Date: Fri, 30 Dec 2005 11:30:02 +0100 From: Koen Martens <fbsd@metro.cx> To: "Julian H. Stacey" <jhs@flat.berklix.net> Cc: =?UTF-8?B?RGVyZWsgS3VsacWEc2tp?= <takeda@takeda.tk>, stable@freebsd.org Subject: Re: "runtime went backwards" message in logs Message-ID: <43B50C2A.9030405@metro.cx> In-Reply-To: <200512300039.jBU0dtYd051657@fire.jhs.private> References: <200512300039.jBU0dtYd051657@fire.jhs.private>
next in thread | previous in thread | raw e-mail | index | archive | help
Julian H. Stacey wrote: >>PS. Not sure if this have anything to do with it (the message mentions CPU >>time, not the clock), but I'm running a ntp daemon, to synchronize time... >> >> > >Highly likely it has a Lot to do with it :-) Maybe the master time >server had `date` run manually, or otherwise shifted, or came back >on net after an outage, & the systems noticed drifted time & corrected etc. > man ntpd > man ntp.conf > etc :-) > > AFAIK, under normal circumstances ntpd should never reset the current time the hard way. Ntpd slows down (slews) or increases (steps) the speed of your clock, such that it will catch up or fall behind enough to become in line with the ntp servers it is listening to (since resetting the time the hard way is usually not a good idea; think makefiles, cronjobs that could run twice, etc..). You could check out the -x flag to nptd, to prevent setting the clock the hard way all together (since network congestion and stuff like that might result in ntpd setting the clock the hard way). Best, Koen
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?43B50C2A.9030405>