[chrony-users] Occasionally high root dispersion shortly after start

[t.barnewski@xxxxxx]

I'm running chrony 4.3 on a router/gateway device that acts both as an NTP
client as well as LAN NTP server.

In somewhere between 33%-50% of all cases, directly after starting chrony, its
root dispersion steadily rises by a second per second until the first polling
interval has elapsed, after which it returns to normal (~0.03s) values.

The relevant part of my chrony.conf could look like this:

    server 5.196.160.139 minpoll 10 iburst
    server 194.95.183.223 minpoll 10 iburst
    makestep 1000 10

I've taken two random addresses from from 2.europe.pool.ntp.org for this
example, but the issue is visible with all public NTP servers I've tested.
For example, after about 7 minutes after starting, the outputs of tracking,
sources and sourcestats are:

    Reference ID    : 05C4A08B (ip139.ip-5-196-160.eu)
    Stratum         : 3
    Ref time (UTC)  : Tue Jan 24 13:33:24 2023
    System time     : 0.000000000 seconds slow of NTP time
    Last offset     : +0.000006036 seconds
    RMS offset      : 0.000006036 seconds
    Frequency       : 0.000 ppm slow
    Residual freq   : +7.203 ppm
    Skew            : 1000000.000 ppm
    Root delay      : 0.038680598 seconds
    Root dispersion : 428.669036865 seconds
    Update interval : 2.1 seconds
    Leap status     : Normal

    MS Name/IP address     Stratum Poll Reach LastRx Last sample
    ===========================================================================
    ^* ip139.ip-5-196-160.eu     2  10    17   433  -9908ns[-3872ns] +/-   20ms
    ^+ afsdb1-dmz.cbs.mpg.de     2  10    17   433   -175us[ -169us] +/-   38ms

    Name/IP Address        NP  NR  Span  Frequency  Freq Skew  Offset  Std Dev
    ==========================================================================
    ip139.ip-5-196-160.eu   4   4     6     +4.787   1694.587  +2208us   262us
    afsdb1-dmz.cbs.mpg.de   4   3     6     +8.709   1337.888  +3593us   230us

Note the root dispersion of more than 400 seconds. As a result, some clients
discard NTP responses from this chrony instance for almost 20 minutes.

Network operations employing our devices have stated they strongly prefer
minimizing NTP traffic and server load, so lowering minpoll is not a viable
option, unfortunately.

Is this behaviour intentional? If so, is there a configuration option I could
set in chrony.conf to work around the issue?

Kind regards,
Tim Barnewski
AVM GmbH

--
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.