It seems that on my SuSE 9.3 system, Yast corrupted /etc/ntp.conf a little more each time I used it to alter my ntp configuration. I emptied the file and used Yast to set up ntp again, upon which /etc/ntp.conf contained only the server line. The traffic is now greatly reduced. Thanks for your help. Regards, John -----Original Message----- From: Carl Hartung [mailto:suselinux@cehartung.com] Sent: Sunday, August 21, 2005 1:37 AM To: suse-kde@suse.com Subject: Re: FW: [suse-kde] What app in KDE to autosync my system clock with NIST? On Sunday 21 August 2005 00:04, John Schwartzman wrote:

When I setup xntp using Yast there is a tremendous amount of traffic. Using tcpdump I see the following continuously repeated:

20:44:53.282811 arp who-has Intrepid.Forte tell 192.168.2.1 20:44:53.282842 arp reply Intrepid.Forte is-at 00:80:ad:c1:35:41 20:44:54.226757 IP Intrepid.Forte.ntp > louie.udel.edu.ntp: NTPv4 client, strat 0, poll 6, prec -18 20:44:54.253991 IP louie.udel.edu.ntp > Intrepid.Forte.ntp: NTPv4 server, strat 2, poll 6, prec -22 20:44:56.228501 IP Intrepid.Forte.ntp > louie.udel.edu.ntp: NTPv4 client, strat 0, poll 6, prec -18 20:44:56.259432 IP louie.udel.edu.ntp > Intrepid.Forte.ntp: NTPv4 server, strat 2, poll 6, prec -22 20:44:58.230146 IP Intrepid.Forte.ntp > louie.udel.edu.ntp: NTPv4 client, strat 0, poll 6, prec -18 20:44:58.279255 IP louie.udel.edu.ntp > Intrepid.Forte.ntp: NTPv4 server, strat 2, poll 6, prec -22

As you can see, I'm using the server louie.udel.edu.ntp and I seem to be submitting my request again and again at 2s intervals. <snip>

Hi John, Did you read the "Where to Start" and various readmes in /usr/share/doc/packages/xntp? Did you go to the maintainer's website to look over any of the topical information available there? I only mention this because I did just that and seem to recall this process is not as cut and dried as it might appear to be at first glance. Your system needs to be brought into sync with the remote server incrementally. This is done to avoid having adjustments to the local clock disrupt other processes that are run automatically based upon time calculations. To accomplish these incremental adjustments, xntp has to perform a few chores: - determine the latency between the two systems; what's the delay between polling the server and receipt of it's response. - determine the actual time difference between the two systems. (Am I ahead or behind and by how much?) - adjust for drift in the local hardware (Am I running fast or slow and by how much). - resolve all of the above into some number and frequency of acceptably small incremental adjustments I suspect the traffic you're observing will probably drop off over time. I think on my system it ultimately ended up only polling the server at boots and maybe once per hour thereafter. regards, - Carl