On Fri, 01 Apr 2005 19:58:32 PST, Gary Funck said:
>
> We'd like to test a disk imaging/copy device/procedure to see how it
> fares if the source drive has one/more bad block errors. Is there a way,
> by communicating with the hard drive, presumably at a very low level,
> to convince the drive that it should mark a sector as having a
> permannent error, and to _not_ have it try and spare this particular
> sector to a good sector on an alternate track? Basically, we'd like
> to insert what appears to be several (physical) bad blocks for the
> purposes of testing.

Get yourself a cheap drive, and communicate with it with a hammer while
it's running. (Seriously.. it's probably the cheapest way to do it...)

Mst drives can only spare a sector on a *write*, because it needs data to write
into the spared block. Also, it's usually silent about doing the sparing,
unless you use SMART or similar to ask it what's going on. If a block was OK
when written but subsequently goes bad, the drive really has no choice but to
report that the block is bad (it certainly can't return the data as it's bad,
and returning a block of zeros is a bad idea as well).

So you really need a way to create a bad block without actually sending
anything down the cable to the drive - thus the hammer...

[ reply ]