Re: Assembly failure

On Tue, 10 Jul 2012 19:59:27 +0100 Brian Candler [off-list ref] wrote:

On Tue, Jul 10, 2012 at 07:38:51PM +0200, Sebastian Riemer wrote:

quoted

Your kernel is similar to v3.4 mainline. Your kernel has been compiled
one day after Linus tagged v3.4. This kernel has major issues. Please
reboot into the old 3.2 kernel.

Your kernel has no tag in the Ubuntu Git repos!

http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-precise.git;a=tags
http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-quantal.git;a=tags

Your kernel is absolutely unstable. Who built this kernel? Can't be
official release!

I don't know who makes ~kernel-ppa packages.

Anyway, box is now on linux-image-3.2.0-24-generic. Same problem:

brian@dev-storage1:~$ cat /proc/mdstat
Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10] 
md127 : inactive sdm[1](S) sdg[5](S) sdh[4](S) sdd[3](S) sdj[9](S) sdl[11](S) sdi[8](S) sdk[10](S) sdb[0](S) sde[7](S) sdf[6](S) sdc[2](S)
      35163186720 blocks super 1.2
       
unused devices: <none>

What's my best next step? There's nothing critical on here, but I would like
to use this as practice of recovering a broken md raid volume.

mdadm -S /dev/md127

Then assemble again with "--force" as you expected.

Don't try to --create --assume-clean, it isn't needed.
And don't worry too much about the kernel - though keep away from any Ubuntu
3.2 kernel before the one you have -there is a nasty bug (unrelated to your
current experience) that you don't want to go near.

When you re-assemble it won't include all the devices in the array - just
enough to make the array functional.  You would then need to add the others
back in if you trust them.


As others have suggested, there is probably some hardware problem somewhere.
It looks like sdr failed first, around "Jul  3 04:21:50 2012".
The array continued working until about 06:36 when sdu then sds failed.
Since then it doesn't look like much if anything has been written to the
array - but I cannot be completely certain.

Do you have kernel logs from the morning of 3rd July?

NeilBrown