Re: ntp and 0.0.0.0 address
On Sun, 30 Jan 2000, Brian May wrote:
> snoopy# ntptrace
> localhost: stratum 16, offset 0.000160, synch distance 0.69316
> 0.0.0.0: *Timeout*
Please send all the relevant configuration files attached when you report
bugs like this one...
ntpd binds to 0.0.0.0 to receive packets from any interface, and this is not
a problem. Now, it should not try to TALK to 0.0.0.0, but you didn't show
ntpd talking to 0.0.0.0, rather *ntptrace* was trying to talk to 0.0.0.0...
A small test shows that the bug is likely to be in ntptrace, and it is
harmless. It should be fixed or documented (if it isn't documented already,
I didn't check), but it isn't a release-critical (important) bug IMO.
The bug can be triggered just after you start ntpd and before it syncs to
the timeservers (in my system this takes about 5-10 minutes, so this is NOT
a small window).
ntptrace tries to connect to the upstream peer (or something like that),
which is reported as 0.0.0.0 (I think), because ntpd doesn't have one
choosen yet. Try ntpq -p -n and you'll probably see your ntpd is working
just fine (if it is correctly configured, anyway).
I didn't read the RFCs, so I don't know if a supposed 0.0.0.0 answer in the
first sync window is right (bug in ntptrace) or wrong (bug in ntpd).
> Also, ntpdate and ntpd often cannot get a response from the server
> (this seems a bit unpredicatable). Other people have been using the
Is your connection to the servers your localhost server is in sync with
down? Are you sure it is not a DNS problem?
If you often cannot connect to the *remote* server, that network path/server
is not suitable for NTP (at least not as a sole server). Try pinging the
servers when you notice you cannot connect to them, to access if the problem
is the network path.
Anyway, if you reply to the above questions, you should probably take the
BTS out of the reply headers...
--
"One disk to rule them all, One disk to find them. One disk to bring
them all and in the darkness grind them. In the Land of Redmond
where the shadows lie." -- The Silicon Valley Tarot
Henrique Holschuh
Reply to: