Modern disk drives contain their own intelligence - SMART - the Self-Monitoring, Analysis and Reporting Technology. Beside a permanent logging of measured values and errors during normal activity, it has also procedures to selftest the device at command. Anyone who takes advantage of these smart features and let the smartmontools daemon monitor the device permanent and carefully, will in many cases get predictive warnings before it fails.

Smartmontools (http://smartmontools.sourceforge.net/) report the measured values, errors and selftests results, that were stored in the device, trigger selftests on a regular base as configured by the user and send warning emails, when values become critical.

On this way you can't ban the worst case of sudden crashes, but even then it is better to detect the damage instantly than later. Otherwise the amount of lost data can get bigger and bigger, since the next complete filesystem check is run. As you see, there are many good reasons to use smartmontools!

I wrote a tutorial for smartmontools beginners in Dec 2004 for the Linux Magazine, with an overview about the features and concrete instructions. Its target audience are home computer users.

Whereas my following German article published in Oct 2005 in Linux-Magazin addresses IT-Professionals and focus smartmontools use on a large number of servers in data centers or ISP companies. Here I gave a short introduction of Munin, a software for central server monitoring.

Today I found a nice tutorial for Ubuntu, which is also written in German.