[ntp:questions] Re: ntp 4.2.0 and connected sockets

Wolfgang S. Rupprecht wolfgang+gnus20031103T143445 at dailyplanet.dontspam.wsrcc.com
Mon Nov 3 23:06:42 UTC 2003


(Wolfgang S. Rupprecht) writes:
> I'm seeing some weirdness with how ntpd 4.2.0 is initializing the UDP
> sockets.  1) It seems to open quite a lot of redundant sockets if one
> has alias IP's on one's interface.  2) It appears to use connected UDP
> sockets with the totally off-the-wall remote address of [64.0.0.0].
> Is this intentional?
>
>     *:* -> 64.0.0.0.ptr.us.xo.net:*

Problem solved.  It was just an artifact of me upgrading my kernel in
preparation to upgrading user-land.  Even thought the external API
didn't change (and the version number stayed the same) some internal
struct that only a few kvm grovelers use changed.

Now off to figure out why my m12 oncore doesn't like the default
ntp-4.2.0 refclock logic.  I can get it to work when I use "mode 1"
with a lat/lon/alt in the /etc/ntp.oncore conf file, but if I ask it
to do a site-survey (mode 2,4) it simply sits there like a lump on a
log for over 8hrs.

Does anyone here use an m12 oncore (firmware v1.3) with ntpd-4.2.0?
It looks like the m12 may not actually do a hardware site-survey,
although it seems to accept the command and answer with the expected
"@@Gd3" .  Neither of the "site survey" or "position hold" bits in the
@@Ha status word get set to anything either.  It is as if the command
is quietly ignored.

-wolfgang
-- 
Wolfgang S. Rupprecht 		     http://www.wsrcc.com/wolfgang/
           The From: address is valid.  Don't mess with it.



More information about the questions mailing list