[ntp:questions] NTP local clock on Windows - updated reference time?
awilt at omneon.com
Fri Nov 11 01:16:12 UTC 2005
I've installed NTP 4.2.0b for Windows using the excellent installer from
Meinberg, and I've set it up to use the local clock as its sole
reference (for isolated networks installed in TV stations; yes, they
should have a reference clock and/or a connection to better time
servers, but that's not something I have control over, and no, it's not
the automation network; that one HAS a good reference clock!).
Using ntpdate -d from various Linux boxes, which will be using the
Windows box as their time server (yes, I know this is backwards; again,
this is not my architectural choice), I see that the reference time
reported by NTP is set once, then never again. Eventually, ntpdate
refuses to sync to it as the server's reference time has gone too long
without an update.
In the event log, I see the message "synchronized to LOCAL(1), stratum
12" when the reference time gets updated and the server's reported
stratum goes from 16 to 13, and then the message "no servers reachable"
appears at the next update, i.e., when the reach becomes 037.
When using Meinberg's NTP Time Server Monitor to watch the status, I see
LOCAL(1) go green and get marked as the source of synchronization ('*')
when reach hits 017, and then go gray and lose its '*' status tag when
reach becomes 037.
When I set up a Linux NTP server with the same configuration:
fudge 127.127.1.1 stratum 12
the reference time that comes back in the ntpdate query is updated every
time the local clock is read. In the Meinberg monitor, LOCAL(1) goes
green and gets the '*' tag when reach becomes 0.17, and then it stays
that way thereafter.
What am I doing wrong? How do I get NTP on Windows to keep updating its
reference time from the local clock?
More information about the questions