Thanks Bob, really can't thank you enough. Just to be clear about this, I'd
do these commands from the rescue disk after I have assembled the arrays
and gone to the bash shell?

Much appreciated,
James


On 2 July 2013 22:44, Bob Proulx <b...@proulx.com> wrote:

> James Allsopp wrote:
> > One other point sda isn't the boot hard drive, that's the partitions
> /sdb1
> > and sdc1, but these should be the same (I thought I'd mirrored them to be
> > honest).
>
> I don't see sda anywhere.  It might be a dual booting Windows disk?
> Or other.  But the BIOS will boot the first disk from the BIOS boot
> order.  BIOS boot order may be different from OS disk order.  It can
> be confusing.  I might assume that BIOS sata0 is the same as the OS
> disk sda but actually it often is different.  Let's ignore this for
> now.
>
> You have sdb1 and sdc1 mirrored into md1.  I can see that because the
> UUID is identical.
>
> > /dev/md1:
> >         Version : 1.2
> >   Creation Time : Thu Jan 31 22:43:49 2013
> >      Raid Level : raid1
> >      Array Size : 1953510841 (1863.01 GiB 2000.40 GB)
> >   Used Dev Size : 1953510841 (1863.01 GiB 2000.40 GB)
>
> >    Raid Devices : 2
> >   Total Devices : 1
> >     Persistence : Superblock is persistent
> >
> >     Update Time : Tue Jul  2 13:49:55 2013
> >           State : clean, degraded
> >  Active Devices : 1
> > Working Devices : 1
> >  Failed Devices : 0
> >   Spare Devices : 0
> >
> >            Name : Hawaiian:1  (local to host Hawaiian)
> >            UUID : a544829f:33778728:79870439:241c5c51
> >          Events : 112
> >
> >
> >     Number   Major   Minor   RaidDevice State
> >        0       0        0        0      removed
> >        1       8       65        1      active sync   /dev/sde1
>
> That info is the same as:
>
> > /dev/md127:
> >         Version : 1.2
> >   Creation Time : Thu Jan 31 22:43:49 2013
> >      Raid Level : raid1
> >      Array Size : 1953510841 (1863.01 GiB 2000.40 GB)
> >   Used Dev Size : 1953510841 (1863.01 GiB 2000.40 GB)
> >    Raid Devices : 2
> >   Total Devices : 1
> >     Persistence : Superblock is persistent
> >
> >     Update Time : Tue Jul  2 13:49:29 2013
> >           State : clean, degraded
> >  Active Devices : 1
> > Working Devices : 1
> >  Failed Devices : 0
> >   Spare Devices : 0
> >
> >            Name : Hawaiian:1  (local to host Hawaiian)
> >            UUID : a544829f:33778728:79870439:241c5c51
> >          Events : 106
> >
> >
> >     Number   Major   Minor   RaidDevice State
> >        0       8       49        0      active sync   /dev/sdd1
> >        1       0        0        1      removed
>
> The UUIDs are identical.  Therefore those two disks are mirrors of
> each other.  And note:
>
> > /dev/md1: (/dev/sde1)
> >     Update Time : Tue Jul  2 13:49:55 2013
> > /dev/md127: (/dev/sdd1)
> >     Update Time : Tue Jul  2 13:49:29 2013
>
> sde1 is newer than sdd1.  This seems consistent with it being the best
> copy to keep.  If it were the other way around I would think about
> using the other one.  But selecting the right master is important
> since it is a component of the lvm.
>
> > How should I proceed from here?
>
> I would proceed as previously suggested.  I would do this:
>
>   mdadm --stop /dev/md127
>   mdadm --manage /dev/md1 --add /dev/sdd1
>   watch cat /proc/mdstat
>
> That will discard the older stale copy of the mirror on sdd1.  It will
> use sdd1 as a mirror of sde1.  After doing the add the mirror will
> sync and you can watch the progress using 'watch cat /proc/mdstat'.
> Use control-c to interrupt it when you want to stop it.
>
> > For ruther information:
> > /dev/sdb3:
> > Preferred Minor : 126
> > ...
> > /dev/sdc3:
> > Preferred Minor : 126
> > ...
>
> That further information looked _okay_ to me.  But I would still
> change the md126 back to md0.
>
>   mdadm --stop /dev/md126
>   mdadm --assemble /dev/md0 --update=super-minor /dev/sdb1 /dev/sdc1
>   cat /proc/mdstat
>
> Since it is clean now it will be stopped cleanly and reassembled
> cleanly and no sync will be needed.  The --update=super-minor will
> reset the superblock with the updated md0 minor device number.
>
> Then update /etc/mdadm/mdadm.conf and rebuild the initrd.
>
> Bob
>

Reply via email to