ntpd and ntpdate

[Bradley Marshall <bmarshal@PIsoftware.com>]

Hi all,

After much screwing around with ntpd, I think starting to
get a handle on what its doing.  When you first start ntpd,
the associations (mappings between hosts listed as ntp servers
and their status) from ntpq looks something like:

$ sudo ntpq
ntpq> ass
ind assID status  conf reach auth condition  last_event cnt
===========================================================
  1  9316  8000   yes    no
  2  9317  8000   yes    no
  3  9318  8000   yes    no
  4  9319  8000   yes    no
  5  9320  8000   yes    no
  6  9321  8000   yes    no

After a little more time, it will look like:

ind assID status  conf reach auth condition  last_event cnt
===========================================================
  1 11940  9014   yes   yes  none    reject   reachable  1
  2 11941  9014   yes   yes  none    reject   reachable  1
  3 11942  9014   yes   yes  none    reject   reachable  1
  4 11943  9014   yes   yes  none    reject   reachable  1
  5 11944  9014   yes   yes  none    reject   reachable  1
  6 11945  9014   yes   yes  none    reject   reachable  1

In either case, it has not synchronized with anything yet, as shown
below.

  $ ntptrace 
  localhost: stratum 16, offset 0.000012, synch distance 0.00003
  0.0.0.0:        *Not Synchronized*

On morris, after a bit more time it looks something like:

ind assID status  conf reach auth condition  last_event cnt
===========================================================
  1 11940  9614   yes   yes  none  sys.peer   reachable  1
  2 11941  9014   yes   yes  none    reject   reachable  1
  3 11942  9014   yes   yes  none    reject   reachable  1
  4 11943  9014   yes   yes  none    reject   reachable  1
  5 11944  9014   yes   yes  none    reject   reachable  1
  6 11945  9014   yes   yes  none    reject   reachable  1

The first line is actually a fake server, setup by:

  server 127.127.1.1
  fudge 127.127.1.1 stratum 10

This isn't actually an ip address, rather it is a fake
identifier for the local clock.  By setting this up, it will
allow ntpd to regard the local clock as somewhat authorative
and give that time out if it can't get anything from elsewhere.

After another few minutes, it will look something like:

ind assID status  conf reach auth condition  last_event cnt
===========================================================
  1 11940  9014   yes   yes  none    reject   reachable  1
  2 11941  9614   yes   yes  none  sys.peer   reachable  1
  3 11942  9414   yes   yes  none  candidat   reachable  1
  4 11943  9414   yes   yes  none  candidat   reachable  1
  5 11944  9314   yes   yes  none   outlyer   reachable  1
  6 11945  9114   yes   yes  none falsetick   reachable  1

This setup is on its way to equilibrium, and has chosen id 2
as the current peer, as shown by ntptrace below.

$ ntptrace 
localhost: stratum 3, offset 0.000011, synch distance 0.17929
cuscus.cc.uq.edu.au: stratum 2, offset 0.026841, synch distance 0.03589
apphys16.mst.csiro.au: stratum 1, offset 0.025527, synch distance
0.00139, refid 'ATOM'

At any point where the ntpd server has chosen a host to
synchronize with you can point ntpdate to it.

Thanks,
Brad
-- 
+=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-+
|Brad Marshall                    |           Plugged In Software|
|Senior Systems Administrator     |     http://www.pisoftware.com|
|mailto:bmarshal@pisoftware.com   |  GPG Key Id: 47951BD0 / 1024b|
+-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=+
 Fingerprint:  BAE3 4794 E627 2EAF 7EC0  4763 7884 4BE8 4795 1BD0