From owner-freebsd-stable Tue Nov 20 12:57:34 2001 Delivered-To: freebsd-stable@freebsd.org Received: from netau1.alcanet.com.au (ntp.alcanet.com.au [203.62.196.27]) by hub.freebsd.org (Postfix) with ESMTP id 4FA1B37B416 for ; Tue, 20 Nov 2001 12:57:30 -0800 (PST) Received: from mfg1.cim.alcatel.com.au (mfg1.cim.alcatel.com.au [139.188.23.1]) by netau1.alcanet.com.au (8.9.3 (PHNE_22672)/8.9.3) with ESMTP id HAA15980; Wed, 21 Nov 2001 07:57:27 +1100 (EDT) Received: from gsmx07.alcatel.com.au by cim.alcatel.com.au (PMDF V5.2-32 #37641) with ESMTP id <01KAYIN8V6FKVFJDLD@cim.alcatel.com.au>; Wed, 21 Nov 2001 07:57:24 +1100 Received: (from jeremyp@localhost) by gsmx07.alcatel.com.au (8.11.1/8.11.1) id fAKKvOL74175; Wed, 21 Nov 2001 07:57:24 +1100 Content-return: prohibited Date: Wed, 21 Nov 2001 07:57:24 +1100 From: Peter Jeremy Subject: Re: ntpd problems? In-reply-to: <3BE8D67E.60C85827@trewitt.org>; from glenn@trewitt.org on Tue, Nov 06, 2001 at 10:36:58PM -0800 To: Mike Harding Cc: Glenn Trewitt , stable@FreeBSD.ORG Mail-Followup-To: Mike Harding , Glenn Trewitt , stable@FreeBSD.ORG Message-id: <20011121075724.A74112@gsmx07.alcatel.com.au> MIME-version: 1.0 Content-type: text/plain; charset=us-ascii Content-disposition: inline User-Agent: Mutt/1.2.5i References: <20011106141816.9F53A13573@netcom1.netcom.com> <3BE8D67E.60C85827@trewitt.org> Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG On Tue, Nov 06, 2001 at 10:36:58PM -0800, Glenn Trewitt wrote: >Well, since your log shows the time getting stepped only one way, it has >to be your clock. It may be useful to enable ntpd's peer and loop logging and have a look at the result. I notice from your other mail that your delay figures are all over the place: The two examples you posted have delays differing by >1000msec. The first report shows differences of 400msec in delay and 180msec in offset between the two servers. The NTP protocol is relatively poor at handling this sort of thing - it assumes that delays are symetric - which they aren't in your case. >If you want to get rid of the messages, you can specify the "-x" flag to >ntpd. However, read the man page first and make sure that ntpdate is >called successfully at boot time, otherwise you could end up with a >badly mis-set clock for a very long time. I'd recommend keeping a close eye on the system if you do this. I've had problems on every system where I've tried this. If the offset ever reaches the point where it would have stepped, the ntpd PLL appears to become unstable. Specifically, once ntpd detects the need to perform a large time slew, it will never regain synchronization - the loop frequency jumps to an incorrect value and the time will then regularly slew to compensate. Peter To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message