[ntp:questions] Re: What's different about these (S)NTP servers?
grosen at cc3.com
Thu Oct 30 00:58:03 UTC 2003
On Thu, Oct 30, 2003 at 12:31:41AM +0000, Harlan Stenn wrote:
> I have a vague recolleciton that the "reference" time on the response from
> the misbehaving server was 0, while it was nonzero in the response from the
> behaving server.
Yes, you're right. I apologize for not reproducing that text:
mix:~# ntpdate -d -q forest
reference time: 00000000.00000000 Thu, Feb 7 2036 1:28:16.000
originate timestamp: c34a63a0.00000000 Wed, Oct 29 2003 11:04:48.000
transmit timestamp: c34a631b.ea52bd3c Wed, Oct 29 2003 11:02:35.915
29 Oct 11:02:35 ntpdate: no server suitable for synchronization found
mix:~# ntpdate -d -q bkupserv
reference time: c34a19e1.019d89e9 Wed, Oct 29 2003 5:50:09.006
originate timestamp: c34a6104.9916872a Wed, Oct 29 2003 10:53:40.598
transmit timestamp: c34a6105.1e612839 Wed, Oct 29 2003 10:53:41.118
29 Oct 10:53:41 ntpdate: step time server A.B.C.160 offset -0.520515 sec
> I also have a vague recollection that a 0 reference time means "I'm not
Yes, I see that in RFC 958 now that I know to look for it:
This is a 64-bit timestamp established by the server or client
host as the timestamp (presumably obtained from a reference clock)
most recently used to update the local clock. If the local clock
has never been synchronized, the value is zero.
So it would be a safe bet that, presuming Microsoft's SNTP
implementation updates its reference timestamp when it's done
synchronizing, ntpdate (and ntpd) will then believe this is a valid
server. Thanks, that's exacty what I was looking for!
I think that it would be useful for ntpdate (and ntpd's) "no
suitable server" messages to be a bit more verbose, especially if
I've asked for verbose (debugging) output. Something like "server is
not synchronized" doesn't seem like much to ask. Perhaps such a
message is present in a more recent version of NTP?
gabriel rosenkoetter / grosen at cc3.com / CC3 Unix & Linux sysadmin
More information about the questions