Re: [tlug] Re: [OT] Intel core duo errata

[<burlingk@example.com>]

On Tue, 10 Jul 2007 09:19:54 +0900,
Darren Cook <darren@example.com> wrote
in <4692D0AA.8080309@example.com>


>
> but linux only uses the low bits of the time. So if a machine has been
> up 500 days it shows (to the external world, via the TCP/IP headers)
as
> only having been up for 3 days. Linux 2.6 uses even fewer bits so
wraps
> after 49 days.
> 

So, basically because they use technological smoke and mirrors
to try to grab information that the servers aren't being told
to give them, they can't get a good read.  That makes sense.

What would make more sense is if they did use a server
deamon for the process and let people who want to participate
submit their machines for monitoring.  That seems a lot more
friendly to me, and a lot more accurate. ^^;

I know I have logged into machines via ssh, and used the local
uptime command (local being a realative term), and gotten times
well beyond 500 days before.  :P  I have a friend who loves to
use Debian, and doesn't reboot often. ^^;