ntpd just sits there and does nothing
doug at fledge.watson.org
Mon Jul 23 04:00:28 UTC 2007
On Sat, 21 Jul 2007, Kevin Oberman wrote:
>> [LoN]Kamikaze wrote:
>>> Doug Hardie wrote:
>>>> On Jul 19, 2007, at 10:08, [LoN]Kamikaze wrote:
>>>>> As the subject says, on my 6-stable systems ntpd just sits there and does
>>>>> nothing. The logs only mention when the daemon gets started or shut
>>>>> down. It
>>>>> complains when servers are not reachable, but does nothing when they
>>>>> are available.
>>>>> The drift file always contains 0.00.
Mostly likely this means you are not communicating with the ntp servers. You
never gave us your ntpd.conf file (that I saw anyway) and what do you get with
'ntpdc -p', or the more complex command suggested earlier?
>> ntpd will not change time if the difference is too big - I think it
>> should be less then 1000s.
>> ntpdate will :)
If ntpd is working your clock will not vary from the server by more than a
second, much less 1000 secs. If ntpdate does reset the clock, it suggests that
your firewalls are not the problem and at least one of the servers will answer
your queries. You can see if ntp packets are being passed by using tcpdump.
I suppose you have made sure its running by something like 'ps -aux | grep ntp'.
> ntpdate is deprecated and is not recommended these days. The proper answer is
> to start ntpd with the -g option and to add the 'iburst' option to one or more
> of the servers in /etc/ntp.conf. The 'iburst' will speed up th initial sync to
> close to that of ntpdate, but have much greater accuracy.
> You can get the '-g' by adding 'ntpd_sync_on_start="YES"' to rc.conf.
yea but so does 'ntpdate_enable="YES"', but I still like nslookup too :)
The problem "clearly" seems to be you are not communicating with the ntp
servers. The possibilities have all been stated: bad ntp.conf, firewall (you
said there were two levels), or the servers you chose are not accepting your
queries. Without seeing the data requested we are all guessing.
More information about the freebsd-stable