I have a few Mikrotiks configured as simple NTP servers using the optional NTP package, like so:
[2400@baud] /system ntp server> print
enabled: yes
broadcast: no
multicast: no
manycast: no
broadcast-addresses:
[2400@baud] /system ntp client> print
enabled: yes
mode: unicast
primary-ntp: x.x.x.1
secondary-ntp: (nearby NIST NTP server)
dynamic-servers:
status: synchronized
Sometimes, at boot, the Mikrotik’s NTP server doesn’t come up. The server module is loaded, but one can’t point a simple NTP client at it (e.g. ntpdate -d on a Unix system) and get responses. In all cases, the NTP client part on the Mikrotik side is happy. The IPs are reachable and serving time (the primary is the upstream router itself) and the status is “synchronized”.
When this occurs, it seems like the only way to fix is to reboot the router. No amount of “disable/re-enable the NTP server/client” jostling seems to fix it. I’ve started seeing this more frequently with 6.3* releases of RouterOS. There’s often no configuration changes between reboots. There’s no extreme traffic or load or anything like that naling these routers at boot-time.
Any advice on what to do the next time I see this to better debug it?