Thanks, please see below. > What does 'sudo systemctl status chrony.service' output? Could you > please tell me more about chronyd's configuration?
Here's the output on startup: Oct 30 19:41:56 systemd[1]: Starting chrony.service - chrony, an NTP client/server... Oct 30 19:41:56 chronyd[1644]: chronyd version 4.8 starting (+CMDMON +REFCLOCK +RTC +PRIVDROP +SCFILTER +SIGND +NTS +SECHASH +IPV6 -DEBUG) Oct 30 19:41:56 chronyd[1644]: Loaded 0 symmetric keys Oct 30 19:41:56 chronyd[1644]: Enabled HW timestamping on enp2s0 Oct 30 19:41:56 chronyd[1644]: Enabled HW timestamping on eno1 Oct 30 19:41:56 chronyd[1644]: Enabled HW timestamping on eno2 Oct 30 19:41:56 chronyd[1644]: Could not open NTP socket on [fd00::1]:123 Oct 30 19:41:56 chronyd[1644]: Using leap second list /usr/share/zoneinfo/leap-seconds.list Oct 30 19:41:56 chronyd[1644]: Frequency 15.913 +/- 0.026 ppm read from /var/lib/chrony/chrony.drift Oct 30 19:41:56 chronyd[1644]: Loaded seccomp filter (level 1) Oct 30 19:41:56 systemd[1]: Started chrony.service - chrony, an NTP client/server. Oct 30 19:42:41 chronyd[1644]: Selected source 198.181.199.86 (time3.mbix.ca) Oct 30 19:42:41 chronyd[1644]: System clock TAI offset set to 37 seconds Oct 30 19:47:00 chronyd[1644]: Selected source 172.97.210.214 (time.web-clock.ca) The error appears to the "Could not open NTP socket" line. The likely relevant bits of configuration include: user _chrony cmdport 0 binddevice bond0 bindaddress fd00::1 fd00::1 is an address on bond0. There's no difference, notably, when I drop "bindaddress" and just try to bind to bond0, except the error becomes: Could not open NTP socket on [::]:123 Now I've noticed that if I use bindaddress but comment out binddevice, then chrony starts properly. So the problem seems to be related to the use of binddevice specifically, at least on this system.
