[TriLUG] mdadm: no devices found for /dev/md0

T. Bryan tbryan at python.net
Thu Nov 30 07:57:53 EST 2006


For the archives...

On Wednesday 29 November 2006 19:51, T. Bryan wrote:

> Some progress!  I think that the problem may be that I once created a RAID
> on these devices incorrectly.  I later reconfigured my RAID, but I think
> that the boot sequence is now detecting the old RAID set.  I could use some
> help from another RAID user out there.
>
> If you have a RAID1 between /dev/hda1 and /dev/hdc1, for example, what do
> the following 4 commands return?

Nevermind.

> Anyway, I think that I might just need to run
>
> mdadm --zero-superblock /dev/hde
> mdadm --zero-superblock /dev/hdg
>
> and reboot.  I just don't want to do that until I'm a little more certain
> that the md superblock isn't supposed to be there.

I decided to give it a try without a confirmation that this guess was what was 
really happening.  As a German friend used to say, "Mut zur Luecke!" (Courage 
through the gap.)  

Zeroing the superblocks fixed my problem and didn't appear to break anything 
else.  

I think that after my recent upgrades, udev was seeing the md superblock on 
the /dev/hde and /dev/hdg devices and not even creating the /dev/hde1 
and /dev/hdg1 devices.  Now that there are no longer obsolete md superblocks 
on /dev/hde and /dev/hdg, udev is making /dev/hde1 and /dev/hdg1 available 
after boot, and mdadm is able to bring up /dev/md0 as a RAID1 across those 
devices.

---Tom



More information about the TriLUG mailing list