Subject: Re: wd.c patch (was Re: NetBSD 3.0_BETA & RAIDframe problems)
To: Manuel Bouyer <bouyer@antioche.lip6.fr>
From: Timo Schoeler <timo.schoeler@macfinity.net>
List: current-users
Date: 06/21/2005 18:57:20
thus Manuel Bouyer spake:
> On Mon, Jun 20, 2005 at 10:10:37PM +0200, Urban Boquist wrote:
> 
>>Timo> please add this quirk for another this, the Seagate ST3160023A
>>Timo> (160GByte, 7.200rpm, 8MByte Cache).
>>
>>Timo> before applying the patch the machine crashed when builing
>>Timo> RAIDframe, with patch applied (i did it for 2.0.2-RELEASE, too)
>>Timo> it runs fine.
>>
>>Are you sure that it really is the drive that is the problem? Maybe
>>that quirk is just masquerading some other problem with your
>>system. Did you try the drive in some other machine?
>>
>>I'm asking because I have been running a "ST3160023AS" for a long time
>>without any problems at all. Yes, I know it's not identical, but
>>still... ;-)
> 
> 
> It depends on how the drive is accessed. You have to access sector 262144
> exactly. If your drive is partitionned in such a way that sector 262144
> in not the first sector of a fragment, I suspect that you won't see the
> problem. A dd with the default block size on the raw device will make it
> show up.

yip, and as the subject describes one of the ways you'll run into this
guaranteed is to build a RAIDframe. my machine crashed when doing this.
not earlier.

regards,

timo