TIME loss

[Bob Johnson]

On 7/13/06, Jean-Paul Natola <jnatola at familycareintl.org> wrote:
> Hi everyone,
>
> I have been trying to figure this one out for a couple of days,  but no can
> do.
>
> My clock on my bsd box currently 19 minutes ahead of the real world.
>
> I have it set to query my w2k box as the time server.
>
> I do have the ntpd running,
>
> So I am not sure how to adjust I tried this
>
> milter# ntpdate -q 192.168.1.3
> Looking for host 192.168.1.3 and service ntp
> host found : fci2003.fci
> server 192.168.1.3, stratum 2, offset -1120.152027, delay 0.03365
> 13 Jul 14:30:19 ntpdate[79951]: step time server 192.168.1.3 offset
> -1120.152027 sec
>
> But the longer the machine stays up  (142 days ) the more time the clock
> loses,
>
> Aside from rebooting ,  is there any way to fix this?
>
>

It doesn't really make sense to run ntpdate if ntpd is already
running.  ntpdate runs once, sets the clock, and then exists.  ntpd
runs continuously and keeps the clock synchronized to the server, but
you must have the config file set up correctly. Do you have a line
like

server 192.168.1.3

in /etc/ntp.conf?

You might also want to make sure rc.conf includes:

ntpd_enable="YES"
ntpd_flags="-g -p /var/run/ntpd.pid"

The -g option lets it do a single large correction when it first
starts, similar to ntpdate.

If ntpd is running, then ntpq -p will tell you what peers it thinks it
is trying to synchronize to, and what the status is. If it shows an
asterisk (*) next to one of the peers, it is synchronized to that one.
In your case, it will be a list of one.  It takes a few minutes to
synch after first starting up.