Magnetic Wobbles Cause Hard Drive Failure

An anonymous reader writes "According to this report by IT PRO, scientists working at the University of California have discovered the main reason of hard drive failure. According to researchers, some materials used in hard drives are better at damping spin precession than others. Spin precession of magnetic material effects its neighbors' polarity and this can spread and cause sections of hard drives to spontaneously change polarity and lose data. This is known as a magnetic avalanche. So next time Windows fails to start, you'll know why!"

Comments

Re:Question

[Speare]

affect

When 'effect' is used as a verb, it means 'to create.' The article writeup has the same primary-school error. It's not that hard, people.

Not "the" but "a lesser known"

[mritunjai]

This phenomenon is only one of the several ways for bit rot to creep in and make you lose data.

In bit rot, bits on HDD spontaneously change. It is generally not observable and the results are often blamed on applications and/or OS.

It is lesser known because in the current state of technology, the aplications, OS, filesystem and even RAID can't even detect the problem much less solve them. (RAID doesn't work because it can't tell which copy is right and which is wrong. It assumed what it got from disk is what it wrote to it.)

ZFS (Solaris/SUN filesystem) solves this problem by using end-to-end checksums. However, it exists for few platforms only.

Re:Which University of California?!

[kf6auf]

Here is a link to the UC Santa Cruz press release and the professor is indeed there (I'm sure you can find him). A little spiel from me: I took a class on nanomagnetism this past term and definitely learned about this effect for individual spins and for domains and it has been known for quite some time. Without reading the PRL article because I'm off campus and don't have a personal subscription ($$$ and, hey, this is /.), my guess is that the model explains the why a lot better than existing ones, and how we get from individual precessing spins to the average spin of the entire domain without brute-force computing it, which is nearly impossible. That being said, different ferromagnetic materials are very different in their interactions between spins and orbits between nearby spins and orbits and so I'm not sure without looking into it how many different ferromagnetic materials this applies too.