Google Disk Failure Analysis Offers Drive Crash Insights

Over the weekend, Google released an interesting study of hard drive failures on its network. After all, the company has about 100,000 of the things lying around. And sooner or later, they crash.

Google did a lot of work analyzing those crashes, and there's good news and bad news about that: The good news is the company released all the information about the crashing to the public. The bad news: It's in the form of a 13-page document written for scientists and researchers, and there's no real digest available for the average reader. (Here's the full document, in PDF form.)

Well, I guess that's why I'm here, right? Here's what the study had to say about hard drives, in a nutshell.

• Drives fail all the time, even brand new ones. (Drives one year old or younger had an annual failure rate of about 6 percent. That means about 1 of 17 new PCs will have a dead drive within a year.)
• Once drives hit two years old, the failure rate really starts to ratchet up.
• Extremely high temperatures can lead to excessive failure rates, but only for older drives. In younger drives, the survey found that excessively cold temperatures can be a bigger problem. Overall, the old tenet that high temperature is a critical factor in drive failure can't be verified.
  
• SMART data (Self-Monitoring, Analysis and Reporting Technology) can be useful in determining whether a drive is going to fail. Up to 30 percent of drives that indicated SMART errors eventually failed, and the probability of crash gets worse and worse the longer an "erroring" drive is left in service. That said, many drives exhibit SMART errors at some point in their lives.
• Google didn't break out drive failures by brand or capacity. Boo!
  

In the end, the study found that SMART data was the best at determining whether a drive would fail soon. How do you monitor SMART data on your computer? There are many inexpensive tools available, but S.M.A.R.T. Monitor 2.1 is free. Get it here (you'll have to scroll about 2/3 of the way down the page). Give it a whirl, but remember, about half of all drives that fail don't display any problems at all before they croak. Back up, back up, back up!