[ntp:questions] ntpd and database servers

David Woolley david at ex.djwhome.demon.invalid
Fri Jan 29 07:29:25 UTC 2010


John Hasler wrote:
in that that.
> 
> I think his idea is that Ntpd should give out the time it is slewing
> towards rather than its current time.  Thus it would compute the time
> it would be giving out had it made a step instead of slewing and then
> give out that time while it slews.  Aside from the implementation
> difficulties, that could lead to the server and its clients using
> substantially different times during the slew.

In which case, the reference to dispersion would be completely bogus.

Incidentally, I believe this is what ntpd V3 did do.




More information about the questions mailing list