Time service question

I have configured our AD PDC emulator to get time updates from one of the
NIST servers. When I try to run w32tm /resync /rediscover, it sets the time
on the server properly, but then after a few minutes it resets back to the
wrong time. The server is about 3 minutes ahead of the correct time. So I
get an warning message in the event log about how the server received a time
stamp that is more than so many msecs off.

How do I get the server to "get over" the fact that the time is off by a few
minutes and catch up?




Jeff Hofstetter

Have you tried manually setting the server time as close as you can (via the
local clock GUI interface)?

-Frank

"Jeff Hofstetter" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed)...
>I have configured our AD PDC emulator to get time updates from one of the
>NIST servers. When I try to run w32tm /resync /rediscover, it sets the time
>on the server properly, but then after a few minutes it resets back to the
>wrong time. The server is about 3 minutes ahead of the correct time. So I
>get an warning message in the event log about how the server received a
>time stamp that is more than so many msecs off.
>
> How do I get the server to "get over" the fact that the time is off by a
> few minutes and catch up?
>

You can look at the default domain GPO, kerberos time policy - Maximum
tolerance for computer clock synchronization. Also, setting any windows time
service policies may also attribute to time differences.

"Jeff Hofstetter" wrote:

> I have configured our AD PDC emulator to get time updates from one of the
> NIST servers. When I try to run w32tm /resync /rediscover, it sets the time
> on the server properly, but then after a few minutes it resets back to the
> wrong time. The server is about 3 minutes ahead of the correct time. So I
> get an warning message in the event log about how the server received a time
> stamp that is more than so many msecs off.
>
> How do I get the server to "get over" the fact that the time is off by a few
> minutes and catch up?
>
>
>