Re: Weakest point of a server?
Jason Lim wrote:
But how about the motherboards themselves? Is it often for something on
the motherboard to fail, after 3-4 years continuous operation without
Normally, I'd say no on this point, particularly if the server is
continuously running. This month's issue of IEEE Spectrum, however, had
an article on failing capacitors on computer motherboards. Apparantly,
somewhere in Asia, a former employee of a capacitor plant stole his
company's formula for the electrolyte, went to work at a new company,
and didn't realize that he had missed part of the process. The new
company's caps have started leaking and failing.
I don't remember all the details offhand, and the article was sketchy on
Or is there some other part(s) we should look out for instead... would the
CPU itself die after 3 years continuous operation? Or maybe RAM? Or even
the LAN cards?
Had the occational LAN card go bad. In many applications, we can
tolerate a quick swap, so we keep a few 3c509 spares around. If it's a
critical system, might want to consider keeping a spare card plugged in.
We have had power supplies die fairly often. Usually traced back to a
dead power supply fan, but sometimes it's just a dead supply.
We keep the systems at between 18-22 degrees celcius (tending towards the
lower end) as we've heard/read somewhere that for every degree drop in
temperature, hardware lifetime is extended by X number of years. Not sure
if that is still true?
You extend the mean time before failure (MTBF). So although you can't
say "I've lowered the temperature from 25C to 22C, so my machine will
last X more years", you can say "there was a 1/x chance of having this
piece of hardware die at 25C, I've lowered the temp to 22C, so now
there's only a 1/y chance".
ETN Systems Inc.
2125 1st Ave East
Hibbing MN 55746