Archived posting to the Leica Users Group, 2009/07/31
[Author Prev] [Author Next] [Thread Prev] [Thread Next] [Author Index] [Topic Index] [Home] [Search]2009-07-31-16:46:00 Nathan Wajsman: > Do you mean both discs failed at the same time (I am assuming it was a > RAID 1 configuration)? Of course it can happen, but the probability is > tiny compared with the probability of one disc going bad. Yeah. Some systemic failure, involving some component other than a disk, is always possible, and it's possible that such a failure might scribble garbage on more than one disk, but most often it seems as if one drive goes at a time. I mean, jeez, I hope he was using a redundant configuration rather than some stripe arrangement without redundancy. Of course, if you stripe a couple of drives together to get the space and speed of both, you're reducing reliability. But nobody who's concerned with safe archiving would do it that way. In my beloved ReadyNAS NV, I've been through two generations of disks (replacing dead ones, and eventually migrating to larger ones, all while maintaining the integrity of the data). One secret is: always have at least one replacement drive on hand, so you can replace a failed drive immediately, and let the array rebuild including the new drive -- because sometimes, if your drives are the same model (and especially production batch), failures might cluster around the same number of hours; but you *usually* get a window for rolling in a replacement. Oh, another thing, which I forgot to mention: if you use drives anywhere near all the time, and have an enclosure with proper cooling, set things up so they never stop spinning. Spin-ups seem to kill drives faster than just keeping them rolling. -Jeff