strange ntpd behaviour when modyfing stratum
I'm running ntp-4.2.4p4 on FreeBSD.
I was observing today the ntpd with ntpq and manipulating the stratum
number of local time sources in ntp.conf.
I've noticed that when I've disabled all pps sources and worked only
with two gps antennas (receiving to ntp nmea frames only, ignoring pps,
I set stratum 8 for antenna0 and stratum 12 for antenna1) the ntp has
chosen antenna0 (so the source with higher stratum) and that's what I've
But then I've set all the sources in ntp.conf (gps nmea and pps, ocxo
pps, rubidium pps) and put stratum 12 option near one of them and
stratum 8 near the others. I saw that ntp has chosen the pps source that
had the lower stratum number (12). The same situation appeared when
setting stratum 12 option to another pps source. Every time the pps
source with the lower stratum number was chosen(even gps pps (having
stratum 12 number) had '*' in ntpq when ocxo and rubidium had '+'
(having stratum 8 number)for all the time). During that all of the
sources had very nice and nearly the same delay, offset and jitter.
Has someone any idea why it behaved so strange with pps sources?
Thank you for your help.