>
> >>>>> On Mon, 10 Apr 2000, "Darren" == Darren Nickerson wrote:
> array.
>
> Ack - too late:
>
> Apr 11 09:58:24 osmin kernel: attempt to access beyond end of device
> Apr 11 09:58:24 osmin kernel: 21:01: rw=0, want=318373889,
> limit=33417184 Apr 11 09:58:24 osmin kernel: dev 09:01 blksize=4096
> blocknr=318373888 sector=636747776 size=1024 count=1 Apr 11 09:58:24
> osmin kernel: raid5: Disk failure on hde1, disabling device.
> Operation continuing on 3 devices Apr 11 09:58:24 osmin kernel:
> raid5: restarting stripe 636747776 Apr 11 09:58:24 osmin kernel:
>
I saw a bunch of this kind of thing once when BIOS and LINUX probe
disagreed about the disk geometry. Using the probe values for the
bios cleared it up. Seems that depending on the driver, Linux may
like the LBA values (logical) instead of the physical values. I
don't know enough about the whys?? of this to write intelligently,
it's just an observation. I simply decided to always use the values
displayed during the probe phase of the kernel start process. This
seems to work every time.
[EMAIL PROTECTED]
