Re: harddisk or kernel problem?

[Nico Schottelius <nico-kernel@schottelius.org>]

Bruce Allen [Wed, Feb 18, 2004 at 10:55:05AM -0600]:
> [...] 
> FWIW, after reading this thread, I've slightly modified smartmontools so
> that when smartctl prints the error log (-l error) it ALSO prints the LBA
> at which a READ or WRITE command failed.

Thank you very much for patching smartctl and explaining howto calulate
the LBA from those values.

So at least my hd crash had some sense.

> [Note that this is a 28-bit sector address.  If a disk is larger than 2^37
> Bytes = 137 GB, then some LBAs can't be written in 28 bits, in which case
> there won't be a summary error log entry.  If the disk is smaller than
> 2^37 Bytes then the failing LBA address should always be logged.]

Why are we bound to a 28 bit value?
As there are currently more and more disks out there with >=260GB, I
think this will be an issue very soon.

Have a nice day,

Nico, currenty working with knoppix, a wonderful live Linux
                           (www.knopper.net/knoppix)