From owner-freebsd-questions Mon Jun 24 09:41:21 1996 Return-Path: owner-questions Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id JAA03995 for questions-outgoing; Mon, 24 Jun 1996 09:41:21 -0700 (PDT) Received: from horst.bfd.com (horst.bfd.com [204.160.242.10]) by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id JAA03989 for ; Mon, 24 Jun 1996 09:41:19 -0700 (PDT) Received: from harlie.bfd.com (bastion.bfd.com [204.160.242.2]) by horst.bfd.com (8.7.5/8.7.3) with SMTP id JAA14257 for ; Mon, 24 Jun 1996 09:41:20 -0700 (PDT) Date: Mon, 24 Jun 1996 09:41:18 -0700 (PDT) From: "Eric J. Schwertfeger" To: questions@freebsd.org Subject: tickadj questions Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-questions@freebsd.org X-Loop: FreeBSD.org Precedence: bulk I'm working on syncronizing the clocks on our network, and as the first step, I wanted to use tickadj to cooerce the clocks to be as accurate as possible by themselves, so that they will stay accurate without an external ntp connection. tickadj seems to be the right command on the Solaris and FreeBSD machines. The questions I have are 1) new_tick: as I understand it, this tells the system how many microseconds should pass before it ticks the hz counter. Is this correct? This gives the ability to correct to 100PPM, or about 8 seconds. 2) new_tickadj: I'm not sure I understand this one. It obviously isn't simply added to tick, otherwise there would be no point. I suspect that it is added to the tick count once a second, for adjustments down to 1PPM. 3) Does FreeBSD read or write the CMOS clock while running? I know with SunOS 4.0, I have to turn this off. For example, the machine that will probably be our internal NTP server is just over 14 seconds fast a day. by my understanding, I should either a) add 1 to the tick value (8.6 seconds per day), and set tickadj to 62 (-62?). b) add 2 to the tick value (overshooting the correction) and set tickadj to 38 (-38?). Figures the hardware with the worst clock, the one that is off not in parts per million, but parts per hundred, would be running Linux, which doesn't have tickadj.