David Chinner wrote:
> On Fri, Jun 08, 2007 at 03:59:39PM +0200, Marco Berizzi wrote:
> > David Chinner wrote:
> > > Where we saw signs of on disk directory corruption. Have you run
> > > xfs_repair successfully on the filesystem since you reported
> > > this?
> >
> > yes.
> >
> > > If you did clean up the error, does xfs_repair report the same
sort
> > > of error again?
> >
> > I have run xfs_repair this morning.
> > Here is the report:
>
> <reports no on disk errors>
>
> > > Have you run a 2.6.16-rcX or 2.6.17.[0-6] kernel since you last
> > > reported this problem?
> >
> > No. I have run only 2.6.19.x and 2.6.21.x
> >
> > After the xfs_repair I have remounted the file system.
> > After few hours linux has crashed with this message:
> > BUG: at arch/i386/kernel/smp.c:546 smp_call_function()
> > I have also the monitor bitmap.
>
> This is sounding like memory corruption is no corruption is being
> found on disk by xfs_repair. Have you run memtest86 on that box to
> see if it's got bad memory?
Yes. I have run memtest for one week:
no errors.
I have also changed the mother board,
scsi controller and ram. Only the cpu
and the 2 hot swap scsi disks were
not replaced. IMHO this isn't an
hardware problem, because the kernel
with debugging options enabled didn't
crash for a long time (>1 month). Just
for record, at this moment this box is
running 2.6.22-rc4 with no debug
options enabled. I will keep you
informed.
Thanks everybody for the support.
|