[ntp:questions] Re: ntpdate steps 86400 seconds

Brian Garrett mgy1912 at cox.net
Sat Oct 1 20:21:47 UTC 2005


"Tom Smith" <smith at cag.zko.hp.com> wrote in message
news:433E8BDF.7000206 at cag.zko.hp.com...
> K.D.P wrote:
> > I ran ntpdate -d time_server_ip (just before to just after midnight)
> > additional to my original ntpdate -sB time_server_ip (every minute) in
cron.
> > Here is the result:
> >
>
> reference time:    c6e84f87.00000000  Fri, Sep 30 2005 23:57:59.000
> originate timestamp: c6e84f88.88fe6cfa  Fri, Sep 30 2005 23:58:00.535
> transmit timestamp:  c6e84f88.897689ca  Fri, Sep 30 2005 23:58:00.536
> offset -0.002518
>
> Not OK:
> reference time:    c6e9a144.00000000  Sat, Oct  1 2005 23:59:00.000
> originate timestamp: c6e9a144.ea4521cf  Sat, Oct  1 2005 23:59:00.915
> transmit timestamp:  c6e84fc4.e9487768  Fri, Sep 30 2005 23:59:00.911
> offset 86400.001896
>
> reference time:    c6e9a180.00000000  Sun, Oct  2 2005  0:00:00.000
> originate timestamp: c6e9a181.3a12cadb  Sun, Oct  2 2005  0:00:01.226
> transmit timestamp:  c6e85001.39a7f3cf  Sat, Oct  1 2005  0:00:01.225
> offset 86400.000962
>
> Both the originate and reference timestamps have skipped ahead by
> 24 hours 1 minute _before_ midnight.
>
> Note also that the reference time only has a resolution
> of 1 second.

It looks like there are either time zone or seconds-since-the-epoch
conversion isues somewhere.  I've had my desktop PC stepped 3600 seconds on
the night of a daylight saving time switch, when the server was in a
different time zone than my desktop.  Could the same thing be happening
here, with the difference in perceived time amounting to a day instead of an
hour?


Brian





More information about the questions mailing list