[ntp:questions] NTPD on Win XP problem

David Mitchell nntpatmitchdj.globalnet.co.uk
Mon Oct 13 10:12:23 UTC 2003


I am running ntpd 4.1.2 on WinXP pro using the NT port on
http://www.norloff.org/ntp/ (I have also used 4.1.72 from the same
source)

There is a 'strange' problem with this, I am not absolutely sure but I
think it started when the worm/virus (MSBlaster?) was going around
using NTP servers.  I have certainly used 4.1.72 on NT 4.0 without
problems.

The configured peers seem to duplicate each polling cycle..

If I have a simple ntp.conf file with the single (active) line
"server europe.pool.ntp.org" then the following happens...

Start the service, running ntpq peer gives:

     remote           refid      st t when poll reach   delay   offset
jitter
==============================================================================
 zirkon.biophys. mars.rz.uni-due  2 u    1   64    1   85.241  -121040
0.004

after the first poll interval the peer command gives:
     remote           refid      st t when poll reach   delay   offset
jitter
==============================================================================
 zirkon.biophys. mars.rz.uni-due  2 u   90   64    2   85.241  -121040
0.004
 zirkon.biophys. 0.0.0.0         16 u    -   64    0    0.000    0.000
4000.00

(IE the server has been duplicated)

after the 3rd poll it gives:
     remote           refid      st t when poll reach   delay   offset
jitter
==============================================================================
 zirkon.biophys. mars.rz.uni-due  2 u  146   64    4   85.241  -121040
0.004
 zirkon.biophys. 0.0.0.0         16 u    -   64    0    0.000    0.000
4000.00
 zirkon.biophys. mars.rz.uni-due  2 u   10   64    1   81.601  -121042
0.004

(Note that in this last case the server was reachable)

and so it continues!  If I configure multiple servers then all of the
servers are duplicated, it seems to make no difference whether the
servers are reachable or not...

I also get 'runt response received 0 octets' errors in the system log
and occasional 'overlapped I/O in progress' errors especially with
multiple servers configured.

The network connection is via ISDN dialup, but even if the line is
held up the above still occurs.  The line is not excessively busy -
the cisco router is in front of me and I can watch the activity
lights!  The router is using NAT but I have looked with a sniffer and
the packets are getting back to the sending PC.  I have tried with and
without the burst and iburst parameters (calldelay doesn't seem to
work...), and also with specific servers coded in the cfg file (I went
to the p.n.o servers to see if it was any better!)...

I am sure it is a dumb error somewhere!

HELP.....!



More information about the questions mailing list