[Top][All Lists]

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


From: Gary E. Miller
Subject: Re:
Date: Thu, 4 Jun 2020 12:33:10 -0700

Yo артур!

On Thu, 04 Jun 2020 21:49:00 +0300
артур хайруллин <> wrote:

> >> but i can’t understand. My logs have  Thu, Jun  4 2020 14:00 —
> >> 14.30 events. But ntpd eats logs and output real time 19:10.   
> > Good thing too. ntpd prevented you from setting a bad system time
> >from bad imput. How to break ntpd is better asked on an NTP list.
> >But for starters, by default ntpd wants 3 sources to agree before
> >changing the system time. Also by default nptd only makes large
> >changes to system time on startup. Both good things to prevent user
> >from shooting themselves in the foot.  
> But i make this experiment because my system had time jumps in case
> of wrong ntpd input from gps receiver 

Yup, so you said.  But without proof, is just a theory.  A good reason
to run the experiment you are running.

> I emulate this situation, i have logs with old data and when i try to
> write them ntp reacts not as i expect.

The number of people that know what ntpd will do is very small.

> Is it a good thing?

Is data good or bad?  I think it is just data.

> Why ntpd
> reads old data from SHM but processes it like nothing have happened,

Because the data was rejected.  I explained before: no cluster
confirmation, large changes prohibited after start.

> updates offset(offset is very small but difference between system
> time and server(SHM) time is large).

Without seeing your ntp logs, hard to say why.

> How is it possible?

Not only possible, but desinged that way, for good reason: to prevent what
you thought was happening.

> or i don’t
> understand the logic at all

Join the club.  NTP is very complex.  It tries hard to not be fooled and
do the right thing.

As I said before, this is now an NTP questions and you should ask on an NTP

Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703  Tel:+1 541 382 8588

            Veritas liberabit vos. -- Quid est veritas?
    "If you can't measure it, you can't improve it." - Lord Kelvin

Attachment: pgp_08r_2V_4P.pgp
Description: OpenPGP digital signature

reply via email to

[Prev in Thread] Current Thread [Next in Thread]