Am Mittwoch, 28. November 2001 04:05 schrieb David Douthitt: > On 11/27/01 at 12:45 PM, Matt Schalit <[EMAIL PROTECTED]> wrote: > > Jeff Newmiller wrote: > > [snip] > > > > > > xntpd steps the clock on my unix box every so often, > > > > but it slews it every day. Not sure what your seeing. > > > > > > David is referring to a startup phenomenon. Read the > > > manpage for ntpdate. > > > > I read that one, and it refers to how ntpdate steps or slews. > > I don't see any mention that xntpd behaves like ntpdate. > > It goes like this: > > You boot your machine. The startup script for xntpd runs ntpdate to > set the system clock (and the kernel clock). It does this in such a > way so that the time is set right away, even if it is hours or days > off. > > Then xntpd runs, and it only slews the clock forwards or backwards to > match the current time server. It may go faster or slower, but it > should never (as far as I know) "jump" in time - this is so that > running processes don't get confused or fail. > > After ntpdate runs during boot, its job is (normally) completed, and > as long as xntpd runs, it's not needed again.
xntpd already knows the switch -g, which prevents xntpd from exiting if local time differs more than 1000 sec from the timeservers time. It also steps and jump in times. It's an incomplete solution since it tooks a long time, your local time gets corrected. According to the FAQ ntp 4 will mimic the ntpdate behaviour during initial synchronisation - see section 6.1.3.2 at http://www.eecis.udel.edu/~ntp/ntpfaq/NTP-s-config.htm Might be worth to test. kp _______________________________________________ Leaf-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-devel
