Re: [chrony-users] Strange "loss of PPS lock" |
[ Thread Index |
Date Index
| More chrony.tuxfamily.org/chrony-users Archives
]
- To: chrony-users@xxxxxxxxxxxxxxxxxxxx
- Subject: Re: [chrony-users] Strange "loss of PPS lock"
- From: Rob Janssen <chrony-users@xxxxxxxxx>
- Date: Thu, 19 Jan 2023 10:06:00 +0100
- Dkim-signature: v=1; a=rsa-sha256; c=simple/simple; d=pe1chl.nl; s=pe1chl; t=1674119161; bh=G6q1zIu2T0UGxhyR8EsNPf1t+PttVH2br4EKF+oj4to=; h=Date:Subject:To:References:From:In-Reply-To:From; b=yvYkPH1sTfKtKmugk2jLpmF/xwQtbLSBeMTMHZlBxGOLngPnmvwOvA1bRhYbJL33+ RarA4nwDcldY+FjSpkwMRDHMfcOlL14V3fdDG9ICIt4+2+L3oLB+ivE81ZeyWlzEro qATc2/HiM2T95Z0KeppsqvDRVCy3zH2NytbNpyo5q+xkKrTdPEBGpkP8HTXnUZEQ4B cI2rUdHC+kB1yYn6oradhsCzkWpfaDQ3AqeTrIcnM4dc6+PYHtjWTB5inJLxXn8gEO ALhE1jFXIJAGFEHbyo09mZVCO2rHh9d1soPGbOVFC/l+BAomEPF9iY87ppgQo5WQaz aF3h/W9O5B4pQ==
- Organization: PE1CHL
On 1/19/23 09:51, Miroslav Lichvar wrote:
> On Wed, Jan 18, 2023 at 07:13:59PM +0100, Rob Janssen wrote:
>> No the PPS for sure did not lose lock. These are either professional GPSDO
>> sources or we have that extra lock source, and it happens in both cases.
>> What SURE could happen is that those NTP sources collectively drift to
>> some offset relative to the PPS due to asymmetric network delay, but
>> in that case the PPS still is the correct reference.
> It wouldn't be a problem if it was just asymmetric delay, because that
> would be covered in the root distance. The servers would have to be
> lying about that.
Note that each of the servers has a local PPS source and a couple of
NTP servers that are all polled via the same network. When there is
an asymmetric delay in that network, all NTP servers will show the same
or similar offset in "chronyc sources", but the system is still locked to
PPS. My hypothesis is that at some point chrony decides that the PPS
must be a falseticker because all other sources have a ~2ms offset that
is within ~100us of the others. At that time it switches to a random
NTP source and unlocks from the PPS. It seems to be satisfied with the
new situation as it takes ~1 hour for it to go back to the PPS time.
>
> Are they stratum 1 servers? What root delay and dispersion is reported
> in chronyc ntpdata?
They are in the region of 15-30us. Indeed all those servers are
stratum 1, all running chrony and with a PPS via serial DCD.
>
>> When a PPS source is marked as "trust" but its lock source is not locked,
>> is it still trusted or will other sources then take over (they should)?
> The PPS source will be unreachable if its lock source is unreachable
> and will be ignored in the source selection unless you add the require
> option.
>
Ok I have added "trust" to the PPS config lines and restarted the services,
now let's see if it still happens.
That is difficult to determine, sometimes it happens a couple of times in a
week, sometimes it does not happen for 2 months. It could be related to
the loading of the network, I could try to "force" that a bit.
Rob
--
To unsubscribe email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
with "unsubscribe" in the subject.
For help email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
with "help" in the subject.
Trouble? Email listmaster@xxxxxxxxxxxxxxxxxxxx.