Tom 'Needs A Hat' Mitchell wrote:
On Mon, Apr 05, 2004 at 11:30:28AM -0500, Jeff Vian wrote:I was working on the machine and noted the time had not changed on the toolbar clock in gnome. Then a little later I noted it had updated.
I do not use rdate.
However I do use ntp to sync with redhat and my time did not update until about 6pn local time. Then it suddenly did the DST adjustment.
This is sort of confusing.
The "system keeps time" UTS (date -u) and the ctime() library presents local time based on /etc/localtime unless the environment variable TZ is set for an individual. ntp works with UTS. I think it does carry daylight saving time hints.
The time of day should have toggled all at once.
How were you watching this, what time zone were you in,
is the hardware RTC set to local or UTS? Did you see any interesting messages
in your log file. My messages file is quiet but does have a gap where
time springs ahead as it should.
After you asked I went back and checked the logs. There are no entries from 01:40 to the normal logrotate @ 04:02 but also no entries between 16:13 and 18:38 yet I was using my cdrom to burn at that time and was haivng difficulty with it. the message at 16:13 and the one at 18:38 are both related to those problems. I was not working with the burner for over 2 hours, although I was doing so when the update appeared on the toolbar clock.
I am in CDT6CST time zone, and /etc/localtime is correct for that.
I did watch my radio controlled clocks and they did not see the TZMy radio controlled clock never did update automatically with the DST change. I had to do that manually. Yes it is set fot DST.
hints until dawn. I moved them to a better location and they almost
immediatly switched.
From the number of messages about problems with the updates, It is possible there was some sort of glitch that prevented all the servers (or at least some of them) from getting/sending updates, However, from what I know about NTP and the localtime file the servers not updating should not have prevented the local machine from doing the DST update. AFAIK the servers all report UCT and the local machine adjusts based on that.I guess we should watch this next year/ next fall.
There may be something I am not aware of in the intricacies of NTP that require an update from the server before the DST switch occurs, which could explain the delay if the local machine could not contact the server.
