1. 12
  1.  

  2. 2

    Given the data they’re looking at there, I feel like they should emphasize the “with a NetApp-like write pattern” qualifier more – I’d guess that other filesystems/DBs/whatever-else may induce significantly different failure patterns than WAFL/ONTAP. That said, it’s still an interesting study (and consistent with my own long-held belief that the risk of wear-out produces a lot more hand-wringing than it generally merits).