> On Feb 22, 2018, at 4:01 PM, Jesse 1 Robinson <[email protected] 
> <mailto:[email protected]>> wrote:
> 
> The most important question is whether your LPAR(s) are running with true 
> GMT/UTC. It was once customary for shops to set both GMT and Local time to a 
> nearby wall clock and specify zero Local offset. Over time most shops have 
> hopped on to the straight and narrow path by defining GMT as UTC with a Local 
> offset that either varies (or does not) by season. If you using a true offset 
> value, you can use 'some timer facility' to adjust the offset twice a year. 
> For us that was once the external sysplex timer (9037). Now it's built-in STP 
> (Server Time Protocol) that performs the time change operation--automatically 
> if desired. An IPL is not necessary for z/OS.
> 
> Otherwise you can issue the SET CLOCK command. However, that changes only the 
> running system. Any desired change to 'parm' fields would have to be done 
> manually. Again, an IPL is not necessary for z/OS.
> 
> As others have said, most modern software like CICS and DB2 logs in UTC, so 
> changes to Local time don't matter except to Bioware. The one-hour overlap 
> when 'falling back' is not a problem except for people reading syslog or 
> operlog, where duplicate time stamps will appear. I'm quite sure that a 
> multiplex requires some kind of hardware timer because XCF cannot handle 
> differing time stamps among members.

That is a Fair statement "most modern software like CICS and DB2 logs in UTC” 
The problem we ran into is there are several old versions of CICS out there. We 
ran into on that was circa 1995 or there about. There are vendors that also 
depend on the old version. We cannot force the users to upgrade, the politics 
are astounding, we tried to push one to upgrade and you would have thought we 
were asking for several million dollars. Needless to say we lost the argument. 
When you run into politics there is little chance of winning. Our CIO was 
amazed at the resistance and refusal He only has just so much power.

Ed


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

Reply via email to