Bug#495580: mdadm: 4 disk raid10 with 1 active and 3 spare possible

[Felix Zielcke]

Am Montag, den 18.08.2008, 19:54 -0300 schrieb martin f krafft: 
> > You need to tell me exactly what you did. What you describe is not
> > possible. I don't contest you are seeing a problem, but I have done
> > these steps hundreds of times without any problem ever.

As I found the bug I even didn't know that you have a `call for testers'
on PTS. This was totally by accident.
I never use mdadm much or RAID at all.

With my little understanding of it, I just think that this is wrong.
RAID 10 with 1 fully synced disk and 3 spare.

In the case of `mdadm' I am a stupid user.
I should have said in the report `Probable I did something stupid, non
realistic and wrong' :)

My mail today morning was probable a bit too short, because I assumed
the question is `how do you get into this situation'.

Am Dienstag, den 19.08.2008, 08:29 +0200 schrieb Felix Zielcke:
> Ok, just did it now from scratch.
[...]
> fz-vm:~# mdadm -Q --detail /dev/md0

Attached is now the full output of everything inclusive my little `vm'
ssh alias and logout message :)

I did it now again like that.
After that `mdadm -Q --detail /dev/md0':

fz-vm:~# mdadm -A
--update=resync /dev/md0 /dev/sdc1 /dev/sdd1 /dev/sde1 /dev/sdf1
mdadm: device /dev/md0 already active - cannot assemble it
fz-vm:~# mdadm --grow -l10 -n4 /dev/md0
mdadm: raid10 array /dev/md0 cannot be reshaped.
fz-vm:~# mdadm -S /dev/md0
mdadm: stopped /dev/md0
fz-vm:~# mdadm -R /dev/md0
mdadm: failed to run array /dev/md0: Invalid argument
fz-vm:~# mdadm -A
--update=resync /dev/md0 /dev/sdc1 /dev/sdd1 /dev/sde1 /dev/sdf1
mdadm: /dev/md0 assembled from 1 drive and 3 spares - not enough to
start the array.

So maybe I should have used `mdadm --zero-superblock' between `mdadm
-r /dev/sdc1' and mdadm `mdadm -a /dev/sdc1' ?
Or maybe I should have done that `mdadm -A --update=resync' in between?

As far as I remember I did this before I reported the bug after I
removed the first disk (sdc1)

But for a stupid user like me in that case, it looks like something
should be changed.
As you can see I just removed and added the disks and didn't check in
between if they get added as spare.
Only in the end I saw `Oh I shouldn't do this' :D
-------------- next part --------------
A non-text attachment was scrubbed...
Name: mdadm.log
Type: text/x-log
Size: 8565 bytes
Desc: not available
Url : http://lists.alioth.debian.org/pipermail/pkg-mdadm-devel/attachments/20080819/7f645cdb/attachment.bin