On Sun, May 27, 2007 at 02:41:33PM +0200, Mario Frasca wrote:
> Package: ntp
> Version: 1:4.2.4p0+dfsg-1
> Severity: important
>
> since quite some time, ntp does not syncronize the internal clock.
> as far as I can see from the logs, it computes the time reset but then
> does not manage to completely sync the time.
>
> this is an example of such a log:
> May 27 13:29:16 localhost ntpd[3221]: time reset -511.395231 s
> May 27 13:29:16 localhost ntpd[3221]: kernel time sync error 0001
This is this code:
if (ntp_adjtime(&ntv) == TIME_ERROR) {
NLOG(NLOG_SYNCEVENT | NLOG_SYSEVENT)
msyslog(LOG_NOTICE,
"kernel time sync error %04x", ntv.status);
ntp_adjtime() is a wrapper around adjtimex(). So it returned TIME_ERROR
(TIME_BAD) saying that the clock isn't synchronised.
> at 14:38 on the same day, the needed time reset is around -140s.
This looks like a rather big offset for such a short time. This seems
to be an error in the order of 30 000 ppm, which is a rather large
number. I think the kernel has a limit of 500 ppm, which is around 1.8
seconds per hour.
Did you always have this problem, or did this happen since a kernel
upgrade? Most of the times I see people complain with such an error,
it's a kernel problem.
Kurt
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
