Re: [tlug] buffer i/o error on usb dev

[Curt Sampson <cjs@example.com>]

On Tue, 8 May 2007, Sigurd Urdahl wrote:

> I'd start by running "badblocks" on it. At least som mk*fs'es support 
> badblocklisting so that you can make sure the new filesystem avoids any 
> blocks that are bad.

If you're seeing bad blocks on a modern drive, it's already failed. For
years now, drives have done 'behind-the-scenes' bad block re-assignment
that the OS doesn't see unless a) you're reading a block that can't
be recovered (where the error will disappear next time you write the
block), or b) your're examining the drive with SMART.

If you're willing to wipe the drive, try "dd if=/dev/zero of=/dev/sda
bs=1m" (this will take some time) and see what happens. If you get any
errors, you've got a problem.

The problem could be controller-related; try pulling the drive from the
case and plug it directly into a system (and don't continue to use the
case power supply, either--make sure it's on a system with plenty of
spare power supply capacity.) You'll then also be able to use smarttools
to query the drive's error log, do self tests, and so on.

cjs
--
Curt Sampson       <cjs@example.com>        +81 90 7737 2974

Mobile sites and software consulting: http://www.starling-software.com