On Mon, Oct 16, 2006 at 08:02:12PM -0300, Henrique de Moraes Holschuh wrote: > On Mon, 16 Oct 2006, Roberto C. Sanchez wrote: > > I have two machines which are running NTP. They both synchronize to > > these servers in /etc/ntp.conf: > > > > server ntp2.usno.navy.mil > > server ntp-1.vt.edu > > server ntp-2.vt.edu > > > > Now, here is where the weidrness comes in. The two servers' clocks are > > perfectly in sync. I have three machines (2 desktops/1 laptop) which > > synchronize to the two above named servers (not the upstream NTP > > servers). Anyhow, two of them are always within five seconds of the two > > local NTP servers. The third machine gains about 5 minutes every two > > Any NTP drift above half a second means something is completely broken, so > *none* of your client machines are working fine. The two servers seem to > work right. Make sure to also configure the two servers to *peer* each > other, btw. vague memories in my head that ntp won't sync more than a half second or so at a time, you have to use something else to get them closer and then ntp can do it. I've used ntpdate in the past as a one-time sync and then its worked after that. Also had a machine that was drifter faster than ntp could keep up with, but after a few days of hitting ntpdate randomly, it was able to calculate the drift enough to keep up after that point. this is all vague memory, ymmv widely. A
Attachment:
signature.asc
Description: Digital signature