Date: Mon, 27 Apr 2015 10:28:35 -0500 From: Jim Thompson <jim@netgate.com> To: Christian Weisgerber <naddy@mips.inka.de> Cc: freebsd-hackers@freebsd.org Subject: Re: System clock always unsynced Message-ID: <0E68BBCF-5E21-4189-9358-F7BA3F5BECDC@netgate.com> In-Reply-To: <slrnmjsigv.8j.naddy@lorvorc.mips.inka.de> References: <slrnmjsigv.8j.naddy@lorvorc.mips.inka.de>
next in thread | previous in thread | raw e-mail | index | archive | help
You can prevent stepping of the clock with the -x option of ntpd. The = issue here is that "can take a long time to converge to an acceptable = offset, about 2,000 s for each second the clock is outside the = acceptable range. During this interval the local clock will not be = consistent with any other network clock and the system cannot be used = for distributed applications that require correctly synchronized network = time.=E2=80=9D http://doc.ntp.org/4.1.0/ntpd.htm Running openntpd is a poor substitute for a real ntpd. Answering, "What *does* an ntpd daemon need to do to sync the clock?=E2=80= =9D likely requires a thesis length posting, and it likely suffices to = state that you=E2=80=99re not the first to discover that openntpd is a = poor implementation of an NTP daemon. Jim > On Apr 27, 2015, at 9:39 AM, Christian Weisgerber <naddy@mips.inka.de> = wrote: >=20 > I run OpenNTPD, from ports/net/openntpd, and I've noticed that after > each reboot, the initial system time is further off. (This is quite > noticeable with OpenNTPD, since by default it does *not* jump the > clock on startup like the base ntpd does.) It's as if the RTC was > never synchronized to the system clock. >=20 > Some digging in sys/kern/kern_ntptime.c shows indeed that the RTC is > only synced if STA_UNSYNC is not set, and dumping the value of the > timex struct... >=20 > offset: 0 > freq: 2730304 > maxerror: 84860000 > esterror: 500000 > status: UNSYNC > constant: 0 > precision: 0 > tolerance: 32500000 > state: ERROR >=20 > ... reveals that the clock remains permanently unsynced. Clearly, > OpenNTPD, which uses adjtime(2) to correct offsets and ntp_adjtime(2) > with MOD_FREQUENCY to correct the frequency, doesn't handle this > quite right. >=20 > What *does* an ntpd daemon need to do to sync the clock? >=20 > The ntp_adjtime(2) man page documents struct timex in detail, but > is very vague on what all of this means. >=20 > --=20 > Christian "naddy" Weisgerber = naddy@mips.inka.de > _______________________________________________ > freebsd-hackers@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-hackers > To unsubscribe, send any mail to = "freebsd-hackers-unsubscribe@freebsd.org"
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?0E68BBCF-5E21-4189-9358-F7BA3F5BECDC>