I just finished transitioning most of my servers to XNTPD. I'd been having
trouble with timesync enabled machines falling in/out of time, forcing me
to restart timesync.nlm to end the in/out loop. I'd expected XNTPD to
resolve that...the day I tranistion to XNTPD one of my machines falls out
of synch.

I'm trying to figure out what might contribute to this kind of problem.
I'd suspect our switch but our network manager assures me it's not. The
switch has been up for fourteen weeks without a single error logged on any

I'm peering three servers with my router. The remaining servers consume
time from one preferred server (with secondary servers listed in NTP.CONF).

Chris Premo put together a nice CoolSolution describing a FIXNTP.NCF script.
"unload xntpd
load ntpdate XXX.XXX.XXX.XXX

Maybe my only means of keeping things in synch is to cron this script and
have each server run the script at non-overlapping intervals?

What can I look at to determine the source of lost sychronization? Or is
edir just THAT picky? Regardless, I'd assume XNTPD could adjust and
compensate for slight variations...