[mythtv-users] Booohooohoooo......

Yeechang Lee ylee at pobox.com
Sun Jul 13 12:02:50 UTC 2008


David Brodbeck <gull at gull.us> says:
> At work I've had to RMA three Western Digital drives in the past
> year.  Two of them were replaced by "recertified" drives, but they
> sent me a brand new one as a replacement for the third.  Amusingly
> enough that last drive was over two years old.

My 19 months-old, all Western Digital-drives MythTV array (the second
one mentioned in the signature) has had three of its 16 drives fail in
the past three months, with two failures overlapping. In each case,
SMART errors showed up ahead of time via smartd. I've replaced each
through WD's advanced-RMA program, and (thanks to RAID 6) my array was
never out of action even during the overlap period. How do I tell the
difference between recertified and new drives? Something in the serial
number?

> I'm spinning about 35 drives, all Western Digital, but that's really
> not a big enough sample size when failures are such rare events.

Agreed. A Slashdot posting on disk drives always degenerates into
endless litanies of "I'll never buy ______ drives again, because they
always fail." No two poster in a row ever lists the same vendor in
______, of course. Heck, these sorts of idiocies have appeared on
mythtv-users. As far as I know the infamous IBM Deskstar/Deathstar
case of several years back is the last time one could legitimately
blackball a particular vendor's offerings for a while.

That said, I will say that WD drives account for 80% of the hard
drives that have ever failed on me. To wit:

* IBM Deskstar 40GB (mid-2001) that failed after five solid years of
  24x7 operation as my primary server drive. Yes, it was one of those
  Deathstars. I was able to copy the disk image to another drive in
  time, then for simplicity's sake purchased a (now Hitachi) exact
  replacement and copied the disk image back.
* Western Digital 100GB (mid-2001) that failed after three solid years
  of 24x7 operation in my old TiVo (although it had begun dying the
  year before, as evinced by occasional playback skips). Heat is
  almost certainly responsible here; a combination of two 7200rpm
  drives inserted into a case that was designed for 5400rpm or slower
  drives and, more seriously, an unventilated apartment that hit 110
  degrees in the summers.
* The aforementioned three Western Digital 500GB server-class
  drives. Meanwhile, no failures among the three Seagate 750GB drives
  that comprise the first-listed array below although it runs in a
  system that has an incredibly hot-running processor while the
  16-disks array runs on a processor that runs more than 10 degrees
  cooler. Again, 24x7 operation for the arrays; both came on line in
  December 2006.

I am not going to stop buying Western Digital drives, or only buy
Seagate drives, based on these data points. I do find them
interesting, though.

-- 
Frontend:		P4 3.0GHz, 1.5TB software RAID 5 array
Backend:		Quad-core Xeon 1.6GHz, 6.6TB sw RAID 6
Video inputs:		Four high-definition over FireWire/OTA
Accessories:		47" 1080p LCD, 5.1 digital, and MX-600


More information about the mythtv-users mailing list