In article <[EMAIL PROTECTED]>, Uwe Klein <[EMAIL PROTECTED]> wrote:
> quality and/or stratum > with/without the local clock > being used as (additional) server This should make no difference to the true quality. It will result in the server reporting a falsely optimistic root dispersion and failing to drop back to stratum 16; the latter when 8 consecutive polls have failed for all servers. (The continuing increase in root dispersion is where I think w32time, Win 2003 SP1, actually gets it right; ntpd assumes an external synchronisation source in its generation of error estimates.) > over time? Depends on how stable the system was when it lost the synchronisation source. If it was dead on frequency and in an air conditioned room, one could reasonably hope for 30 seconds a year (the best way of getting this is to calibrate the frequency over several days, rather than using ntpd to measure it). _______________________________________________ questions mailing list [email protected] https://lists.ntp.isc.org/mailman/listinfo/questions
