Date: Thu, 05 Aug 1999 14:36:12 -0600 From: Chris Fedde <cfedde@fedde.littleton.co.us> To: Glenn Johnson <gljohns@bellsouth.net> Cc: freebsd-questions@FreeBSD.ORG Subject: Re: How to get xntpd to work Message-ID: <199908052036.OAA94406@fedde.littleton.co.us> In-Reply-To: Your message of "Wed, 04 Aug 1999 21:52:31 CDT." <19990804215231.A1974@gforce.johnson.home>
next in thread | previous in thread | raw e-mail | index | archive | help
It has been my habit to chime my masters to more than one public time source. Xntpd uses an interpolation algorithm to converge on the right time so more than one server makes the time more accurate. Also more than one source makes the configuration more robust. this way you can have a server fail and still stay in sync. Go to http://www.eecis.udel.edu/~mills/ntp/clock2.htm and pick three public servers to chime your primary. If you are going to use 127.127.1.1 (the hardware cmos clock) then you need to have it listed as a server. server 127.127.1.1 fudge 127.127.1.0 stratum 10 Stratum numbers give an indication about how close the clock is to a real clock. Stratum 1 is synced to an atomic clock, stratum 2 syncs from a stratum 1 clock, and so on. Stratum 16 marks an unusable clock. Your error message indicates that the client thinks that server is unusable. You should be able to achieve 1 millisecond stability quite easily on a modern Intel based PC. That accracy is usually more than reasonable when you are attempting to merge syslog style logs from separate systems. If you are a small network of 1 to say 20 systems then I'd set up one master chiming three external sources and have all the internals list the master as a server and eachother as peers. That prevents drift when the master source is unreachable. For more than 20 systems I'd have at least two systems that chime to external servers and each other. The set of external servers should be disjoint between the servers. Some people prefer to have one server chime externally then use timed internally or use periodic calls to ntpdate. This can work but it is vulnerable to sync rot when the master is unreachable. There are good discussions of all this stuff available on http://www.eecis.udel.edu/~ntp chris Glenn Johnson writes: I am trying to set up time synchronization on 5 machines. What I want to do is have one machine, call it host0, sync to a public ntp server and have the other four machines, host{1,2,3,4} sync to the local xntpd server. I have read the xntpd(8) and searched the archives but can't get it. On host0 I have the following in '/etc/ntp.conf': server black-ice.cc.vt.edu driftfile /etc/ntp.drift fudge 127.127.1.1 stratum 4 After starting xntpd on host0 I tried 'ntpdate -q host0' on the other machines. I get a message that host0 is stratum 16, followed by a 'no host suitable for synchronization message'. Why is host0 stratum 16? I did an ntpdate to the public server prior to starting xntpd to get the time right at the start. I have had xntpd running for several hours now and the driftfile that was created has the following contents: 0.000 0 That does not seem right. Does anyone know what I am doing wrong? Thanks in advance. -- Glenn Johnson gljohns@bellsouth.net To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message __ Chris Fedde <cfedde@fedde.littleton.co.us> 303 773 9134 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199908052036.OAA94406>