doesn't do that and it 
doesn't have an interface by which can notify all Interested Parties.

You need guest whose sole purpose in life is to await said signal.  Upon 
receiving it, would signal your automation software (e.g. VMUTIL) to Do 
Something (E.g. generates MSG VMUTIL TIMEZONE-CHANGED.  The VMUTIL 
server's most likely response is to restart CMS and GCS servers.

The PIPE delay stage is documented to use the clock comparator to wake up. 
 That is, it sets the comparator with a TOD clock value in the future. 
When the current TOD matches the comparator, an interrupt is generated. 
Changing the timezone does not change the TOD, so there is no automatic 
wakeup.  Just interrupting a Pipe delay won't help since I think Pipe 
retrieves the tz (via diag 0) when the PIPE command is issued, not when 
the delay stage i started.

Alan Altmark
z/VM Development
IBM Endicott

Reply via email to