On Tue, Nov 8, 2016 at 7:52 AM, Christopher Brown <[email protected]> wrote:

> Hello Listers,
>
> A curious situation from Saturday night. At an acquaintances site, they
> usually do a POR (no idea why and I'm afraid to ask) for the time change.
> They also like to sit for an hour as they believe the DB2 logs will be
> corrupted (not a DB2 user but I have my doubts about that).  Anyway, the
> sysprog (who is offsite) had OPS shut the machine down and he would call
> back an hour later to IPL. But due to a medical issue he never did call
> back or IPL the machine. So through a chain of events, the acquaintance (a
> mainframer but not a sysprog) called me for help.  So its now 0630 and they
> are hours behind. I get them logged onto the HMC and walk them through an
> IPL.  During this process I am questioned about changing the time as they
> do it on the HMC (what??).  I tell them we don't have time (no pun
> intended) to figure out why or where he usually changed the time on the HMC
> and we were going to use the SET CLOCK command from the console.  So a
> miracle occurs and the IPL goes without incident, we set the clock, and
> they start to unravel their application mess.
>
>
> But they are now in a state I've never seen before.  Local time is good,
> but GMT/UTC is daylight savings time.  The only thing I can come up with is
> that they set local and GMT equal. And there is no prompt or timezone
> parameter in the CLOCKxx parmlib member.  So there're asking me what if
> they IPL again and the time hasn't been changed on the HMC??  As an STP
> user (the greatest thing since sliced bread) its been years since I've had
> to manually do anything with changing clocks, but as I recall we never did
> anything other than a SET CLOCK command.
>
> So if they IPL, what will the local time be?  Is there a reason to change
> the time on the HMC?  And the only time I've ever seen GMT set is at the
> TOD prompt at IPL, true?
>
>
​Doing an IPL does not affect the value of the TOD clock. That is what will
set the z/OS time at IPL. So, if you ended up needing to do a SET CLOCK=
(or, better SET TIMEZONE=) to get the local time correct, then if you IPL,
somebody will again​

​need to set the local time. What does a "D T" say? As an example, mine
says:
IEE136I LOCAL: TIME=08.10.51 DATE=2016.313  UTC: TIME=14.10.51 DATE=2016.313

If you have no timezone parameter in the CLOCK member, then the local time
will be set to the "UTC" time as shown in the above message.​

Your client is wallowing in the deep, dark past. DB2 logs run off of the
TOD clock (usually set to UTC) and pays _no_ attention to the local time.
The same with IMS. CICS does pay attention to the local clock, but a simple
CEMT PERFORM,RESET command will fix that right up. Or simply recycle the
CICS region (likely what your client does if they run CICS<shudder>). UNIX
processes are "who knows" but given what you've said, I'd think it likely
the client doesn't do much with UNIX.


-- 
Heisenberg may have been here.

Unicode: http://xkcd.com/1726/

Maranatha! <><
John McKown

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to