[ntp:questions] NTP not syncing

David Lord snews at lordynet.org
Sun Nov 3 02:33:55 UTC 2013


Antonio Marcheselli wrote:
>>
>> My guess is you're not meant to do it that way but then I'm
>> no expert at configuring debian.
> 
> As pointed out by John, it's how the manufacturer implemented the system.
> 
>>
>> Are you completely unable to resolve ip address to hostname
>>
>> eg.
>>
>>      $ host 130.88.200.4
>>      4.200.88.130.in-addr.arpa domain name pointer dir.mcc.ac.uk.
> 
> Yes, I get a timeout.
> 
> Sorry but I can't really amend network settings. The network on this 
> machine is being used for critical secure packages and anything wrong 
> would have bitter consequences.
> 
> I appreciate that using pools would be better but in my case I'm happy 
> with just the IP numbers. If/when the servers go offline I'll change the 
> configuration.
> 
> But thanks for your input anyway
> 
>> You also probably need to stop ntpd if it's not synced with
>> a low offset of a few ms, delete the driftfile, run ntpdate
>> then restart ntpd.
> 
> Which I did an hour-ish ago.
> 
> Stopped ntpd, deleted drift file, amended configuration as you suggested 
> as below:
> 
>     server 158.43.128.66 iburst
>     server 81.168.77.149 iburst
>     server 130.88.200.4 iburst
>     restrict 127.0.0.1
>     restrict 169.254.1.1 mask 255.255.0.0 nomodify
>     tos orphan 6
>     driftfile /status/etc/ntp/ntp.drift
>     logfile /var/log/ntp.log
>     multicastclient
>     broadcastdelay 0.008
>     enable monitor
> 
> Now after just an hour the drift is -30, and all offsets are near zero.
> 
> Same server, same network, what happened??

Ntpd is intended to run continuously, here I've seen uptimes
of +200 days. Older releases of ntpd don't behave at all well
if repeatedly restarted so above method is same as I use when
restarting after an outage, especially if the battery backed
RTC is not keeping good time during the outage.


David


> 
> Antonio



More information about the questions mailing list