mdadm udev rules don't appear to work correctly

[Bruno Wolff III]

Since upgrading to mdadm 3 I have had to remove the udev rules file for mdadm
as they were having arrays set up twice duruing the udev step resulting in
booting not completing.
Even with them removed if something happens to an array that drops an
element, both parts result in an array being created instead of leaving
the failed elements alone. However, this does stop the boot process
from completing.
Gnome has been fubar'd since the mass rebuild making a bit harder to
fill in bug reports (and a lot of other stuff), but I'll eventually
get around to filing a real bug report.
[root@cerberus ~]# cat /proc/mdstat
Personalities : [raid1]
md126 : active raid1 sdb3[1]
      41945600 blocks [2/1] [_U]

md127 : active raid1 sdb6[1]
      149500736 blocks [2/1] [_U]

md0 : active raid1 sda1[0] sdb1[1]
      264960 blocks [2/2] [UU]

md3 : active raid1 sda2[0] sdb2[1]
      41945600 blocks [2/2] [UU]

md4 : active raid1 sda6[0]
      149500736 blocks [2/1] [U_]

md1 : active raid1 sda5[0] sdb5[1]
      10482304 blocks [2/2] [UU]

md2 : active raid1 sda3[0]
      41945600 blocks [2/1] [U_]

unused devices: <none>

The md126 and md127 arrays shouldn't exist. They started showing up after
sdb3 and sdb6 had issues during a boot that got them kicked out of their
arrays.
With the mdadm udev rules file in place I also get an md128.

[Adam Williamson]

On Fri, 2009-03-06 at 11:21 -0600, Bruno Wolff III wrote:
>  Since upgrading to mdadm 3 I have had to remove the udev rules file
> for mdadm
>  as they were having arrays set up twice duruing the udev step resulting in
>  booting not completing. 
Have you filed a bug?
-- 
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org
http://www.happyassassin.net

[Bruno Wolff III]

On Fri, Mar 06, 2009 at 09:43:15 -0800,
  Adam Williamson <awilliam(a)redhat.com&gt; wrote:
>  On Fri, 2009-03-06 at 11:21 -0600, Bruno Wolff III wrote:
>  > Since upgrading to mdadm 3 I have had to remove the udev rules file for mdadm
>  > as they were having arrays set up twice duruing the udev step resulting in
>  > booting not completing.
>  
>  Have you filed a bug? 
I started to in lynx, but after losing it once by hitting too many
left arrows, I think I'll wait until I get home. (Though the phone broke
there this morning so I am not sure I'll have access from home tonight
either.) At home I have rawhide from before the mass rebuild and things
are working a lot better.

[Robin Laing]

Bruno Wolff III wrote:
>  Since upgrading to mdadm 3 I have had to remove the udev rules file
> for mdadm
>  as they were having arrays set up twice duruing the udev step resulting in
>  booting not completing.
>  Even with them removed if something happens to an array that drops an
>  element, both parts result in an array being created instead of leaving
>  the failed elements alone. However, this does stop the boot process
>  from completing.
>  Gnome has been fubar'd since the mass rebuild making a bit harder to
>  fill in bug reports (and a lot of other stuff), but I'll eventually
>  get around to filing a real bug report.
>  [root@cerberus ~]# cat /proc/mdstat
>  Personalities : [raid1]
>  md126 : active raid1 sdb3[1]
>        41945600 blocks [2/1] [_U]
>  
>  md127 : active raid1 sdb6[1]
>        149500736 blocks [2/1] [_U]
>  
>  md0 : active raid1 sda1[0] sdb1[1]
>        264960 blocks [2/2] [UU]
>  
>  md3 : active raid1 sda2[0] sdb2[1]
>        41945600 blocks [2/2] [UU]
>  
>  md4 : active raid1 sda6[0]
>        149500736 blocks [2/1] [U_]
>  
>  md1 : active raid1 sda5[0] sdb5[1]
>        10482304 blocks [2/2] [UU]
>  
>  md2 : active raid1 sda3[0]
>        41945600 blocks [2/1] [U_]
>  
>  unused devices: <none>
>  
>  The md126 and md127 arrays shouldn't exist. They started showing up after
>  sdb3 and sdb6 had issues during a boot that got them kicked out of their
>  arrays.
>  With the mdadm udev rules file in place I also get an md128.
>   
This thread hit home this weekend.  My RAID drives decided to show up 
like the above.  Some had both drives working and others just one.

When I was setting up F10, I had problems with the raid arrays and it 
took some time to figure out that I needed to manually create the 
mdadm.conf file.  This is what I needed to get rid of the md126/127 type 
numbers that started to appear.

As long as I tried to depend on the udev, I couldn't create RAID 1 arrays.

-- 
Robin Laing

[Emmanuel Galanos]

On Mon, Mar 09, 2009 at 01:54:04PM -0600, Robin Laing wrote:
>  
>  This thread hit home this weekend.  My RAID drives decided to show up 
>  like the above.  Some had both drives working and others just one. 
I'm getting the same problems with my RAID 1 /boot. Commenting out
udev rules in 70-mdadm-assemble.rules prevented the problem.

--
e