General NAS-Central Forums

Welcome to the NAS community
It is currently Fri Aug 18, 2017 12:41 pm

All times are UTC




Post new topic Reply to topic  [ 13 posts ] 
Author Message
PostPosted: Thu May 26, 2016 7:32 pm 
Offline

Joined: Thu May 26, 2016 7:05 pm
Posts: 7
I have two 7.5TB 5big units which have served me "well-ish" for several years backing up my iTunes library and work archives. Within the last year both drives have suffered what appears to be an internal hardware or software failure. The first unit started with what seemed like a bad drive. I replaced the drive and waited. It appeared to be rebuilding the new drive, then the entire NAS disappeared from my network. I can shut it down and wait a day and it will reappear for a short time, but never long enough to enable me to retrieve the files it holds. The second unit just randomly disappeared from my network a couple weeks ago, and none of the usual tricks (new power supply, connect directly to computer, etc) help at all. I'm not as worried about the first unit, as I'm confident the files are intact and can be retrieved with time and patience. The second unit, however, holds the archives of my work for the last 20 years many of which are probably not backed-up elsewhere.

So... long story short, what do you guys recommend for accessing and retrieving the files from the second unit which just fell off the radar? I've researched several theories and software options, but most are risky and/or expensive. Right now I see my top three options as:

1. Remove drives from the unit and connect directly to computer then use the Seagate recovery app

2. Remove drives from the unit and connect directly to computer then use Ubuntu to try and access the RAID.

3. Remove drives from the unit and connect directly to computer then use R-Studio to try and recover the RAID

Does anyone have any advice or other options I should look into?

Thanks in advance!

jv

Specs: iMac / El Capitan


Top
 Profile  
 
PostPosted: Fri May 27, 2016 10:10 am 
Offline

Joined: Mon Jun 16, 2008 10:45 am
Posts: 6022
I guess you have a hardware problem. As you already exchanged the powersupply, the most likely problem are bad capacitors.

Anyway, I'd give your 2nd option a higher rank. Any decent Linux distro will recognize and mount the raid array just fine. The biggest problem (in most cases) is to connect 5 disks to a PC. But you are allowed to mix up USB and SATA, if that is desirable.


Top
 Profile  
 
PostPosted: Thu Jun 02, 2016 7:45 am 
Offline

Joined: Thu May 26, 2016 7:05 pm
Posts: 7
Mijzelf,

Thanks for the pointers! I've been gearing up for this for a while, so I already had 5 SATA to USB adapters ready to go.

So, I got Ubuntu to recognize the RAID (it seems at least 4 of 5 drives are intact), but I can't figure out how to mount it. I've scoured the Internet, but can't seem to find a simple answer. My knowledge of Terminal commands is limited at best. Any pointers for that part of the process? I just need to mount the RAID in Ubuntu so I can see it on my network and transfer the files to a new server. Easy, right..?

Thanks again!


Top
 Profile  
 
PostPosted: Thu Jun 02, 2016 4:50 pm 
Offline

Joined: Mon Jun 16, 2008 10:45 am
Posts: 6022
Run (as root)
Code:
mdadm ---examine /dev/sd[abcdefg]2
to find out which partitions have a valid raid header. If you have found at least 4 of them, you can assemble the array with
Code:
mdadm --assemble /dev/md0 /dev/sd[abcd]2 --run
where you have to exchange abcd with the actual recognized partitions.

After assembling the array, you can mount it:
Code:
mkdir /tmp/mountpoint
mount -o ro /dev/md0 /tmp/mountpoint


Top
 Profile  
 
PostPosted: Fri Jun 03, 2016 1:29 am 
Offline

Joined: Thu May 26, 2016 7:05 pm
Posts: 7
Ok, so it seems to have assembled the RAID... GParted is reading all five drives and recognizing a "/dev/md123" which matches the size of the RAID at 5.45TB. When I try to mount it using the Terminal commands you gave me it says it can't read the superblock. GParted is also saying it can't determine the filesystem. I think we are close, do you have any more advice?


Top
 Profile  
 
PostPosted: Fri Jun 03, 2016 8:29 am 
Offline

Joined: Mon Jun 16, 2008 10:45 am
Posts: 6022
I *think* the filesystem is XFS, so you can try to specify it to the mount command:
Code:
mount -o ro -t xfs /dev/md123 /tmp/mountpoint


Top
 Profile  
 
PostPosted: Fri Jun 03, 2016 12:23 pm 
Offline

Joined: Thu May 26, 2016 7:05 pm
Posts: 7
So, before I got your last message there was a power outage in the middle of the night which outlasted my battery backup... Once I got everything up and running again this morning, I had lost the seemingly assembled array. Now when I try to assemble it again it tells me each of the drives are busy or already part of another array (Ex: found some drive for an array that is already active: dev/md3).

thanks again for your patience. Any further advice for this new issue?


Top
 Profile  
 
PostPosted: Fri Jun 03, 2016 1:22 pm 
Offline

Joined: Mon Jun 16, 2008 10:45 am
Posts: 6022
If you execute
Code:
cat /proc/partitions
you can see all 'block devices' (disks, partitions, raid arrays, ...), and see if a 5.45TB raid array is available. If yes, try to mount that explicitely as xfs.


Top
 Profile  
 
PostPosted: Mon Jun 06, 2016 3:45 pm 
Offline

Joined: Thu May 26, 2016 7:05 pm
Posts: 7
Since that power failure knocked everything offline I haven't been able to get Ubuntu to recognize the RAID as a whole again. I've tried everything you told me and a few other things I found on the net. Even though all 5 drives are visible in gParted and 4 of them test out ok with mdadm, it refuses to recreate the RAID. The one time I got close, it gave me the superblock error. Is there a way to recreate the superblock?

Unfortunately I'm getting close to just forking out the cash for R-Studio and see what I can rescue before another drive dies.

Mijzelf, I'm hoping you have a few more tricks up your sleeve.


Top
 Profile  
 
PostPosted: Tue Jun 07, 2016 8:38 am 
Offline

Joined: Mon Jun 16, 2008 10:45 am
Posts: 6022
If the raid array is assembled without errors, the raid array should contain a valid filesystem. Apparently it doesn't, hence the 'superblock error'.

I don't know if that can be recreated/repaired. I *think* the filesystem is xfs, as all older Lacie boxes use that. For xfs there is a tool 'xfs_repair' (part of the package 'xfs_tools') which might do the job.
But repairing a filesystem is a dangerous job. A damaged filesystem is in an undefined state, and so repairing might cause more damage, if it starts with wrong assumptions.
If your data is valuable, you should do a low level copy of the block device, and try to repair the copy. (If there are bad sectors on the original, chances are that it can't be repaired anyway, on that hardware). Unfortunately your volume is huge, so you'll need a 6TB disk (or a raid array) to contain the filesystem.

Of course you can try R-Studio (are you sure that is a data recovery tool? I don't read that on their website), but unless they wrote their own xfs implementation, I guess they use the engine Linux has, and so it can't mount the volume either.

Depending on the nature of your data PhotoRec could recover a lot. It can recognize about a 100 different file types on bare metal, without any help of the filesystem. Big drawback is that it doesn't recover metadata from that filesystem, so it recovers only the contents, not the name, directory, timestamp, ...


Top
 Profile  
 
PostPosted: Tue Jun 07, 2016 12:32 pm 
Offline

Joined: Thu May 26, 2016 7:05 pm
Posts: 7
I thought I had found a "cheaper" solution that seemed hopeful called ReclaiMe, but then I realized it only works on Windows and I only have Macs.

How do I go about making a low level copy? I assume this isn't done with Mac OS, as it only sees the drives as being mounted, and not a readable format. I have a new NAS with plenty of room just waiting to hold the disks' contents.

I'm still on the fence about R-Studio, and perhaps I need to do more research on it before spending the money.


Top
 Profile  
 
PostPosted: Tue Jun 07, 2016 1:50 pm 
Offline

Joined: Thu May 26, 2016 7:05 pm
Posts: 7
There are two pieces of software called R Studio. The one I'm looking at is this one:

http://www.r-studio.com/


Top
 Profile  
 
PostPosted: Tue Jun 07, 2016 7:54 pm 
Offline

Joined: Mon Jun 16, 2008 10:45 am
Posts: 6022
fuzzydisco wrote:
How do I go about making a low level copy?
Something like
Code:
dd bs=16M if=/dev/md0 of=/path/to/file
where /dev/md0 is the raid array node, and /path/to/file can be a disk or partition node, or a file (on a volume big enough). Make sure the if (InputFile) and of (OutputFile) are right. Dd will without asking overwrite your precious data if you copy in the wrong direction.

Because this is Linux, you are free to pipe it everywhere. For instance:
Code:
dd bs=16M if=/dev/md0 | gzip >/path/to/zipped/disk
will write the data to a compressed file.
Code:
dd bs=16M if=/dev/md0 | ssh user@nas "dd bs=16M of=/path/to/file"
will write the image to a file on a remote NAS, through ssh. Of course this will take a lot of time. (It will take a lot of time anyway). As the data is encrypted, you might have a higher throughput if you compress it first. (Encryption is expensive):
Code:
dd bs=16M if=/dev/md0 | gzip | ssh user@nas "cat | gzip -d | dd bs=16M of=/path/to/file"


fuzzydisco wrote:
There are two pieces of software called R Studio. The one I'm looking at is this one:

http://www.r-studio.com/
Ah. Indeed I was looking to the other piece. According to that website the tool doesn't support xfs, so I think it's useless for you.


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 13 posts ] 

All times are UTC


Who is online

Users browsing this forum: No registered users and 4 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB® Forum Software © phpBB Group