ntpd not synchronizing - again :-(

I have just installed Linux (Slackware 10.2) on a box that used to
run an older Linux distribution (Slackware 10.0). With this older
distribution ntpd was working fine. In the new one it isn't.

With ntpd up and running, I decided to launch a small script to
print out regularly the time according to the hardware clock, and the
output from the ntpdate -q xxx command, where xxx
is my local NTP server, with IP address 192.168.0.1. This server in
turn takes its NTP feed from some external ones, and other boxes in my
LAN synchronize with it without any problems. The first lines of
output from that script are

Thu 06 Jul 2006 08:13:15 AM PDT -1.269066 seconds
server 192.168.0.1, stratum 3, offset -0.103281, delay 0.02702
6 Jul 08:13:15 ntpdate[3431]: adjust time server 192.168.0.1 offset
-0.103281 sec

The very first line is the time as reported by the hardware clock
(the output from the hwclock --show command) whereas the last two are
the output from ntpdate -q xxx. In two minute intervals, this is the
output I got:

Thu 06 Jul 2006 08:15:19 AM PDT -0.919212 seconds
server 192.168.0.1, stratum 3, offset -0.384778, delay 0.02705
6 Jul 08:15:19 ntpdate[3488]: adjust time server 192.168.0.1 offset
-0.384778 sec

Thu 06 Jul 2006 08:17:23 AM PDT -0.969468 seconds
server 192.168.0.1, stratum 3, offset -0.777177, delay 0.02704
6 Jul 08:17:23 ntpdate[3655]: step time server 192.168.0.1 offset -0.777177 sec

Thu 06 Jul 2006 08:19:27 AM PDT -0.948820 seconds
server 192.168.0.1, stratum 3, offset -1.171127, delay 0.02704
6 Jul 08:19:28 ntpdate[3864]: step time server 192.168.0.1 offset -1.171127 sec

Thu 06 Jul 2006 08:21:31 AM PDT -0.958195 seconds
server 192.168.0.1, stratum 3, offset -1.543037, delay 0.02704
6 Jul 08:21:32 ntpdate[4045]: step time server 192.168.0.1 offset -1.543037 sec

Thu 06 Jul 2006 08:23:35 AM PDT -0.962411 seconds
server 192.168.0.1, stratum 3, offset -1.957947, delay 0.02705
6 Jul 08:23:36 ntpdate[4247]: step time server 192.168.0.1 offset -1.957947 sec

When I finally stopped it the ouput was

Thu 06 Jul 2006 08:35:00 AM PDT -0.731368 seconds
server 192.168.0.1, stratum 3, offset -4.185915, delay 0.02702
6 Jul 08:35:00 ntpdate[24616]: step time server 192.168.0.1 offset
-4.185915 sec

Can anybody explain why the OS clock is drifting so dramatically?
Also, is it the case that it running in lockstep with the hardware
clock? I thought that the OS clock and the hardware clock were two
different things.

My ntpd configuration file is as follows:

server 127.127.1.0 # local clock
fudge 127.127.1.0 stratum 10
server 192.168.0.1
driftfile /etc/ntp/drift
multicastclient # listen on default 224.0.1.1
broadcastdelay 0.008
_______________________________________________
questions mailing list
[email]questions@lists.ntp.isc.org[/email]
[url]https://lists.ntp.isc.org/mailman/listinfo/questions[/url]

Hi eveybody;

Previously I asked NTP server implementation status or related
recommendations in Vxworks, but I havent get response any one yet except
Harlan Stenn... That's okay, thanks all.

During the that time I learned that vxWorks has SNTP client and server
support.

In our project there is an izaolated network and one an other computer act
as NTP server and we havent get yet this computer and our computer which has
vxWorks OS have to act as NTP client. First thing first I want to test our
side. Because of this I need NTP server to use it in our local izolated
network to test our software application which will act as NTP client or for
beginning act as SNTP client...

Could anyone recommend where can I find ntp server that can be used in local
izolated network and how I can setup test enviroment to test our software
application which will act as NTP client or SNTP client and which
communicates with NTP server in local izolated network...

Please let me get response...

See you all...

Thanks in advance...
EGE


Configuration:
Tornado 2.02 -> vxWorks 5.4
Motorola Powerpc target
Codebase language C++

_______________________________________________
questions mailing list
[email]questions@lists.ntp.isc.org[/email]
[url]https://lists.ntp.isc.org/mailman/listinfo/questions[/url]

ntpd is an ntp client or server. If it is not ported to your machine you
can either port it (we can help) or see if there is some other ported
application that wil do it.

Plan B would be to find an old x86 box, install FreeBSD (or some other free
os of your choice) and put ntpd on it and use it as your local time server.

H

JCA wrote:[color=blue]
> Can anybody explain why the OS clock is drifting so dramatically?
> Also, is it the case that it running in lockstep with the hardware
> clock? I thought that the OS clock and the hardware clock were two
> different things.
>
> My ntpd configuration file is as follows:
>
> server 127.127.1.0 # local clock
> fudge 127.127.1.0 stratum 10[/color]

Are you serving time to other systems? If not, it's not needed.
[color=blue]
> server 192.168.0.1[/color]

You were querying this server, yet that's the address in here, so I'm
not sure what you were getting.
[color=blue]
> driftfile /etc/ntp/drift
> multicastclient # listen on default 224.0.1.1[/color]

There is no default. The address is required. Do you have a multicast
server set up somewhere? Otherwise this option has no meaning and since
you didn't disable authentication it will in any case ignore any
multicast packets that it does receive even if it were configured properly.
[color=blue]
> broadcastdelay 0.008[/color]

On a LAN why would you put this in at all? This should not be here
unless you have a good measure of the propogation delay AND ntpd is not
doing its own calculation because it is not querying the multicast
server at all (it's part of what authentication would do for you).

Danny
_______________________________________________
questions mailing list
[email]questions@lists.ntp.isc.org[/email]
[url]https://lists.ntp.isc.org/mailman/listinfo/questions[/url]