[clug] strange output from df

[Ian McCulloch]

Hi,

On Thu, 29 Jul 2004, Gough, Chris wrote:

> Strange behaviour from a Debian box (2.4.20-idepci kernel).
> 
> The following output from df seems a bit nuts
> --
> Filesystem           1k-blocks      Used Available Use% Mounted on
> /dev/hda2            -579653282885         1         0  28% /
> /dev/hdb1            -2557280710065         1         0  73% /usr/SmbShare
> --
> 
> and the kernel is spewing millions of these into /var/log/messages 
> --
> Jul 29 18:24:22 myMachine kernel: hdb: status error: status=0x00 { }
> Jul 29 18:24:22 myMachine kernel: hdb: drive not ready for command
> --
> 
> and quite a lot of these
> --
> Jul 29 16:13:57 myMachine kernel: ide0: reset: success
> Jul 29 16:13:57 myMachine kernel: hdb: status error: status=0x00 { }
> Jul 29 16:13:57 myMachine kernel: end_request: I/O error, dev 03:41 (hdb),
> sector 32
> Jul 29 16:13:57 myMachine kernel: hdb: drive not ready for command
> Jul 29 16:13:57 myMachine kernel: bread in fat_access failed

Hi,

No idea if this is relevant (probably not actually ;), but for some 
kernels I see the same messages on my thinkpad R50.  The problem is the 
host protected area (where M-bloody-S store their 'recovery' software).  I 
disabled the HPA and put the /home partition there instead, but some 
(broken?) kernels mis-report the size of the drive as having the HPA still 
active, resulting in similar looking log-spam.  Do your disks have a host 
protected area?

Cheers,
Ian