home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.protocols.time.ntp
- Path: sparky!uunet!charon.amdahl.com!pacbell.com!tandem!UB.com!quack!mrapple
- From: mrapple@quack.sac.ca.us (Nick Sayer)
- Subject: Re: xntp3, DOSYNCTODR_SUCKS
- Message-ID: <fU8Y2Uo@quack.sac.ca.us>
- Organization: The Duck Pond public unix: +1 408 249 9630, log in as 'guest'.
- References: <1992Nov9.214008.2068@elroy.jpl.nasa.gov>
- Date: 11 Nov 1992 03:34:43 UTC
- Lines: 48
-
- stevo@elroy.Jpl.Nasa.Gov (Steve Groom) writes:
-
- >In xntp3, this DOSYNCTODR_SUCKS thing has a comment that seems
- >to indicate that if I don't define this, the hardware clock doesn't get
- >updated by xntpd.
-
- No, it's that if you DON'T add DOSYNCTODR_SUCKS it DOESN'T update the
- hardware clock.
-
- > Is this something new in v3?
-
- Well, I made the patch originally in v2, but v3 came out very shortly
- after that.
-
- > Also, there is a comment
- >that says that I probably only need this if I'm not on a network.
-
- I disagree with that comment. I think it should be done on all suns
- that have _dosynctodr in the kernel. However, subordinate clocks
- can do as well by running 'ntpdate' before starting xntpd. Though
- the clock will be hosed between boot time and the time ntpdate is
- run (and ntpdate may hang momentarily if one of its sources is
- down).
-
- > It seems
- >to me that if this is indeed a problem, the machine could wake up
- >with the clock far enough out of whack that xntpd would give up, making
- >this an issue for "networked" machines as well.
-
- Exactly, only your reasoning is 180 degrees out of phase. :-)
-
- >I initially xntpd (v3) with this definition disabled. When I booted the
- >machine after being powered off over the weekend, the clock was 14
- >seconds off. This isn't great, but it sounds reasonable for a
- >battery-powered CPU clock. It looks to me like the CPU's hardware
- >clock must have been OK when the system went down. Since I got this
- >fairly reasonable behavior with DOSYNCTODR_SUCKS undefined, what is
- >the benefit of having it defined?
-
- The benefit of having it defined is that when you reboot after having
- the system up for several months, you'll find it's only maybe a few
- msec off.
-
- --
- Nick Sayer <mrapple@quack.sac.ca.us> | "What are we gonna do, Stimpy?"
- N6QQQ @ N0ARY.#NOCAL.CA.USA.NOAM | "We could get some work..."
- 37 19 49 N / 121 57 36 W | "Work?! Have you lost your MIND?!"
- +1 408 249 9630, log in as 'guest' | -- Ren & Stimpy
-