[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: OpenNTPD does not 'pull-in' wrong time
- To: misc_(_at_)_openbsd_(_dot_)_org
- Subject: Re: OpenNTPD does not 'pull-in' wrong time
- From: Uwe Dippel <udippel_(_at_)_uniten_(_dot_)_edu_(_dot_)_my>
- Date: Fri, 16 Dec 2005 06:51:14 +0800
On Wed, 14 Dec 2005 15:32:13 +1100, Darren Tucker wrote:
> By my rough calculations, your system clock is drifting at about 1.6%,
> which is more than adjtime can correct for (roughly 0.5%).
All fine, and no flame intended:
There have been people bragging about openntpd to be advantageous compared
to standard ntpd (in the ports) for 'pilling in' any offset. ntpd
simply doesn't start when it is off by more than a certain amount.
If openntpd can correct around 0.5%, how many days / weeks would it take
then to pull in a very wrong clock !?
I wonder if the refusal of ntpd to start does not make some sense at least
With all respect to openntpd: Isn't it a false 'warm feeling' to have
it running, when it can't possibly never ever pull in a drift of 1.6 %
- but rather drifts apart ever further ? Or has to catch in a badly
offset clock of 8 hours (e.g.) and runs off sync for ages before
reaching correct time ?
In short: should it not have a control built-in warning the user
(syslog at least) that it will never catch up or only within days ?
(and suggests or does run rdate) ?
It would have at least saved me early on in this case.