Re: mdadm can't start RAID0 volume after power failure

Re: mdadm can't start RAID0 volume after power failure

[AndyLiebman]

Looking at your mdadm.conf file, and at the results of your mdadm -QE --scan, 
I can see that the UUID numbers are different. Which UUID numbers do you get 
when you run mdadm -E /dev/sd{e,f,g,h,i,j,k}1? I tried to follow the web page 
link in your post and I couldn't get to it. Are you trying to start your array 
with the wrong UUID number by chance? 

Regards, 
Andy Liebman

># cat /etc/mdadm/mdadm.conf
>
>DEVICE /dev/sd[efghijk]1
>ARRAY /dev/md0 level=raid0 num-devices=7 
>UUID=46c66ef8:8acf2810:48adb2d3:85b04900
>

>Still, every # mdadm -E /dev/sd[efghijk]1 looks ok, output can be found 
here: >http://philou.org/mdadm.log
>mdadm.conf is here : http://philou.org/mdadm.conf
>
># mdadm -QE --scan
>ARRAY /dev/md0 level=raid0 num-devices=7 
>UUID=3c1ad678:d3a7a1ed:1112143d:68b4e569

Re: mdadm can't start RAID0 volume after power failure

[Philippe Gramoullé]

Hello Andy,

On Fri, 17 Dec 2004 12:44:20 EST
AndyLiebman@aol.com wrote:

  | Looking at your mdadm.conf file, and at the results of your mdadm -QE --scan, 
  | I can see that the UUID numbers are different. Which UUID numbers do you get 
  | when you run mdadm -E /dev/sd{e,f,g,h,i,j,k}1?


  |  I tried to follow the web page  link in your post and I couldn't get to it.

The link is ok and works.

  |  Are you trying to start your array with the wrong UUID number by chance? 

Indeed !! starting it with the detected UUID allowed me to start it:

# mdadm -v -A /dev/md0 --uuid=3c1ad678:d3a7a1ed:1112143d:68b4e569
mdadm: looking for devices for /dev/md0
mdadm: /dev/sdk1 is identified as a member of /dev/md0, slot 6.
mdadm: /dev/sdj1 is identified as a member of /dev/md0, slot 5.
mdadm: /dev/sdi1 is identified as a member of /dev/md0, slot 4.
mdadm: /dev/sdh1 is identified as a member of /dev/md0, slot 3.
mdadm: /dev/sdg1 is identified as a member of /dev/md0, slot 2.
mdadm: /dev/sdf1 is identified as a member of /dev/md0, slot 1.
mdadm: /dev/sde1 is identified as a member of /dev/md0, slot 0.
mdadm: added /dev/sdf1 to /dev/md0 as 1
mdadm: added /dev/sdg1 to /dev/md0 as 2
mdadm: added /dev/sdh1 to /dev/md0 as 3
mdadm: added /dev/sdi1 to /dev/md0 as 4
mdadm: added /dev/sdj1 to /dev/md0 as 5
mdadm: added /dev/sdk1 to /dev/md0 as 6
mdadm: added /dev/sde1 to /dev/md0 as 0
mdadm: /dev/md0 has been started with 7 drives.

and it mounted fine ! 

Still, this is the first time that such a thing happens to me. What could be the reason that UUID would be changed ?

Thanks much,

Philippe

  | 
  | Regards, 
  | Andy Liebman
  | 
  | ># cat /etc/mdadm/mdadm.conf
  | >
  | >DEVICE /dev/sd[efghijk]1
  | >ARRAY /dev/md0 level=raid0 num-devices=7 
  | >UUID=46c66ef8:8acf2810:48adb2d3:85b04900
  | >
  | 
  | >Still, every # mdadm -E /dev/sd[efghijk]1 looks ok, output can be found 
  | here: >http://philou.org/mdadm.log
  | >mdadm.conf is here : http://philou.org/mdadm.conf
  | >
  | ># mdadm -QE --scan
  | >ARRAY /dev/md0 level=raid0 num-devices=7 
  | >UUID=3c1ad678:d3a7a1ed:1112143d:68b4e569
  | -
  | To unsubscribe from this list: send the line "unsubscribe linux-raid" in
  | the body of a message to majordomo@vger.kernel.org
  | More majordomo info at  http://vger.kernel.org/majordomo-info.html
  | 
  | 

Re: mdadm can't start RAID0 volume after power failure

[AndyLiebman]

Phillipe

>Still, this is the first time that such a thing happens to me. What could be 
the reason that >UUID would be changed ?

If I were a betting person, I would say that the mdadm.conf file was old, and 
that you had created a new array since the time you made that conf file. And 
you just didn't update the conf file. 

As I'm sure you know, you don't need to use a conf file anyway to use mdadm. 
That's what makes it so beautiful! All kinds of things -- like device ids -- 
can change on your system and mdadm just figures it out (provided you give it a 
minimal amount of info, like the correct uuid!). 

Glad to help :)
Andy Liebman

Re: mdadm can't start RAID0 volume after power failure

[Philippe Gramoullé]

Hello Andy,

On Fri, 17 Dec 2004 13:06:56 EST
AndyLiebman@aol.com wrote:

  | 
  | Phillipe
  | 
  | >Still, this is the first time that such a thing happens to me. What could be 
  | the reason that >UUID would be changed ?
  | 
  | If I were a betting person, I would say that the mdadm.conf file was old,

Yes it was, the filesystem was created last February.

  |  and  that you had created a new array since the time you made that conf file. And 
  | you just didn't update the conf file. 

Nope, nothing has changed since, no new array was created nor deleted nor modified.

  | 
  | As I'm sure you know, you don't need to use a conf file anyway to use mdadm. 

Yes, i used a straightforward mdadm command to create the array, 

  | That's what makes it so beautiful! All kinds of things -- like device ids -- 
  | can change on your system and mdadm just figures it out (provided you give it a 
  | minimal amount of info, like the correct uuid!). 

Sure, but again, _nothing_ had changed between last February and today, at least regarding hardware
and kernel. I regularly updated the Debian packages, including mdadm, but i don't think that was the problem

The mdadm.conf is the one from when the array was created, and until today i have shutdown and powered on the box
right fine, for weeks and months.

Maybe someone will be able to explain how the UUID can change on the disks, so that what's on disk and
what's in the conf can get out of sync.
  | 
  | Glad to help :)

You made my day :)

Thanks,

Philippe

mdadm can't start RAID0 volume after power failure

[Philippe Gramoullé]

 Hello,

i'm looking to understand why i can't start my RAID0 volume since i had a power failure this afternoon.
Nothing regarding hardware was changed/added/removed.

System    : Debian Unstable + Kernel 2.6.7-mm4 (worked rock solid for months)
Hardware  : Dell Poweredge 2600 + 2xPowervault 210S diskshelves + PERC3/QC RAID adapter
Filesystem: ReiserFS, /dev/md0 built last February, 500 Go
mdadm     : 1.7.0

The underlying Hardware RAID 1 part looks ok : each /dev/sd[efghijk]1 is a hardware RAID 1
which is set on the PERC RAID adapter. Every logical volume is reported as "optimal" state

Out of these 7 logical volumes, i've built a Software RAID-0 using mdadm:

# cat /etc/mdadm/mdadm.conf

DEVICE /dev/sd[efghijk]1
ARRAY /dev/md0 level=raid0 num-devices=7 UUID=46c66ef8:8acf2810:48adb2d3:85b04900
   devices=/dev/sdk1,/dev/sdj1,/dev/sdi1,/dev/sdh1,/dev/sdg1,/dev/sdf1,/dev/sde1

After the power failure, i can't start the RAID0 device anymore:
 
# mdadm -v -A /dev/md0
mdadm: looking for devices for /dev/md0
mdadm: /dev/sdk1 has wrong uuid.
mdadm: /dev/sdj1 has wrong uuid.
mdadm: /dev/sdi1 has wrong uuid.
mdadm: /dev/sdh1 has wrong uuid.
mdadm: /dev/sdg1 has wrong uuid.
mdadm: /dev/sdf1 has wrong uuid.
mdadm: /dev/sde1 has wrong uuid.
mdadm: no devices found for /dev/md0


Still, every # mdadm -E /dev/sd[efghijk]1 looks ok, output can be found here: http://philou.org/mdadm.log
mdadm.conf is here : http://philou.org/mdadm.conf

# mdadm -QE --scan
ARRAY /dev/md0 level=raid0 num-devices=7 UUID=3c1ad678:d3a7a1ed:1112143d:68b4e569
   devices=/dev/sdk1,/dev/sdj1,/dev/sdi1,/dev/sdh1,/dev/sdg1,/dev/sdf1,/dev/sde1


Any hope to get my data back ?

Thanks,

Philippe