pasdepoil wrote:
Hi
We have a system working only in a LAN. We install ntp on each machine
(windows 2000). We choose one of them to be the time server, and then
we configure all others machines to synchronize on this time server.
We put the same ntp.conf on every machine, but some of them are not
synchronized.
When I used monitor ntp, we can see that these machines have between 23
secondes and 150 secondes of offset.
I would like to know if the problem that I have is due to the big time
difference between machines ?
And How can I resolve the problem ?
Ntpd cannot correct errors greater than 1024 seconds! Large errors less
that 1024 seconds can take a very long time to correct; the highest rate
at which ntpd can slew the clock is 500 parts per million or about 43
seconds per day. If a clock has a frequency error greater than 500
parts per million, ntpd cannot correct it at all.
If machines are not synchronized:
a. ntpd is not running, or
b. ntpd cannot contact the server and receive a reply, or
c. ntpd cannot correct the clock because either the offset or the
frequency error is too large, or
d. ntpd does not believe that the server is "suitable for synchronization".
If you have ntpq installed, try "ntpq -p". This should show whether the
server is reachable or not, what the offset is, the poll interval, etc.
Last, but not least, NTP was not designed to be used the way you are
using it; e.g. without an external source of time or a hardware
reference clock. If it works for you, you are lucky. If not, you will
find sympathy in the dictionary!
_______________________________________________
questions mailing list
[email protected]
https://lists.ntp.isc.org/mailman/listinfo/questions
