Date: Sat, 10 Nov 2007 21:56:36 +0200 From: "David Naylor" <blackdragon@highveldmail.co.za> To: "Jeremy Chadwick" <koitsu@freebsd.org> Cc: freebsd-stable@freebsd.org Subject: Re: Harddisk failure causes system crash, please help Message-ID: <b53f6f940711101156l7dbb9a09re4b382ed6653f3f9@mail.gmail.com> In-Reply-To: <20071109065201.GA47328@eos.sc1.parodius.com> References: <b53f6f940711081240q7100a08djae76b560cddfed6f@mail.gmail.com> <20071108212921.GA34721@eos.sc1.parodius.com> <b53f6f940711082229l67f9a77ch497ee6270490249a@mail.gmail.com> <20071109065201.GA47328@eos.sc1.parodius.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 09/11/2007, Jeremy Chadwick <koitsu@freebsd.org> wrote:
> Okay, so it's probably that area of the disk which has some problem...
It may but I can confirm that FreeBSD is not handling it properly (see below)
> There's a free utility called HDTune which has a sector scanner which
> explicitly looks for bad sectors ("Error Scan"). I would *uncheck* the
I got it and it works well. Thank you. The first time I used it
there was a corrupt sector in approximately the area ad0e occupied
(where the crashes from reads were from). I then did a dd
if=/dev/urandom of=/dev/ad0e bs=64k and the error disappeared (from
HDTune). Something strange did appear though, a bad sector near the
beginning of the drive however I have not seen it since. I have run
multiple tests since and all were green.
> You might also be able to use that utility to get SMART stats for the
> drive, although smartctl -a /dev/ad0 should suffice too. The disk
Sorry but I have not used smartctl since as the dd wiped out large
portions of FreeBSD and I have not reinstalled it.
In summary: I have been having problems with FreeBSD when reading from
a certain area of the hard disk (a system crash occurs). I have been
able to determine that the hard drive is not corrupting (or the
corrupting has stabilized...?). More importantly I have determined
that other operating systems have not been having problems (Windows
and Linux (openSUSE 10.2 Rescue)) thus I conclude that FreeBSD is in
fact not handling the hardware properly (malfunctioning or buggy
driver? or more likely not with a required quark).
When I first installed FreeBSD I did a dd if=/dev/zero of=/dev/ad0
bs=1M. I think this may have contributed to the problem. I read
somewhere that optical drives do not handle 0's and 1's in continues
succession well and this may be applicable to hard drives (comments?)
and that FreeBSD is not handing correctly.
Should I submit a PR or is this better handled on the mailing list
(if it should be handled at all)?
Thank you for your help.
David
P.S. After further testing the occurrence of the crashes seems to be
less consistent. I will continue some testing to see if I can
determine a consistent pattern.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?b53f6f940711101156l7dbb9a09re4b382ed6653f3f9>