RAID failure already!!!!!

Robin Laing Robin.Laing at
Thu Nov 16 18:52:43 UTC 2006

James Pifer wrote:
>>>i think you need something like
>>>      mdadm --assemble /dev/md0 /dev/hdb /dev/hdc
>>>you should then be able to add the third device using "mdadm --manage"
>>> (it's unclear if you need to "start" the array before doing this).
>>Yeah, tried that. It tries for a few seconds, then I get:
>>[root at storage ~]# mdadm --assemble /dev/md0 /dev/hdb /dev/hdc
>>mdadm: /dev/md0 assembled from 1 drive - not enough to start the array.
>>If I cat /proc/mdstat
>>[root at storage ~]# cat /proc/mdstat
>>Personalities :
>>md0 : inactive hdc[1] hdb[0]
>>      312581632 blocks super non-persistent
>>unused devices: <none>
> Any other suggestions here? Is there any way to determine 100% that I
> have no chance at getting to any of the data? It would be nice to get to
> some of it. It seems as though hdb and hdc are acting ok right now. I
> don't know if the reference to "dev/hdd" is causing issues or not. Is
> hdb marked "failed" somehow? If so, is there a way to modify it to not
> be failed so I can try and bring up the array?
> Sorry, just don't want to give up unless there's absolutely no hope. 
> Thanks,
> James

Read through the mdadm manuals and see if you can add the drives to the 
array and rebuild.  I don't know if it works in RAID5 as I had to 
rebuild a RAID 1 array.  Thinking that your RAID array was running with 
only two disks in the first place, then it may not have tried writing to 
the third disk.


Look at the testing methods for removing and adding drives and testing.

Robin Laing

More information about the users mailing list