[ntp:questions] "ntpd -q" is slow compared to ntpdate

Richard B. Gilbert rgilbert88 at comcast.net
Sun Oct 19 19:17:40 UTC 2008


Unruh wrote:
> David Woolley <david at ex.djwhome.demon.co.uk.invalid> writes:
> 
>> Steve Kostecke wrote:
> 
>>> This discussion dates back to August, 2002.
>>>
>>> Please see
>>> http://groups.google.com/group/comp.protocols.time.ntp/msg/3387eda49518407a?dmode=source
>>>
>> The argument there seems to be not that ntpdate is worse than SNTP, but 
>> that there is an implied promise that it is almost as good as ntpd.  The 
>> reason for having SNTP is to have something which has no pretensions of 
>> being anything other than crude.
> 
> ??? The rfc on sntp seems to say that sntp should be just as good as ntp in
> disciplining a local clock, it just should not be used as a server (
> uneless it gets its time from an atomic clock). Ie, the promises for sntp
> seem far stronger than for ntpdate. ntpdate makes no promise to discipline
> the frequency of a computer's clock. sntp does as far as I can see. 
> 
> The claim is that ntpdate should be retired because a) it is written in a
> way which is hard to maintain, and  support, and b) it is flawed. The first
> is a perfectly valid concern. The second I would like to know how it is
> flawed. AFAIK all it does is to set the local clock to the time as
> determined via ntp packet exchange from some server. it does not set the
> frequency, it just sets the time. Is there some flaw in the way in which it
> sets the time? Could I run ntpdate with a reliable server as a source and
> suddenly discover that my local clock is out by 79 days, or that the
> frequency has been reset to 1 tick per century? Ie, is there a flaw in what
> it claims to do?
> 
> 

I believe that one of the biggest problems with ntpdate is that it has 
no volunteer maintainer; no one is responsible for it.  Wrapping your 
mind around someone else's code can be damned near impossible unless the 
code is well written and well documented.  I haven't looked at the code 
but I suspect that both attributes are lacking.

Since sntp can do most, if not all, of what ntpdate does, there is 
little or no incentive to bring ntpdate up to standard!

Since ntpdate does something useful and at least appears to do it 
correctly, it will probably be around for quite a while.





More information about the questions mailing list