On Tuesday 30 December 2008, Jos van de Ven wrote:
> And you should use at least 4 upstream servers to protect against
> falsetickers.
>
> http://support.ntp.org/bin/view/Support/SelectingOffsiteNTPServers
>
> Best wishes to all,
> Jos van de Ven
>
> Op 31 dec 2008, om 00:18 heeft Brian Foddy het volgende geschreven:
> > After only 20 minutes, its many seconds off already.
> > ntpq -pn
> >     remote           refid      st t when poll reach   delay
> > offset  jitter
> > =
> > =
> > =
> > =
> > =
> > =
> > =
> > =
> > ======================================================================
> > *127.127.1.0     .LOCL.          10 l   32   64  377    0.000
> > 0.000   0.977
> > 128.105.37.11   128.105.201.11   2 u   26  256  377   64.117
> > 10503.0 3750.77
> > 173.16.32.174   99.150.184.201   2 u   18  256  377   58.125
> > 10599.0 3771.55
> > 209.51.161.238  .CDMA.           1 u   16  256  377   68.408
> > 14408.1 6707.28


Well I'm totally dumbfounded, I removed the eth2 interface, restarted ntpd and 
no help.  Finally I simply rebooted the machine and it synced up in < 2 
minutes.  My best guess, when I added the interface there probably was a 
reboot in there and somehow the clock got programmed with some incorrect 
value from the Linux kernel on startup that caused the clock to drift so far 
off so fast the ntp couldn't pull it back.  I was seeing a 100 second drift 
in only a few minutes.  Now even before the sync its only off by a almost 
nothing.
At least my graph is moving up again...
http://www.pool.ntp.org/scores/216.160.0.218

Thanks for the help, I'm sure some of the config recommendations should have 
been done anyway.  If I find another possible cause as I re-do some of what 
I've undone I'll post a followup.

Brian

_______________________________________________
timekeepers mailing list
[email protected]
https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers

Reply via email to