Re: broken raid level 5 array caused by user error

[Phil Turmel <philip@turmel.org>]

On 11/10/2015 06:47 PM, Mathias Mueller wrote:
> Hi Phil

> sorry, here you go:
> 
> sdb -> sde  Serial Number: JK1170YBHYV6MD
> sdc -> sdf  Serial Number: JK1100YAG64A1T
> sdd -> sdg  Serial Number: JK1121YAG7YDLS
> sde -> sdh  Serial Number: ML0220F30PZUVD
> 
> 
>> We know the possible order combinations:
>>  sdc1 sdb1 sdd1 sde1
>>  sdc1 sdd1 sdb1 sde1
>>  sde1 sdb1 sdd1 sdc1
>>  sde1 sdd1 sdb1 sdc1
> 
> so I used this combinations:
> 
> sdf1 sde1 sdg1 sdh1
> sdf1 sdg1 sde1 sdh1
> sdh1 sde1 sdg1 sdf1
> sdh1 sdg1 sde1 sdf1
> 
> 
>> For now, use data offset 2048 for all of them.
> 
> output of all eight combinations:

Hmmm.  Very strange.  Let's see the dump of the start of the data area of all four partitions:

for x in /dev/sd[efgh]1 ; do echo -e "\n$x\n" ; dd if=$x skip=2048 count=16 2>/dev/null |hexdump -C ; done

Phil