[ntp:questions] Ntp not syncing after powerfail of server

M. Giertzsch mgiertzsch at mobotix.com
Thu May 19 15:52:02 UTC 2011

Am 19.05.2011 16:01, schrieb Heiko Gerstung:
> Hi,
> Am 18.05.2011 11:57, schrieb M. Giertzsch:
>> Hello together,
>> I want to synchronise a local network and there is no uplink to the outside world.
>> So I have one dedicated ntp server and all other devices are client of this server.
>> Now if there is a powerfail, there is no guarantee that the server has finished the boot
>> before the clients have. In this case the time will not be synced.
>> Is anybody aware of options taking account to this? Does anyone know threads
>> or howtos for further reading? Am I doing a mistake or do I have my clients missconfigured?
> Maybe I am missing something, but if you use ntpd on the client machines, they will poll the server in intervals. So even if
> they do not get a response the first time they ask, they should receive the time on the second try (and all consecutive calls).
> We have no chance to tell you whether you have misconfigured your clients, because we do not know
> - what OS your clients are running on
> - what kind of NTP software you are using (ntpd version X or something else?) on the clients and on the server
> - how your current configuration looks like (ntp.conf?)
> I could imagine that your clients probably use ntpdate or a similar software to query the time once during startup, but that
> is just me guessing your setup. If that would be the case, all you could probably do is inserting a sleep before the ntpdate
> call or some other ugly workaround like this.
> Regards,
>   Heiko
> _______________________________________________
> questions mailing list
> questions at lists.ntp.org
> http://lists.ntp.org/listinfo/questions

Hello Heiko,

thanks for your help. You are totally right, I have forgotton to give 
you all these informations required.
The NTP software in use is ntpd 4.2.0a on both client and server. Well, 
I can't tell you exactly the OS in use,
but it's a linux like OS running on an embedded arm device.

The NTP configuration looks like follows:

restrict default notrap nomodify nopeer noquery
driftfile /var/status/ntp.drift
pidfile /var/run/ntpd.pid
server 0.de.pool.ntp.org iburst
server 1.de.pool.ntp.org iburst
server 2.de.pool.ntp.org iburst
fudge stratum 13

May it be that iburst is the problem? Client gives up because of too 
many non-responses, or will iburst
ignore these and send another bunch a little later?

If you need any more information please ask me.
Thank you in anticipation.

Best regards,
    M. Giertzsch

More information about the questions mailing list