For smaller drives, there is lower probability of a URE during rebuild, but the probability is not zero. It's very embarassing to explain to a client that his "infallible" RAID system has indeed collapsed.
It's absurd to have massive amounts of stored data dependent on a 100% recovery. The only way to avoid increasing failures of this type as storage needs expand is to design storage architectures so that a few lost bits do not translate into global failures.
BTW: from the ratings of the original article, it appears to me that IT-ers have quite a case of denial. Thanks for rocking the boat!
Discussion on:
Message 9 of 1
The best of ZDNet, delivered
ZDNet Newsletters
Get the best of ZDNet delivered straight to your inbox
