mdadm array troubles.
I have an intel storage array, 4-disk nfs/smb NAS job. A little while ago it failed, and I started down the path of trying to recover my data.
It turns out that under the covers, the thing is linux, and uses mdadm to build the array. So took three of the disks (one of them is actually dead) and mounted them in a machine with adequate sata ports, and booted it up off of a fedora live disc. It see's the mdadm array, but i cant get mdadm to assemble the array. It keeps telling me that a minimum of 3 disks are required, and that it can only detect 2. using mdadm --examine on the individual disks gives me a hint that things are out of whack, and i'm wondering how to resolve them. I'll paste output from the live image in a bit, i'm posting from a different machine. Basically, each disk tells me that a different number of drives are available, and i'm wondering if I can override that. Essentially, i want to fail the bad disk which isnt even present, and then make sure that the three working disks are active. |
Ok, here's some output.
First, attempting to assemble the array based on uuid. Code:
[root@localhost liveuser]# mdadm --assemble --scan --uuid "dbeab1fc:fbec2bad:b0a43543:72ffeee2" Code:
[root@localhost ~]# mdadm -v --assemble /dev/md1 /dev/sda3 /dev/sdb3 /dev/sdd3 Code:
[root@localhost liveuser]# for i in a b d;do mdadm --examine /dev/sd${i}3;done; |
I notice that the checksum is different on sda3. Makes me wonder if that disk was somehow not a member of the array with the 4th disk failed.
Which makes me think my data is likely lost. Any information would be useful. Thanks! |
All times are GMT -5. The time now is 06:20 AM. |