OT: Re: [CLUE-Admin] Website Updates

[Jed S. Baer]

On Sun, 07 Sep 2003 20:02:56 -0600
David Anselmi <anselmi at americanisp.net> wrote:

> Jed S. Baer wrote:
> [...]
> > Nope. But what I do know, unfortunately, is that my computer is
> > approaching the unreliable stage -- bad hardware clock.
> [...]
> > BTW, the other symptom was watching XDaliClock jump back/forth by an
> > hour at something like 1 minute intervals -- very strange.
> 
> Doesn't that clock show system time?  And isn't Linux system time kept 
> by the kernel?  I guess you might have meant CPU clock rather than RTC 
> (which I assumed).  Oh well, glad you got it fixed.
> 
> I hate hardware, so hard to figure out what it's doing.

Yeah, so do I, anymore. Too many things to know about these days. But,
AFAIK, there are a couple of "clocks". One is the "time-of-day" clock,
which is part of the BIOS. There's also the RTC, which seems to be a
different thing. But yeah, system time is kept by the kernel.

The failure mode which got my attention was the failure (and hang) at
shutdown for the "synching hardware clock to system time" step. That wound
up being a poweroff event, with the subsequent fsck of all partitions at
boot. Yeah, I should upgrade to a journaled filesystem.

The clock issue, though, is a bit muddied in my mind by the nntp docs,
which talk a lot about some wierd terminology to speed up / slow down the
clock, to get it to where it keeps proper time. Now, this would almost
have to be a hardware issue, because I've noticed that on my system (or at
least the old one) the "drift" (as evidenced by how far off the clock was
after the system being down for several hourse) got corrected after I'd
run nntp for a while. Or maybe I'm hallucinating.

jed
-- 
... it is poor civic hygiene to install technologies that could someday
facilitate a police state. -- Bruce Schneier