Date: Mon, 31 Jan 2000 15:39:15 +1100 From: Peter Jeremy <peter.jeremy@alcatel.com.au> To: Igor Roshchin <igor@physics.uiuc.edu> Cc: security@FreeBSD.ORG Subject: Re: ntpd configuration and strange time "jumps" Message-ID: <00Jan31.153917est.115231@border.alcanet.com.au> In-Reply-To: <200001310018.SAA07833@alecto.physics.uiuc.edu>; from igor@physics.uiuc.edu on Mon, Jan 31, 2000 at 11:18:59AM %2B1100 References: <200001310018.SAA07833@alecto.physics.uiuc.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
On 2000-Jan-31 11:18:59 +1100, Igor Roshchin <igor@physics.uiuc.edu> wrote: >1. I've noticed that there were a few rather strange time steps: >(it's the first time I see it changing back and force without any >visible reason) >Jan 28 19:58:45 <ntp.notice> myhost xntpd[144]: time reset (step) -0.244614 s >Jan 28 21:04:09 <ntp.notice> myhost xntpd[144]: time reset (step) 0.353294 s ... NTP assumes that the path delays between your daemon and it's peers (or servers) are symmetric - it halves the RTT to determine the peer delay. This is a flaw in the protocol, but I don't believe there's any way around it. If all your peers share a common bottleneck, which has a large traffic asymmetry, it is possible for NTP to see this asymmetry as a peer offset and adjust the local time to suit. When the traffic asymmetry goes away, xntpd will then skew the time the other way. Many years ago, I added a fudge to detect (or try to) and ignore this situation. At the time Dave Mills wasn't interested in the patches. Peter To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-security" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?00Jan31.153917est.115231>