[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#574817: libata transition breaks mdadm arrays



On Sun, 2010-03-21 at 15:20 +0100, Mathieu Parent (Debian) wrote:
> On Sun, Mar 21, 2010 at 3:03 PM, Bastian Blank <waldi@debian.org> wrote:
> > On Sun, Mar 21, 2010 at 01:38:53PM +0100, Mathieu Parent (Debian) wrote:
> >> On Sun, Mar 21, 2010 at 1:35 PM, Bastian Blank <waldi@debian.org> wrote:
> >> > This is expected and not an error. Don't list device-names in
> >> > /etc/mdadm.conf.
> >> I am not using device names, see attached file.
> >
> > Hmm, the file is /etc/mdadm/mdadm.conf and mdadm may have generated a
> > config during a long ago upgrade, see
> > /usr/share/doc/mdadm/README.upgrading-2.5.3.gz.
> 
> The attached file was actually /etc/mdadm/mdadm.conf (I don't have
> /etc/mdadm.conf). And I have already read the
> README.upgrading-2.5.3.gz and done what is expected. i have already
> done "rm -f /var/lib/mdadm/CONF-UNCHECKED".
> 
> I have the problem with 2.6.32-10, not with 2.6.32-9. I really don't
> were mdadm find device-names, I only use uuid in conffiles.

The files you sent show that in the broken state mdadm has found some
but not all of the devices.  So I would say the configuration is fine
but perhaps mdadm is not waiting long enough to find all the devices
that make up the array at boot time.

If /etc/mdadm/mdadm.conf is likely to contain device names then I should
deal with this in the upgrade script.  However this does not seem to be
an issue in this case.

Ben.

-- 
Ben Hutchings
If you seem to know what you are doing, you'll be given more to do.

Attachment: signature.asc
Description: This is a digitally signed message part


Reply to: