[4427] in linux-scsi channel archive
Re: Bad blocks: How many?
daemon@ATHENA.MIT.EDU (Dr. Michael Weller)
Tue Jul 21 14:06:29 1998
Date: Tue, 21 Jul 1998 19:17:49 +0200 (MESZ)
From: "Dr. Michael Weller" <eowmob@exp-math.uni-essen.de>
To: Karl-Heinz Herrmann <k.-h.herrmann@fz-juelich.de>
Cc: linux-scsi@vger.rutgers.edu
In-Reply-To: <XFMail.980721153222.k.-h.herrmann@fz-juelich.de>
On Tue, 21 Jul 1998, Karl-Heinz Herrmann wrote:
> I have a Samsung Ultra (non wide) SCSI drive (2GB) and started with
> some hundred bad blocks (thats sectors, not tracks. scsiinfo tells
> about whole bad tracks by adding a -1 as sector number for that track
> and normal bad blocks are adressed as track:sector).
> The grown list was clean.
> After reinstalling linux I had -- much to my surprise -- 34 grown
> defects on that drive. ARRE is off and AWRE in ON (manufacturer
> setting), which meens only errors while writing are relocated
> silently by the drive.
>
> I had absolutley no warning or error message from the linux kernel.
Well, actually you can control things like that with the other buttons in
that mode page. Now, things are complex though as due to buffers the
driver might already confirmed sectors ok although they can't be written
later. There are various things like reporting errors on another sector
(or asynchronously) although that sector was ok (or even the actual
failing sector could be written, but at another place). It is also
possible that the drive reports something like: sector x written ok but
reallocated. I'm not sure if the kernel driver would actually be able to
handle such things. A problem is of course also to test such things, as
you don't have a drive usually which reproducible produces such errors.
> Obviously while writing to the disk it found some bad blocks and
> instead of complaining to the linux kernel the drive just marked it
> bad and relocated the block.
> I suppose if a read error occurs the error will be reported, since
> the ARRE Flag is off.
Yes.
>
> BTW both my drives (the Samsung and a Quantum Saturn Wide 2GB) have
> some hundred bad blocks listed in the manufacturer list. The quantum
> has zero grown defects and the samsung now 34.
Well, with regards to my last post to the list, this is at least some
positive feedback showing the defect display code works (and at least
sometimes reallocated sectors show up in the defect list).
Michael.
--
Michael Weller: eowmob@exp-math.uni-essen.de, eowmob@ms.exp-math.uni-essen.de,
or even mat42b@spi.power.uni-essen.de. If you encounter an eowmob account on
any machine in the net, it's very likely it's me.
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.rutgers.edu
