[ntp:questions] Test internal clock
hundoj at comcast.net
Thu Mar 19 22:24:39 UTC 2009
On Thu, 19 Mar 2009, Rob wrote:
> Richard B. Gilbert <rgilbert88 at comcast.net> wrote:
>> Towli wrote:
>>> I would like to test my internal clock (from my domain pc) against a
>>> public ntp server, to see if there is a discrepance (i suspect our
>>> domain time is not synchronised properly).
>>> Is there a way to do this?
>>> Bst rgds from Denmark
>> You can install ntpd and/or ntpdate. Ntpdate will either set your clock
>> or tell you how far off it is, depending on the options you select.
>> Ntpdate is deprecated.
>> Ntpd will discipline your clock so that it synchronizes with your
>> selected time source(s).
>> Meinberg Funkuhren offers a pre-built ntpd for Windows together with a
>> Windows "installer". Our Meinberg contact is "Martin Burnicki"
>> <martin.burnicki at meinberg.de>
> I am a little confused about this group...
It's about the NTP protocol. Which seems to be
faithfully implemented only by the NTP reference
implementation from udel.edu.
> Is it really only about ntpd as written by Dr. Mills?
> I believed it is in the "protocols" hierarchy and is about the "ntp"
yes. Dr. Mills is the author of the NTP protocol,
accept no substitutes.
> When somebody asks a question about a Windows system, where a native
> NTP implementation is available that can be debugged with a simple
> commandline program (w32tm), the response to any question invariably
> is to "install ntpd".
That's about as good as it gets on Windows, which is a
*terrible* platform for precision time-keeping. Various
folk (Thanks, Martin and Heiko!) have made it less odious,
but you can't polish a turd.
The definition of 'NTP implementation' comes into dispute,
among many, for the Windows platform. It is not at all a
settled issue, and many would argue that MicroSoft implementation
is non-compliant. Hence another reason for the udel ntpd
> It may have some advantages, but is it really necessary for everyone?
> Wouldn't it be better to help everyone with the software they are using,
> instead to first suggest them to convert?
No. They are asking for assistance in improving their
ability to track the One True Time. No other response is
either reasonable or indicated. The folk that make do with
native Windows commands are perfectly happy with their
benefits and failings. Those that are not, need to be open
to other solutions, else their questions have no point.
> It almost seems like a religious group. Most people try to convert
> the world to Mills' NTPD, and then there is the dissident who tries
> to push Chrony in every thread.
Chrony has an impulse response that is ill-advised in
a network of NTP servers. It's really rather crude, IMHO.
Your position in a network of servers would indicate
whether Chrony is acceptable or not - there are certainly
situations where one could convincingly argue that Chrony
is a ''better'' (FSVO) solution than ntpd.
> Not really necessary, I think.
ntpd is proven to work. Others cannot make the same
claim, across the same extent of environments,
conditions and requirements. Analysis and historical
performance back the claim that ntpd is the best.
Other contenders have only hand-waving and moaning
for their justifications. Which will you trust?
> questions mailing list
> questions at lists.ntp.org
More information about the questions