Comment 4 for bug 734894

Revision history for this message
Steve Langasek (vorlon) wrote : Re: [Bug 734894] Re: FFe: DBus time API should control ntpdate, not ntpd

On Mon, Mar 14, 2011 at 04:16:03PM -0000, Martin Pitt wrote:
> Oh, and thirdly, I don't really see enough of an use case for having
> ntpd on a desktop system that warrants exposing this in the GUI. ntpd
> has nonzero overhead in terms of extra CPU/memory/battery/start up
> time/CD space, and also opens a network port.

GUIs notwithstanding, ntpdate is not a proper substitute for ntpd. At best,
ntpdate gives you a one-time sync with a network time source when the
network interface comes up, after which the clock will resume drifting
unchecked. For a true "desktop system", which may have a wired network
connection that never goes down, this can easily result in clocks drifting
well outside the requirements of certain time-synced network protocols (such
as Kerberos).

I don't think this changes anything wrt the FFe or the decision about how to
handle time syncing on the desktop, I just want to be clear about the
rationale and note that there are use cases that ntpdate will not satisfy.

--
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.
Ubuntu Developer http://www.debian.org/
<email address hidden> <email address hidden>