hddtemp: does exactly what it says on the tin.

Overly hot hard disks are bad. I think we all know that. It doesn't stop us worrying about our data, but a problem can't be fixed unless we know it's a problem. Opening cases and touching hard drive to see if they are hot is the most low tech of approaches, but it's cheap and completely cross platform.

There is a wide range of hardware solutions, often tied to a fan controller and usually taking up a drive bay. Prices range from less than a beer to several weddings, and whilst many of the cheapest are capable of sounding alarms they rely on someone being there to hear them.

But why spend money on what's already built in? Most recent hard drives have built in temperature sensors, and the easyest way to extract them is hddtemp.

Exactly what it says on the tin, and the -n switch is handy for feeding the output in to a script that does something useful, but that's already taken care of. A quick edit of /etc/default/hddtemp and disk temperatures can be sent to the syslog.

Of course extracting useful warnings from logs is a timely fashion is a problem of it's own. Anyone know of a demon that watches logs for problems and then e-mails an administrator?

/me wonders whether a text-to-speech announcements of warnings is appropriate for a media box.