[ntp:questions] Re: limitation on the client time/date at ntp startup

user at domain.invalid user at domain.invalid
Mon Sep 11 19:17:11 UTC 2006


Per,

I may have to wash out my mouth with soap after telling you this.

The xntp daemon maintained two timescales, one for the daemon and the 
other for the kernel. There could be a large discrepancy between these 
timescales that was not revealed by the monitoring functions. Secondly, 
late xntpd and and eraly ntpd used two independent loops, one for "slew" 
mode and the other for "normal" modes. This was a mistake; the slew is 
now incorporated in the normal model. The only difference is that slew 
normally means simply increasing the step threshold; the loop functions 
remain the same.

Note that some systems, including Solaris, have modified the adjtime() 
syscall function to more quickly respond to large adjustments. This adds 
an extra poll to the feedback loop response. Setting a large step 
threshold can result in large overshoot and possible unstable behavior.

Dave

Per Hedeland wrote:

> In article <T1157455870 at djwhome.demon.co.uk> david at djwhome.demon.co.uk
> (David Woolley) writes:
> 
>>In article <ecl43a$n38$1 at scrotar.nss.udel.edu>,
>>David L. Mills <mills at udel.edu> wrote:
>>
>>
>>>Classic way to test NTP functionality is to stop ntpd, set the time by 
>>
>>This isn't a classic way because it hasn't been an available option for
>>long enough.
> 
> 
> The option of stopping ntpd?:-) If you mean -g, I think it has been
> around since xntpd/v3 days, though possibly with slightly different
> semantics back then (I'm not sure if it was limited to *one* step).
> Note, I'm not saying it was *documented*.:-)
> 
> --Per Hedeland
> per at hedeland.org




More information about the questions mailing list