Re: System clock running too fast

[Roel Schroeven <rschroev_nospam_ml@fastmail.fm>]

Arjen Meek wrote:

> Hello,
>
> I'm running Debian Sid on a system that seems to be having clock
> problems: the system clock is running too fast by about 5 seconds per
> hour. The system is always powered on, so from my limited understanding
> the hardware clock doesn't play a role in this.
>
> I have installed ntp-simple without result, the clock remains too fast
> even though ntpd seems to be syncing correctly. "ntpdate eskarina"
> (eskarina = my ntpd-running, properly-synced server) does seem to sync
> the clock correctly, but the time jumps resulting from running that
> regularly, even automated, don't seem like a very attractive idea to me.

In my experience, chrony does a better job of coping with clocks that 
run very fast or very slow.

Also, I don't know much about ntpd, but something in the log doesn't 
look right:

> Apr 30 03:25:13 greebo ntpd[30014]: synchronized to 193.77.147.115, stratum 3
> Apr 30 03:26:16 greebo ntpd[30014]: synchronized to LOCAL(0), stratum 13
> Apr 30 03:28:27 greebo ntpd[30014]: synchronized to 193.77.147.115, stratum 3
> Apr 30 03:32:38 greebo ntpd[30014]: synchronized to 10.0.0.1, stratum 3
> Apr 30 03:32:48 greebo ntpd[30014]: synchronized to LOCAL(0), stratum 13
> Apr 30 03:34:58 greebo ntpd[30014]: synchronized to 193.77.147.115, stratum 3
> Apr 30 03:41:24 greebo ntpd[30014]: synchronized to LOCAL(0), stratum 13

It looks like 10.0.0.1 is your local machine (since it is on stratum 
13). That would mean that ntpd syncs with 3 machines, only one of which 
is a real timeserver (and chances are that LOCAL(0) and 10.0.0.1 are the 
same machine, or at least on the same network).

--
If I have been able to see further, it was only because I stood
on the shoulders of giants.  -- Isaac Newton

Roel Schroeven