Solution 1:

There is a good summary here.

The reasons include:

  • Head crash, where the read-and-write head touches a platter.
  • Faulty air filter, allowing dust to land on the platter
  • Actual failure of the controller electronics.
  • Mechanical failure or worn parts

So being knocked during shipping is a possible reason for disk failures, and prolonged use in intense conditions will cause wear to mechanic parts and dust getting in can cause a failure too.

The most likely cause of a disk failure if the disk is not defective in some way is just getting worn.

Solution 2:

Some causes for harddrive failure include:

  • poor handling, e.g. impact, static electricity
  • vibration
  • high accelration
  • overvoltage/power surge
  • dust or other particular matter, e.g. metal filings
  • rapid changes in temperature - I think EMC says more than 10C per hour temperature change - but check with your vendor
  • plain old high temperature e.g. ambeint temp over 45degrees C or there abouts - check with - your vendor
  • in the old days buggy drivers, but nowadays I believe the firmware/physical interlocks tends to prevent that - although I believe there has occasionally been buggy firmware released that has caused drive failure problems.
  • faulty components/other manufacturing defect

You might find this paper interesting: http://labs.google.com/papers/disk_failures.pdf Basically some Google Sysadmin/SRE did some research into drive failures over a large population of drives (i.e. the google data centres).

Solution 3:

Bugged firmware: yes, like velociraptors. being bitten by that stupid thing at the moment. Have to change firmware on a dozen drives.