[PLUG] automatic running of ntpdate

[Denis Heidtmann]

On Wed, Dec 2, 2009 at 8:45 AM, Denis Heidtmann
<denis.heidtmann at gmail.com> wrote:
> On Tue, Dec 1, 2009 at 9:41 PM, Mike Connors <mconnors1 at gmail.com> wrote:
>> Denis Heidtmann wrote:
>>> Related to my long dialog about ntpdate, I found two types of entries in syslog:
>>> Those caused by a script related to if_up:
>>> Dec  1 10:30:05 R2D4 ntpdate[3316]: step time server 74.207.251.121
>>> offset 0.407052 sec
>>>
>>> And those of an unknown cause:
>>> Dec  1 12:00:21 R2D4 ntpdate[6823]: step-systime: Operation not permitted
>>>
>>> The first type is the default invocation in ubuntu. It invariably
>>> reports an offset of in the order of .5 seconds.  Yet when I run the
>>> command from a terminal the offset is near zero.  This seems strange.
>>> One possibility is that ubuntu does not set the hardware clock to the
>>> system clock on exit.  How can I explore this?
>>>
>>> About the second type,  I am trying to find what precipitates it.
>>> What does it mean?  I have searched for files in /etc containing the
>>> text ntpdate, but have come up with nothing which seems to execute
>>> ntpdate except /etc/network/if-up.d/ntpdate.  I thought that the noon
>>> time was significant, but an identical entry appeared at around 1829.
>>>
>>> Ideas on how to crack this nut?
>
>  >From my research, the 2nd entry can be caused by 1 of 2 things.
> Either another ntp daemon is running when ntpdate attempts to step the
>> system time. See below:
>>
>> ntpdate will decline to set the date if an NTP server daemon (e.g., ntpd) is running on the same host.
>>
>> Or if you're using ntpdate on a virtual server.
>
> ntpd is not installed. I searched in synaptic and found nothing
> installed save ntpdate and ntpdate-debian containing the letters ntp.
> Besides, doesn't the 2nd type log entry mean that the executable
> ntpdate was launched? By what? A virtual server?  I have to confess I
> have no idea what this means.  I do have VirtualBox, but not running.
> I assume you mean something more esoteric that VB.
>
> BTW, the time offset in the log this AM is near 1 second.  So my idea
> that the kernel is not setting the hardware clock on shutdown is
> looking more plausible.
>
> -Denis
>
This AM the log showed an offset of 0.366 seconds, so it looks like
the hardware clock was reset.  Is it possible that the hardware clock
is reset (on shutdown) to the nearest second?  That would account for
the offsets bouncing around, but staying under 1 second.

-Denis