On the server, I stop the chronyd daemon (ver 3.5).
I start it and send a NTP3 CLIENT query, to localhost port 123/udp:
NTP_VERSION << 3 | NTP_MODE_CLIENT
The first answer has the status 0xdc : the bit NTP_LEAP_NOTSYNC is set. Good: the daemon is starting and looking for the Ntp servers.
I query again the daemon, and in the following replies from the server:
- the status is OK : 0x1c
- the server time (reference timestamp, at offset 16 of the Udp packet payload) is incorrect.
The time values are random, below 1/1/1970 (Ntp time 2208988800). After a few queries, the server time changes to a sane value.
Why does chronyd report random time values, without NTP_LEAP_NOTSYNC ?
- is this compliant to RFC 1305 ?
- is there any other info in the response payload I have to check ?