[ntp:questions] Re: Monitoring the leap second tonight

mike michael.no.spam.cook at wanadoo.fr
Sun Jan 1 00:43:04 UTC 2006


Anyone miss it?
  I caught it ok.
Sun Jan 1 00:59:55 CET 2006
Sun Jan 1 00:59:56 CET 2006
Sun Jan 1 00:59:57 CET 2006
Sun Jan 1 00:59:58 CET 2006
Sun Jan 1 00:59:59 CET 2006
Sun Jan 1 00:59:59 CET 2006   <=== leap second occured
Sun Jan 1 01:00:00 CET 2006
Sun Jan 1 01:00:01 CET 2006
Sun Jan 1 01:00:02 CET 2006

  but my ntp service went a bit wonky:

on linux, using pool servers and my DCF ref.

Sun Jan  1 00:59:32 CET 2006
      remote           refid   st t when poll reach   delay   offset  jitter
==============================================================================
  127.127.1.0  LOCAL(0)  10 l   13   64  377    0.000    0.000   0.001
*127.127.28.0  .DCF.     0 l   34   64  377    0.000    0.316   0.498
+80.96.120.252 .GPS.     1 u    3   64  377  109.862   -8.316   0.709
+192.36.143.150 .PPS.    1 u   26   64  377   74.167   -6.790   0.323
+84.16.227.201   172.20.20.37    4 u   32   64  377   49.461   -2.908 
2.053

on the next sample , one server is seen with a step.

Sun Jan  1 01:00:36 CET 2006

  127.127.1.0     LOCAL(0)  10 l   12   64  377    0.000    0.000   0.001
*127.127.28.0    .DCF.      0 l   34   64  377    0.000   -0.069   0.507
+80.96.120.252   .GPS.      1 u    3   64  377  109.862   -8.316   0.698
+192.36.143.150  .PPS.      1 u   27   64  377   74.167   -6.790 377.940
+84.16.227.201   172.20.20.37     4 u   34   64  377   49.461   -2.908 
  2.056

...
   the server showing the step jitter then gets offset 1s, and it and 
the other pool are declared bust.

Sun Jan  1 01:09:08 CET 2006
      remote           refid      st t when poll reach   delay   offset 
  jitter
==============================================================================
  127.127.1.0     LOCAL(0)  10 l    8   64  377    0.000    0.000   0.001
*127.127.28.0    .DCF.      0 l   29   64  377    0.000    0.188   0.681
x80.96.120.252   .GPS.      1 u   61   64  377  111.450   -8.042   0.328
x192.36.143.150  .PPS.      1 u   26   64  377   74.366  993.388   0.456
x84.16.227.201   172.20.20.42 3 u   34   64  377   49.371   -2.903   0.487

On an NTP client of the above server, also using pool servers,

Sun Jan  1 01:07:30 CET 2006
      remote     refid      st t when poll reach   delay   offset  jitter
==============================================================================
  127.127.1.1     LOCAL(1)  12 l   43   64  377    0.000    0.000   0.001
x81.169.154.44   81.169.158.205  3 u   61   64  377   49.317  -1018.0 
377.508
+194.145.249.108 192.36.143.150   2 u   22   64  377   64.932   -3.852 
  7.070
*192.168.123.102 .DCF.      1 u    4   64  377    0.538   -2.696 999.686
Sun Jan  1 01:08:34 CET 2006
      remote     refid      st t when poll reach   delay   offset  jitter
==============================================================================
  127.127.1.1     LOCAL(1)        12 l   43   64  377    0.000    0.000 
   0.001
x81.169.154.44   81.169.158.205   3 u   59   64  377   49.317  -1018.0 
13.397
x194.145.249.108 192.36.143.150   2 u   22   64  377   64.935   -4.006 
  6.992
*192.168.123.102 .DCF.      1 u    3   64  377    0.539  -1002.9   0.972


   The linux server remained selected even though the jitter stepped out 
to 1 sec and then offset reflected that.

  30 mins later , the systems are still settling down.

this server is still showing the offset.. Did it miss the leap?
x192.36.143.150  192.36.143.164   2 u    2   64  377   73.835  992.975 
  0.270


Markus Kuhn wrote:
> If you have nothing better to do around midnight UTC tonight, then
> install
> 
>   http://www.cl.cam.ac.uk/~mgk25/time/leap/timelog.c
> 
> and run it with
> 
>   ./timelog -w2005-12-31T23:59:00 -d180 >/tmp/leapsecond.log
> 
> to record the last minute of this year and the first to of the next on
> any potentially interesting NTP machine that you can get hold of.
> 
> This records the BSD API, the POSIX API, and on Pentium CPUs also the
> Pentium timestamp counter. Command-line options allow you specify start
> time and duration of the recording.
> 
> This is meant to be portable POSIX code, but has only been tested under
> Linux so far.
> 
> Markus
> 




More information about the questions mailing list