recover data on defective buffalo link station

[nextel]

Hello everybody,

a power failure in our office caused a problem on our nas.

It seems that at least one of two disks in the device is failing....

I removed both disks from the device in order to make at least an image, one has few bad sector, the other one is not accessible.

Looking in the good-one I saw a partition table as follow:

Number  Start        End            Size           File system     Name     Flags
 1      1048576B     1025507327B    1024458752B    ext3            primary
 2      1025507328B  6145703935B    5120196608B    xfs             primary
 3      6145703936B  6146752511B    1048576B                       primary
 4      6146752512B  6147801087B    1048576B                       primary
 5      6147801088B  7172259839B    1024458752B    linux-swap(v1)  primary
 6      7172259840B  492004259839B  484832000000B                  primary

 

Hexdumping the partition 6 I saw some of my data in it, but I can't recognize any known (by me) file system; I wonder If someone could give me some hint to get me move from my hanging point.

 

any help would be appreciated.

[franbour]

Try with UFS Explorer ! It gave me a good help a few weeks ago ...

[ESCHER]

You can also try using your favorite Linux distro, it is free.

Partition 6 is a RAID component, XFS formatted.

You need the package mdadm.

If you your NAS was running RAID0, you need to mount both drives.

If RAID1, theoretically you have two chances, try one disk first.

If "normal" then mount one disk first and then the other.

[Jotin]

Escher... have you ever done this? Would you be willing to take the time to post a youtube video here so everyone can see what you've done? Or could you post a link to the software and a walk through?

[nextel]

Thank you for reply.

after almost successful imaged the two disks I discovered the nas was set up by mistake as raid0. both disks has several errors and it took to fully image one day work. Now I managed to have the 2 disks image. I'm not confident on the sectors allocation in raid disks, but on the image of the second disk, data partition begins with 64kb of zeros, my fear is that I lost the super block on this partition.

the following are tho operation I did:
disk1.img image of disk 1

disk2.img image of disk 2

 

# parted disk1.img
GNU Parted 2.3
Using /mnt/disk1.img
Welcome to GNU Parted! Type 'help' to view a list of commands.
(parted) u
Unit?  [compact]? b
(parted) p
Model:  (file)
Disk /mnt/disk1.img: 500107862016B
Sector size (logical/physical): 512B/512B
Partition Table: gpt

Number  Start        End            Size           File system     Name     Flags
 1      1048576B     1025507327B    1024458752B    ext3            primary
 2      1025507328B  6145703935B    5120196608B    xfs             primary
 3      6145703936B  6146752511B    1048576B                       primary
 4      6146752512B  6147801087B    1048576B                       primary
 5      6147801088B  7172259839B    1024458752B    linux-swap(v1)  primary
 6      7172259840B  492004259839B  484832000000B                  primary

(parted) q

# parted disk2.img
GNU Parted 2.3
Using /mnt/disk2.img
Welcome to GNU Parted! Type 'help' to view a list of commands.
(parted) u
Unit?  [compact]? b
(parted) p
Model:  (file)
Disk /mnt/disk2.img: 500107862016B
Sector size (logical/physical): 512B/512B
Partition Table: gpt

Number  Start        End            Size           File system     Name     Flags
 1      1048576B     1025507327B    1024458752B                    primary
 2      1025507328B  6145703935B    5120196608B                    primary
 3      6145703936B  6146752511B    1048576B                       primary
 4      6146752512B  6147801087B    1048576B                       primary
 5      6147801088B  7172259839B    1024458752B    linux-swap(v1)  primary
 6      7172259840B  492004259839B  484832000000B                  primary

(parted) q

# losetup --offset 7172259840 /dev/loop3 /mnt/disk1.img

# losetup --offset 7172259840 /dev/loop4 /mnt/disk2.img

# mdadm -A /dev/md0 /dev/loop3 /dev/loop4
mdadm: no recogniseable superblock on /dev/loop3
mdadm: /dev/loop3 has no superblock - assembly aborted

 

looking into the partitions I observe the following:

# cat /dev/loop3 | hexdump -C | more
00000000  49 4e 41 ff 01 01 00 07  00 00 00 00 00 00 00 00  |INA.............|
00000010  00 00 00 07 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00000020  4d ca 24 d5 28 7f db e0  4e e6 10 d3 23 9b 2f f1  |M.$.(...N...#./.|
00000030  4e e6 10 d3 23 9b 2f f1  00 00 00 00 00 00 00 58  |N...#./........X|
....

 

root@ubuntu:/mnt# cat /dev/loop4 | hexdump -C | more
00000000  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
*
00010000  64 53 74 61 6d 70 0d 0a  23 65 6c 73 65 0d 0a 23  |dStamp..#else..#|
....

 

which means that at the beginning of data partition in disk 2 image contains 64kb of zeros

 

could someone confirm me I lost the superblock of disk 2 or point me errors I made to reach this conclusion?

Again, any suggestion would be appreciated.

Alessandro

 

 

[ESCHER]

Maybe you can post your questions in this forum:

http://forum.buffalo.nas-central.org/

Your skills are above average for the members of this forum.

[ESCHER]

Hi Jotin,

 

Mounting Linkstation drives in a Linux computer is surprisingly easy.

For LS-CHL, WVL or XHL just attach the drive and boot.  Partitions will be automatically mounted.

 

For LS-WXL I have just successfully tested with worst case scenario RAID0 (using spare 320 Gb drives). 

Should work also for any other model.

 

Hardware: Old desktop computer with two free SATA ports (USB ports should work also).

Software:  Ubuntu Linux 11.10 64 bit, make sure to have package mdadm installed. (32 bit version should work also).

 

Procedure:

Remove the drives from the LS and connect to the PC, boot.

Don't know why, but you get a blank screen, reboot (Alt+Ctr+Del).

You get a text screen asking if you want to mount the degraded RAID array, answer yes.

Linux boots normally and you have partition 6 mounted properly, strangely, the other partitions mount in degraded mode:

 

output form command "cat /proc/mdstat":

Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10]
md124 : active raid0 sdc6[0] sdb6[1]
      595372032 blocks super 1.2 512k chunks
      
md125 : active (auto-read-only) raid1 sdb2[0]
      4999156 blocks super 1.2 [2/1] [U_]
      
md126 : active (auto-read-only) raid1 sdb5[0] sdc5[2] sdc2[3](S)
      1000436 blocks super 1.2 [2/2] [UU]
      
md127 : active (auto-read-only) raid1 sdb1[0]
      1000384 blocks [2/1] [U_]
      
unused devices: <none>

 

For sure, this is not the most elegant procedure, but really easy.  RAID1 and "normal" mode work as well.  I would mount one drive first.   "normal mode" partitions are also recognized as RAID components.

 

It is also possible to hot swap the drives and the partitions mount automatically.

 

[Jotin]

Thanks Escher. Now if anyone asks I can just point them to this thread. I'll give it a try myself on a test unit and see what I get.

[wmalyoh]

I also had a power failure.  6 flashes.

I can't see the unit with NAS navigator, but see the MAC address through my router (without an IP assigned). 

 

Appears the controller card on the unit is bad.

Attatched the drives to PC, see healthy partitions.

 

Plan:

1. I have an identical NAS, I am going to attempt connecting the drives to this, long enough to get my data.

if that does not work.

2. use the method described by another poster, attaching to linux.

 

[axe]

Here some more info how to read / backup a whole drive, even without having a linux installation.  This is the most easiest way.

 

What do you need:

 

1. Knoppix Live linux image -> http://knopper.net/knoppix

2. Optional, but makes live even more easier:  a SATA-USB adapter (costs about 15 euro)

 

Howto:

 

1. Download Knoppix image and burn on CD

2. Remove hard drive from Linkstation

3. Connect harddrive to your computer using the SATA-USB Adapter  (if you do not have adapter, you have to install the drive inside your computer using SATA-connector)

4. Boof from Knoppix CD

5. If Linux is completely started, click in icon in left corner (there is only one icon on desktop, comparable to "workplace" in windows)

6. There you will find alle partitions of your harddrives.  Just open each, if you have the right one, you will see your Shares you had setup on your linkstation.

 

Now you can copy all data to a external backup drive by drag and drop.  External backup drive can e.g. be a USB hard drive.  Just connect to your computer, Knoppix will automatically mount it.

 

 

[erpilot]

I looked over the items in this thread, but none worked for me. I used this program:

http://www.ufsexplorer.com/download_stdr.php

and was able to retrieve all my files very easily. It cost me $55, but it was very simple to execute.  I tried the knoppix thing listed elsewhere and had zero success.

 

Seems like the problem wasn't the hard drive, but maybe the circuit board dealing with the power. I don't know, but this drive wasn't very good and didn't last very long.

[Chrishannah8386]

i have a very simlar issue i've tried to do most things above including

 

UFS - took 24 hours to read the drives and then gave me almost nothing all the files it found were inacessable

 

Ubuntu RAID mount - this did not work gave me errors regarding unable to mount the raid from the 6th partition. I did manage to mount the other smaller partitions and see the files there in typical luck.....

 

Knoppix - i tried to load this from a DVD however this did not work it just kinda waited at a dark black screen and didnt do anything.

 

i've tried cloneing the 6th partion onto a fresh drive that was already set up in a linkstation but this did not work i'm not clueless and lost and would appreciate some help? the linkstation basically said the raid was not configured

 

[businesslegions]

If anyone is interested, I managed to get my files back by following these instructions: Files Recovery for Failed Buffalo Hard Disk