[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Clock running slow



At 04:51 PM 12/05/02 +1100, David Cureton wrote:
>I had this problem also, never really put my finger on what was causing it, 
>however I have a feeling that is was due to the NTP server being a different 
>NTP version.

>Running ntpdc utility:
>	entering in the 'sysstats' command and observing the 'old version packets' 
>'new version packets' fields

Interesting.  ntp-simple doesn't have ntpdc, unfortunately, so I can't
check without installing ntp.


>Anyway to cut a long story short i changes the NTP server I was syncing to 
>and the problem went away. 

Darn.  I have the same ntp servers listed on my laptop and another desktop
and they are accurate.  

I'm not blocking port 123 -- all outgoing connections are allowed on this
machine.

Hum, looking at my laptop now I see nptd writing to syslog often, yet on
the other machine it's not writing to syslog.  But ntp is running.

Let's try another restart.

Both my laptop and this other "time delayed" machine are running same
version of ntpd, same exact config file.  I just restared both and see the
exact same startup sequence in syslog.

Dec  4 21:42:29 burn ntpd[1109]: ntpd exiting on signal 15
Dec  4 21:42:31 burn ntpd[12291]: ntpd 4.1.0 Mon Mar 25 23:39:47 UTC 2002 (2)
Dec  4 21:42:31 burn ntpd[12291]: signal_no_reset: signal 13 had flags 4000000
Dec  4 21:42:31 burn ntpd[12291]: precision = 7 usec
Dec  4 21:42:31 burn ntpd[12291]: kernel time discipline status 0040
Dec  4 21:42:31 burn ntpd[12291]: frequency initialized 12.294 from
/var/lib/ntp/ntp.drift
Dec  4 21:42:31 burn ntpd[12292]: signal_no_reset: signal 17 had flags 4000000

Oh! and a few minutes later:

Dec  4 21:46:00 burn ntpd[12291]: time correction of 1625 seconds exceeds
sanity limit (1000); set clock manually to the correct UTC time.

Well, that explains why ntp is not updating.  But it doesn't explain why
it's getting behind in the first place.

I'll run ntpdate again and watch more closely.

I ran /etc/init.d/ntpdate restart and that reset the system clock that was
about 25 minutes behind.  I ran hwclock --show and that showed it was
*ahead* of current time by a few minutes.  I ran hwclock --systohc to
update that and restarted ntp.

Now, as they say, time will tell.

Thanks for all the responses!



-- 
Bill Moseley
mailto:moseley@hank.org



Reply to: