X-no-Archive:yes
We have been having great fun trying to think through the implications of
this. We have concluded that so long as nothing else time sensitive is
running between 2AM (for the POSIX clock) and issuing SETCLOCK TIMEZONE (or
vice versa), we do not expect to see any impact. Conversation with NOBIX has
confirmed that anything scheduled to run at or after 1AM (but before issuing
SETCLOCK TIMEZONE) would run twice. We have chosen to live as dangerously as
possible on one system, and have the NOBIX JobPak scheduler stream the job
that does our SETCLOCK, at 2AM, which NOBIX recommends against. We have
nothing else scheduled for Sunday AM, so the worst thing that should be
possible is that this SETCLOCK job is scheduled and ran twice. Based on
Patrick's testing, we have added date and showclock before and after the
setclock, for proof of concept. If there is anything interesting to report,
I'll repost Monday; if anything seriously bad happens, I will attach a plain
text copy of my current resume ;-). Until then, trick or treat.

Greg Stigers
http://www.cgiusa.com

-----Original Message-----
From: Patrick Santucci [mailto:[log in to unmask]]
Sent: Friday, October 29, 1999 2:23 PM
To: [log in to unmask]
Subject: Re: when worlds collide

We did some testing on our Y2K box after we got the same info from Nobix.
We found that the Posix time (using the $date command) changed instantly
at 2 am back to 1 am. We then did a SETCLOCK TIMEZONE=W6:00 (we're on
Central time so our fall-back is from W5 to W6) the Posix time changed
again to match the system clock!
<snip>