isw attempt to access beyond end of device

Dan Williams dan.j.williams at intel.com
Thu Apr 9 16:47:42 UTC 2009


On Sun, Apr 5, 2009 at 11:11 AM, Greg Watson <gwatson at linuxlogin.com> wrote:
> Hey guys,
>
> I'm having a problem on boot and at random times:
>
>
> [   36.231451] attempt to access beyond end of device
> [   36.231462] sda: rw=0, want=156424715, limit=125045424
> [   36.231467] Buffer I/O error on device sda2, logical block 156264064
> [   36.231473] attempt to access beyond end of device
> [   36.231477] sda: rw=0, want=156424716, limit=125045424
> [   36.231481] Buffer I/O error on device sda2, logical block 156264065
> [   36.231486] attempt to access beyond end of device
> [   36.231489] sda: rw=0, want=156424717, limit=125045424
> [   36.231493] Buffer I/O error on device sda2, logical block 156264066
> [   36.231498] attempt to access beyond end of device
> [   36.231501] sda: rw=0, want=156424718, limit=125045424
> [   36.231505] Buffer I/O error on device sda2, logical block 156264067
> [   36.231509] attempt to access beyond end of device
> [   36.231512] sda: rw=0, want=156424719, limit=125045424
> [   36.231517] Buffer I/O error on device sda2, logical block 156264068
> [   36.231521] attempt to access beyond end of device
> [   36.231525] sda: rw=0, want=156424720, limit=125045424
> [   36.231530] Buffer I/O error on device sda2, logical block 156264069
> [   36.231534] attempt to access beyond end of device
> [   36.231537] sda: rw=0, want=156424721, limit=125045424
> [   36.231542] attempt to access beyond end of device
>
> < snip about 300 lines >
>
> [   36.242813] sda: rw=0, want=156424907, limit=125045424
> [   36.242820] attempt to access beyond end of device
> [   36.242823] sda: rw=0, want=156424907, limit=125045424
>
>
> This then repeats a few times during boot, and onces or twice while
> system is running.  The system becomes unresponsive when it happens.
> Any ideas on how to fix this?  The system has been running, for the
> most part, without other issues for a while.  I'd just like to get it
> fixed if possible.
> Thanks

This is not a dmraid problem, it is a kernel problem with recognizing
raid members.  The kernel sees a partition table on device sda, but
the partition table is for the raid0 and extends out to sdb.  A fix
for this went in for 2.6.28: "block: sanitize invalid partition table
entries" [1].

--
Dan

[1]: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=ac0d86f5809598ddcd6bfa0ea8245ccc910e9eac




More information about the Ataraid-list mailing list