 |
 |
Dear list, James is right about the way that drives re-map sectors. In addition, different drives do this in different ways. A drive that is intended for use in a RAID array will timeout after a few seconds when accessing a problematic sector, to allow the RAID controller to look for the data elsewhere in the array, and possibly initiate some kind of sector remapping in the background. On the other hand, a drive that is designed for use outside a redundant array will keep its host system waiting for a long time while it attempts to access/recover problematic sectors before finally giving up. In a RAID array, this kind of behaviour would cause the RAID controller to mark the disk as bad. On some disks, this behaviour can be switched on or off, on others not. We have had just this situation a while back, where a hardware supplier insisted that cheap drives would be fine in a RAID5 setup, but we had no end of problems until we switched to slightly more expensive ones that were designed for RAID use. See for example: http://en.wikipedia.org/wiki/Time-Limited_Error_Recovery http://download.intel.com/support/motherboards/server/sb/enterprise_class_versus_desktop_class_hard_drives_.pdf (in particular section 1.1.4.1) Regards, Peter. On Wed, 27 Oct 2010, James Holton wrote: > I recently bought 40 drives of the Western Digital 2 TB "Caviar Green" > variety (WD20 EARS). So far, 8 of them are "bad" inasmuch as they have > unrecoverable bad sectors, or simply won't mount. In several cases, this > began as the drive having no bad sectors, but just getting very very slow. A > "normal" one of these drives will easily get 25 MB/s or more sustained > transfer rates, but the first sign of trouble is that this rate drops to ~1 > MB/s or less. Eventually, you start seeing bad sectors, and these continue > to increase until the drive becomes unusable. > > Just a guess, but what I think is going on is that most drives these days > have an internal reserved area for re-mapping bad sectors. I imagine this > re-mapping slows things down, but no bad sectors are reported. This could > all just be a particle of dust floating around inside, and once you make one > scratch, that makes more bits of dust. etc. > > Yes, there are more expensive drives that are "server certified" and cost > twice as much. Note that at 25 MB/s it takes about a day to simply > read-check all 2e12 bytes on the drive, and about a week to do a full > "badblocks" run. This may be why "better" drives are more expensive. A > colleague of mine here bought a few of those, but one of them is now in my > "BAD" pile. Transfer rate suddenly slowed down to ~2 MB/s. > > Sigh. Technology just doesn't work. > > -James Holton > MAD Scientist > > On 10/27/2010 1:43 AM, Frank von Delft wrote: >> I've been told by my (frighteningly geek-competent) colleague that the >> platters are identical, but the cheaper ones are those at the bottom of the >> Quality Control pile, which are therefore spun more slowly, and don't get >> any claims of reliability. >> >> (Have you checked the disk rotation speed? I imagine the Dell ones go much >> faster.) >> >> phx >> >> >> On 27/10/2010 02:52, Edward A. Berry wrote: >>> Another question about computer hardware- If I configure a computer at the >>> Dell site, it costs about $700 to add a 2TB SATA drive. >>> On amazon.com or Staples or such, a 2TB drive costs ~$110. to $200 >>> depending on brand. >>> >>> Are the Dell-installed drives much faster, or more reliable, or have >>> a better warranty? After all, RAID is supposed to stand for redundant >>> array of inexpensive disks, and we could afford a lot more redundancy >>> at the Amazon.com price. >>> >>> And, are there any brands or models that should be avoided due to known >>> reliability issues? >>> >>> Thanks, >>> eab > -- Peter Keller Tel.: +44 (0)1223 353033 Global Phasing Ltd., Fax.: +44 (0)1223 366889 Sheraton House, Castle Park, Cambridge CB3 0AX United Kingdom