RAID problem

RAID problem

[fixitdad5]

If you are still offering help with raid corruption due to the bug in kernels 3.2 and 3.3, I would like some help.

I load Ubuntu 12.04 precise (with kernel 3.2 since it was in the package) just 2 weeks ago. I had pre-existing raids and the raid's meta data was corrupted on the first boot after the successful installation so I have lost my data. I tried most of the suggestions that I found on the web, employing the intelligence in mdadm -assemble, etc. without success.

I have now updated the kernel to 3.4.0 rc7 (3.4.0-030400-generic #201205210521 SMP) and built an empty raid. However, I cannot boot without disconnecting the RAID drives. With the RAID drives hooked up and the updated kernel with the additional parameters listed below, I I fall into busybox and cannot boot successfully. What do I do now?
 
My configuration is 
Intel(R) Core(TM) i5 CPU 750 @ 2.67GHz
7983 MiB
running SELinux
Raid is a simple stripping (level 0) for speed
I have set the noautodetect and nodmraid parameters for the kernel

Re: RAID problem

[NeilBrown]

[-- Attachment #1: Type: text/plain, Size: 1825 bytes --]

On Wed, 14 Aug 2013 15:49:37 -0500 (CDT) fixitdad5@verizon.net wrote:

> If you are still offering help with raid corruption due to the bug in kernels 3.2 and 3.3, I would like some help.
> 
> I load Ubuntu 12.04 precise (with kernel 3.2 since it was in the package) just 2 weeks ago. I had pre-existing raids and the raid's meta data was corrupted on the first boot after the successful installation so I have lost my data. I tried most of the suggestions that I found on the web, employing the intelligence in mdadm -assemble, etc. without success.
> 
> I have now updated the kernel to 3.4.0 rc7 (3.4.0-030400-generic #201205210521 SMP) and built an empty raid. However, I cannot boot without disconnecting the RAID drives. With the RAID drives hooked up and the updated kernel with the additional parameters listed below, I I fall into busybox and cannot boot successfully. What do I do now?
>  
> My configuration is 
> Intel(R) Core(TM) i5 CPU 750 @ 2.67GHz
> 7983 MiB
> running SELinux
> Raid is a simple stripping (level 0) for speed
> I have set the noautodetect and nodmraid parameters for the kernel

We would need lots more details to be able to help.
I assume you have given up on the old data - is that correct?  Sorry about
that.

Do you just have the one RAID (the RAID0)?  Is it for the root filesystem or
something else?

Maybe there is something in your initrd that is causing confusion.


 zcat /boot/initrd | cpio -idv
will explode it for you.  Is there an 'mdadm.conf' in there?  What does it
contain?

Maybe just run
   mkinitrd
(or whatever the Ubuntu command is - mkinitramfs maybe).
That might fix it.

Are there  any error messages during boot?

You might need to get Ubuntu specific help as different distros have quite
different boot sequences.

NeilBrown

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 828 bytes --]

RAID Problem

[colli419]

Hello, I am having problems with my RAID as follows:

I installed the drives in an old computer and created the RAID and it 
worked great. When I tried to install the same RAID set in a new computer I 
could get it to mount manually with some messing around but couldn't get it 
to mount automatically as in the old machine. I was wondering if zeroing 
the superblocks and trying to re-create the RAID would be a good idea. The 
data that is stored on the drives should otherwise be intact because I had 
it running just fine the other day. Any help at all would be most 
appreciated.

Thank you so much,
Michael Collins

Re: RAID Problem

[Neil Brown]

On 16 Mar 2010 18:32:39 -0500
colli419@umn.edu wrote:

> Hello, I am having problems with my RAID as follows:
> 
> I installed the drives in an old computer and created the RAID and it 
> worked great. When I tried to install the same RAID set in a new computer I 
> could get it to mount manually with some messing around but couldn't get it 
> to mount automatically as in the old machine. I was wondering if zeroing 
> the superblocks and trying to re-create the RAID would be a good idea. The 
> data that is stored on the drives should otherwise be intact because I had 
> it running just fine the other day. Any help at all would be most 
> appreciated.
> 

You haven't provides a lot of concrete information, like error messages
during boot or "mdadm -E" output of devices or even kernel/mdadm versions, so
I can only guess, but my guess would be that you can fix it by assembling
the array with "--update=homehost"

i.e.
   mdadm --assemble /dev/mdwhatever  \
   --update=homehost /dev/device1 /dev/device2 ....

NeilBrown

Re: RAID Problem

[Bill Davidsen]

Neil Brown wrote:
> On 16 Mar 2010 18:32:39 -0500
> colli419@umn.edu wrote:
>
>   
>> Hello, I am having problems with my RAID as follows:
>>
>> I installed the drives in an old computer and created the RAID and it 
>> worked great. When I tried to install the same RAID set in a new computer I 
>> could get it to mount manually with some messing around but couldn't get it 
>> to mount automatically as in the old machine. I was wondering if zeroing 
>> the superblocks and trying to re-create the RAID would be a good idea. The 
>> data that is stored on the drives should otherwise be intact because I had 
>> it running just fine the other day. Any help at all would be most 
>> appreciated.
>>
>>     
>
> You haven't provides a lot of concrete information, like error messages
> during boot or "mdadm -E" output of devices or even kernel/mdadm versions, so
> I can only guess, but my guess would be that you can fix it by assembling
> the array with "--update=homehost"
>
> i.e.
>    mdadm --assemble /dev/mdwhatever  \
>    --update=homehost /dev/device1 /dev/device2 ....
>
>   

My take on hostname is that it causes more users problems than it helps, 
and that it would have been better as an option to the assemble function 
for those who actually need that checking. More people have little RAID 
USB boxes than have some shared storage.

Do *NOT* read this as a suggestion to change, just a comment on it as an 
example of "unintended results."

-- 
Bill Davidsen <davidsen@tmr.com>
  "We can't solve today's problems by using the same thinking we
   used in creating them." - Einstein