On Fri, Jun 25, 2010 at 12:50 PM, Joe Tinney <jtin...@lastar.com> wrote: > It seems I need to educate myself some more on how hard drives and > their electronics flag bad sectors independent of the file and operating > system!
Modern hard drives reserve some spare blocks, which are not visible to the host. When the onboard electronics detect a bad block, the data is automatically and transparently relocated to a spare block. Ideally, bad blocks are detected and relocated on write. Slightly less ideally, you read from a block, the read succeeds but the drive detects the block is marginal, and the drive relocates the block then. Much less ideally, you try to read, the drive retires a few times but still can't do it, and so the drive returns an error to the host. The drive will keep retrying that read every time you ask it. Sometimes it will succeed eventually, and the drive can relocate it then. Otherwise, if you write new data to that block, the drive will relocate it then, since you've overwritten what was there anyway. Modern hard drives also store redundant data and employ ECC to recover from the occasional bit error. This happens all the time, even on a brand new drive. SpinRite's core read/test/rewrite algorithm will trigger the above. But so will "badblock -s -v -n" from a Linux boot CD. SpinRite can also repeatedly try to read a bad block on a failed disk. So will "dd_rescue" and "dd_rhelp". Etc. SpinRite was special back when when MS-DOS was popular, because MS-DOS had all the intelligence of a not-so-bright rock. These days, SpinRite is just another overpriced utility. -- Ben ~ Finally, powerful endpoint security that ISN'T a resource hog! ~ ~ <http://www.sunbeltsoftware.com/Business/VIPRE-Enterprise/> ~
