Corrupted RAID 1

Corrupted RAID 1

[Leslie Rhorer]

	I have a corrupted bootable RAID 1 array on a pair of SDD drives, and I 
fear I need some assistance.  Actually, there are four partitions on 
each drive, and two RAID 1 arrays were assembled from each drive.  When 
working properly, the second pair of partitions were mounted as / and 
the first pair were mounted as /boot.  The OS is Debian Buster.  When I 
attempt to boot the system, it goes directly to the GRUB prompt.

	I pulled the drives and attached them to an active system.  Mdadm 
reports both partition tables to be intact with partition labels of fd, 
83, ef, and 82, respectively and MBR magic of aa55.  When I try to 
assemble any array says the partitions exist but are not md arrays.

	Fdisk reports the partition types as Linux raid autodetect, Linux, EFI, 
and Linux swap / Solaris.  The EFI partitions are marked bootable and 
contain the following:

total 3472
-rwxr-xr-x 1 root root     108 May 28  2022 BOOTX64.CSV
-rwxr-xr-x 1 root root   84648 May 28  2022 fbx64.efi
-rwxr-xr-x 1 root root     152 May 28  2022 grub.cfg
-rwxr-xr-x 1 root root 1672576 May 28  2022 grubx64.efi
-rwxr-xr-x 1 root root  845480 May 28  2022 mmx64.efi
-rwxr-xr-x 1 root root  934240 May 28  2022 shimx64.efi

	Any suggestions?  I should say that the running system
(Bullseye) is not the same version as the failed one (Buster).  Of 
course the failed system does need to be upgraded, but there are 
specific reasons why this is quite undesirable at this point.

Re: Corrupted RAID 1

[Yu Kuai]

Hi,

在 2023/07/14 4:58, Leslie Rhorer 写道:
>      I have a corrupted bootable RAID 1 array on a pair of SDD drives, 
> and I fear I need some assistance.  Actually, there are four partitions 
> on each drive, and two RAID 1 arrays were assembled from each drive.  
> When working properly, the second pair of partitions were mounted as / 
> and the first pair were mounted as /boot.  The OS is Debian Buster.  
> When I attempt to boot the system, it goes directly to the GRUB prompt.
> 
>      I pulled the drives and attached them to an active system.  Mdadm 
> reports both partition tables to be intact with partition labels of fd, 
> 83, ef, and 82, respectively and MBR magic of aa55.  When I try to 
> assemble any array says the partitions exist but are not md arrays.
> 
>      Fdisk reports the partition types as Linux raid autodetect, Linux, 
> EFI, and Linux swap / Solaris.  The EFI partitions are marked bootable 
> and contain the following:
> 
> total 3472
> -rwxr-xr-x 1 root root     108 May 28  2022 BOOTX64.CSV
> -rwxr-xr-x 1 root root   84648 May 28  2022 fbx64.efi
> -rwxr-xr-x 1 root root     152 May 28  2022 grub.cfg
> -rwxr-xr-x 1 root root 1672576 May 28  2022 grubx64.efi
> -rwxr-xr-x 1 root root  845480 May 28  2022 mmx64.efi
> -rwxr-xr-x 1 root root  934240 May 28  2022 shimx64.efi
> 
>      Any suggestions?  I should say that the running system
> (Bullseye) is not the same version as the failed one (Buster).  Of 
> course the failed system does need to be upgraded, but there are 
> specific reasons why this is quite undesirable at this point.
> .

There really is not enough information, what is the kernel version?
And in the active system, what mdadm cmds you're using and what's the
result? (And please show us the result of mdadm -E /dev/[partition]).

Thanks,
Kuai
> 

Re: Corrupted RAID 1

[Wols Lists]

On 18/07/2023 10:25, Yu Kuai wrote:
> Hi,
> 
> 在 2023/07/14 4:58, Leslie Rhorer 写道:
>>      I have a corrupted bootable RAID 1 array on a pair of SDD drives, 
>> and I fear I need some assistance.  Actually, there are four 
>> partitions on each drive, and two RAID 1 arrays were assembled from 
>> each drive. When working properly, the second pair of partitions were 
>> mounted as / and the first pair were mounted as /boot.  The OS is 
>> Debian Buster. When I attempt to boot the system, it goes directly to 
>> the GRUB prompt.
>>
>>      I pulled the drives and attached them to an active system.  Mdadm 
>> reports both partition tables to be intact with partition labels of 
>> fd, 83, ef, and 82, respectively and MBR magic of aa55.  When I try to 
>> assemble any array says the partitions exist but are not md arrays.
>>
>>      Fdisk reports the partition types as Linux raid autodetect, 
>> Linux, EFI, and Linux swap / Solaris.  The EFI partitions are marked 
>> bootable and contain the following:
>>
>> total 3472
>> -rwxr-xr-x 1 root root     108 May 28  2022 BOOTX64.CSV
>> -rwxr-xr-x 1 root root   84648 May 28  2022 fbx64.efi
>> -rwxr-xr-x 1 root root     152 May 28  2022 grub.cfg
>> -rwxr-xr-x 1 root root 1672576 May 28  2022 grubx64.efi
>> -rwxr-xr-x 1 root root  845480 May 28  2022 mmx64.efi
>> -rwxr-xr-x 1 root root  934240 May 28  2022 shimx64.efi
>>
>>      Any suggestions?  I should say that the running system
>> (Bullseye) is not the same version as the failed one (Buster).  Of 
>> course the failed system does need to be upgraded, but there are 
>> specific reasons why this is quite undesirable at this point.
>> .
> 
> There really is not enough information, what is the kernel version?
> And in the active system, what mdadm cmds you're using and what's the
> result? (And please show us the result of mdadm -E /dev/[partition]).
> 
@kuai

Point them at the wiki ...

https://raid.wiki.kernel.org/index.php/Linux_Raid#When_Things_Go_Wrogn

Cheers,
Wol