> -----Original Message----- > From: centos-bounces at centos.org [mailto:centos-bounces at centos.org] On > Behalf Of Peter Brady > Sent: Wednesday, April 13, 2011 18:39 > To: CentOS mailing list > Subject: Re: [CentOS] CentOs 5.6 and Time Sync > > On 14/04/11 7:42 AM, Mailing List wrote: > > remote refid st t when poll reach delay > offset > > jitter > > > ======================================================================= > ======= > > > > bindcat.fhsu.ed 132.163.4.101 2 u 1015 1024 377 49.987 - > 15082. > > 6919.88 > > 216.45.57.38 108.71.253.18 2 u 998 1024 377 83.112 - > 15139. > > 6900.14 > > javanese.kjsl.c 69.36.224.15 2 u 1 1024 377 109.083 - > 29233. > > 7285.83 > > *LOCAL(0) .LOCL. 10 l 13 64 377 0.000 > 0.000 > > 0.001 > > <snip> > > Glad you've got a fix but you should keep an eye on it. > > If you look at the output for ntpq you have three stratum 2 servers > which differ by ~15s from both you and each other. Stratum 2 servers > should be a lot closer than 15s given that they are only one link > removed from some form of atomic clock. > I think you may be comparing a couple of rotting apples to one that is just now ripening. when offset 1015 15082 998 15139 1 29233 i.e., the samples that are 17 seconds apart in the taking are .057S apart in value, and trending longer.