On Thu, 2004-05-06 at 13:18, Nathan Scott wrote:
> On Thu, May 06, 2004 at 09:26:38AM -0700, Jeffrey W. Baker wrote:
> > We've been having tons of trouble with whatever version of XFS is merged
> > into the 2.4 kernel. At first (some months ago, as reported on this
> > list) we believed the XFS code was having trouble recovering hard I/O
> > errors on our SCSI-attached RAID, which is easy to understand. But
> > yesterday, on a 2.4.26 machine with an SATA-attached 4-way RAID-0 stripe
> > set, we got this error from XFS, without any I/O errors at all:
> >
> > xfs_force_shutdown(md(9,0),0x8) called from line 1070 of file xfs_trans.c.
> > Return address = 0xc0295edc
> > xfs_iunlink_remove: xfs_itobp() returned an error 5 on md(9,0). Returning
> > error.
>
> errno 5 is EIO - this could be coming from either XFS or it
> could be percolating up from MD.
Could be, but md logged nothing.
> > Filesystem "md(9,0)": Corruption of in-memory data detected. Shutting down
> > filesystem: md(9,0)
> > Please umount the filesystem, and rectify the problem(s)
> >
> > XFS just decided out of the blue it was hosed. This has been reported
> > on this list a number of times, normally with NFS involved. But this
> > filesystem was not exported with NFS, it was simply running bonnie,
> > memtest.sh, and dd md0 all at once. That's not even a remarkable I/O
> > load, and it was less than 1 hour of this load before failure.
>
> yeah, what Eric said - are you reading the raw device while
> reading/writing to the filesystem? why?
Yes, and just to exercise the device.
-jwb
|