Old things do work.

Sergio Lenzi lenzi at k1.com.br
Mon Mar 12 13:39:46 UTC 2007


As I said some emails ago....

I have a 200 machines all freebsd network that runs 24/7 with a mtbf
of 15 years...

how I got the numbers?

well   I have an "issue"  every 3 or 4  weeks... (the machine hangs,
stops, the power supply explodes....) the machines are aways on...

if you count 200 machines and one machine crashes once every 200 days,
I will have a 200 day running time on each machine... well I have almost
once a month
problem... that is 30 times 200 days -> 6000 days for each machine...
the
machines have now 3 to 5 years old...

Now I am replacing the machines for AMD64 with mainboard ECS  that
proved
the more reliable hardware... (asus k8v is also a good choice)
(gateway is bad...) 98% of the machines are AMD, some are cyrix
It continues to run even with the cpu cooler stopped
by dust (yes!!! some machines are in rural areas, gas stations and have
no way to
offer maitenance) and I have not seen them for years... but sometimes
the machines emails me.. (when they finally
get dial up connection... and have some issue like: no space, no
swap...) it is
like sending them to mars...

You can see some pictures of the running system on FreeBSD + gnome +
windowmaker...

at http://www.k1.com.br/images/sistemas

these are old pictures, but you can have an idea.
the new versions have more enhancements....

It is all coded in FreeBSD with python + gtk + postgres...

and runs in windows too... 


Sergio


More information about the freebsd-questions mailing list