Manolis Kiagias wrote:
From what I can see, the 32bit timer part used in TCP rolls over after
497 days. That is why they can not detect usage over 497 days.
Netcraft says "Additionally HP-UX, Linux, NetApp NetCache, Solaris and
recent releases of FreeBSD cycle back to zero after 497 days,"
Evren Yurtesen wrote:
I see that netcraft uptime doesnt function because the kernel timer is
set to 1000 by default in freebsd 6.x
So perhaps setting kern.hz=100 fixes the problem? What about things
getting wrapped in 497 days? How are there some FreeBSD hosts which
have 1600+days uptime? how do they manage it? I am just trying to
figure out how to do the same in my systems :)
Yes, I believe the problem is the kernel timer. Older versions of
FreeBSD had a 100Hz timer. A Debian server of mine started showing
uptime again in Netcraft when I recompiled the kernel with 100Hz timer
As for the 497 days problem, I think this is Linux specific. I don't
know however if it will exist in FreeBSD 6.X with timer set to 100Hz.
Managing to get 1600 days uptime requires very solid hardware, possibly
with redundant (maybe hot swap) components and stable power. A UPS is
necessary and probably a generator, as during this time there is going
to be a blackout that will probably outlast your battery. FreeBSD is
actually the easiest (and less costly) part of the setup: It will,
simply, run all this time,
Well, I could do over 500 days with a small ups and a 10 year old
machine. There is almost no blackouts around here. When there is, it doesnt
no sweat :)
last more than 1-3 minutes once in 2-3 years.
Anyway, we will see if netcraft will show 0 days after 497 but it would
be a shame really :)
email@example.com mailing list
To unsubscribe, send any mail to "[EMAIL PROTECTED]"