badblocks

[Michael Hipp]

The Small Box Admin wrote:
> I've found that on modern drives, once you start getting bad blocks
> the end is near for the drive.  I've even read that smartctl isn't a
> good predictor of failure.  There was a Google study on drive failure
> that reported this, though at the time I read it I questioned some of
> the conclusions.
> 
> It's not like the good old days with MFM and RLL drives where you
> could do a low level format if you started getting soft or even hard
> errors.  The low level format would remap the bad blocks and you'd be
> back in business.
> 
> As for a suggestion, if you want to proceed with the suspect drive, I
> would get the manufacturers utility do do a low level fromat or test.
> That is if the manufacturer has the utilities, you didn't indicate
> drive manufacturer, type or model.

Thanks. These drives all come out of Windows boxes where MSW couldn't seem to 
read part of the disk. In those cases it's usually cheaper to just swap the 
drive and reload rather than try to determine if Windows is lying or not. But 
in a fair number of cases the drive turns out to be ok.

I've used Seagate's Seatools quite a bit but it requires a reboot every time 
and the hokey UI often hangs on me just after a long test. So I don't use it 
much anymore. (The version of it that runs on Windows is actually more stable 
than the bootable CD.)

Anyway, I generally trust Linux tools to be more thorough so I am wanting to 
use them. What I really want is to be able to do something like this:

   # isthisdiskgood /dev/sda
   chug ... chug ... chug
   Yes

Michael