[ntp:questions] To check tim
kostecke at ntp.org
Wed Aug 29 19:07:58 UTC 2007
On 2007-08-28, joachim.zobel at novartis.com <joachim.zobel at novartis.com> wrote:
> to synchronize time stamping in pharmaceutical enviromental it is possible
> to connect all production machines via NTP to one Time Server.
There are some problems with using only one time server. Especially if
you don't control that server or if that server is not connected to a
real ref-clock (no, the Undiciplined Local Clock does not count).
> Question is: Is there a procedure developped to validate that all
> connected production machines do have the same time adjusted than the
> time server?
If the 'production machines' are running ntpd the simplest method is to
add all of the 'production machines' to another system, such as your
time server, as 'noselect servers'. Then the ntpq peer billboard for the
server ('ntpq -p') will show the offsets (in milliseconds) between all
of your production machines and your time server.
Use the following line for each production system in your monitoring
server production_machine_name_or_ip_address noselect
It's up to you to decide how you wish to use the information generated
in this manner, how to collect it, etc.
> This procedure must be initiated centralized e.g. on the Time Server
> and not on the connected production machines to reduce the effort for
> this test.
Polling or quering ntpd adds virtually nothing to the system load.
Steve Kostecke <kostecke at ntp.org>
NTP Public Services Project - http://support.ntp.org/
More information about the questions