Michael Marsh wrote:
> I run fedora as a samba server and have just recovered from my second
> hard disk crash in only 8 months. Is there a way that i can monitor
> and fix hard disk health between disasters?
Other people have mentioned smartd.
Then there's the physical questions. How clean is the power in your
location? Do you get power surges or spikes? Are there any industrial
machines in the vicinity?
If so, buy a UPS.
How hot do the drives get? There are two main things to consider: how
hot does the server room get in summer? Are there windows which trap
the heat from the sun? Or is it next to a boiler room?
Then is there adequate cooling inside the server? Partly that's a
question of having enough fans: consider a front case fan. Partly that's
having cables tidied out the way and having sensible airflow around the
hard drives.
How much dust is there in your machine? (Don't laugh: I've just had a
hard drive die completely because the office it was in gathered dust,
the PCs sit on the floor, the cleaners weren't up to much, and the
office got really hot this summer.
What else generates heat in the PC? If it's just doing Samba, consider
a Via C3 CPU with on-motherboard graphics. All your PC will be doing is
taking files off the hard drive and squirting them onto the network, and
occasionally taking files off the network and storing them on the hard
drive. Network performance and hard drive performance are the only
major considerations.
Hope this helps,
James.
--
E-mail address: james | "!" sez I. And "?". After a few speechless seconds
@westexe.demon.co.uk | I come out with "%^&*". Unless I come up with
| something plausible soon I'm going to run out of
| special characters. -- Ben at lspace.org
