after re-flashing microSD card with PiAware 7.1, noticed that NTP is being terminated and restarted every 15 minutes. Seems like this prevents optimal system clock sync.
repro steps:
flash PiAware 7.1
add flight feeder id
enable ssh
put card in RPi system
from looking at ntpq ; peer; see that reach drops periodically for all servers to low values
look at log to find, for example:
pi@piaware:~ $ cat /var/log/daemon.log | grep "ntpd exit"
Jan 23 18:23:18 piaware ntpd[654]: ntpd exiting on signal 15 (Terminated)
Jan 23 18:38:18 piaware ntpd[1409]: ntpd exiting on signal 15 (Terminated)
Jan 23 18:53:18 piaware ntpd[1569]: ntpd exiting on signal 15 (Terminated)
Jan 23 19:08:18 piaware ntpd[1672]: ntpd exiting on signal 15 (Terminated)
Jan 23 19:23:18 piaware ntpd[1773]: ntpd exiting on signal 15 (Terminated)
Any idea why this occurs at 15 minute intervals? Or how to prevent the 15 minute restart of NTP?
ntp is restarted if the network config changes (e.g. new IP). (I’m not sure this is entirely intentional - from a quick look at the hook scripts, the intention seems to be to restart only if the DHCP-provided NTP server list changed, but the actual implementation seems to restart always if no servers are provided via DHCP)
Poor peer reachability reported by ntpq implies network problems.
Maybe you have network problems and the DHCP lease is getting dropped/refreshed every 15 mins or something like that? I’d look in syslog to see what else is going on at around the time that ntp restarts.
(FWIW, I do not see 15-min restarts of ntp on a fresh 7.1 install here)
After reading this post, I noticed my system exhibited the same behavior. Though the restart times varied slightly - about every 27 minutes. The simple workaround in the link below fixed the issue for me. Comment out a single line in /lib/dhcpcd/dhcpcd-hooks/66-ntp.conf.
