Bug#361408: Further information

[Detlef Schmicker]

> I suggest using UUIDs instead, as that will be a lot more robust
> when the kernel decides to rename devices. It happens.

I was not able to start with UUIDs

> Yes, this *is* weird. Could you please post the output of
> 
>   for dev in /dev/[hs]d[a-z]*; do mdadm -E $dev 2>/dev/null && echo --; done
> 

Since my last post, I repartitioned part of hdc and added /dev/hdc2
to /dev/md0 and /dev/hdc3 to /dev/md1 and all problems vanished. Now I
can use UUIDs. The wired output is ok now.
Both hdc partitions belonged to md devices in the past, but were not
used anymore. In case you need other information: I have original hard
disks (with on device of every raid) of the problem after upgrade lying
on my table). Seems to be difficult to give good information to you?!
Thanks detlef
-- 
Detlef Schmicker <d.schmicker at physik.de>