Anyone know how to recover crashed Linux...

Gimpy

Gimpy

The Living Force
...RAID 5 array...in specific the use of the mdadm command?

Background: we had a drive crash here that wiped out years worth of my photographs/artworks etc. Out of the blue. Hubby swore up and down to me this would never happen, but he did not ask me and put those things on this machine anyway. Now all is poofed, I'm crying my eyes out (but attempting to move on) and he's trying to find a way to restore what may be restorable.

I have no idea what I've typed here, I'm a luddite when it comes to computers, but he's at the 'grasp at straws' point. :rolleyes:

Hubby asks: "Should I use --re/add ? "
 
Gimpy said:
...RAID 5 array...in specific the use of the mdadm command?

Background: we had a drive crash here that wiped out years worth of my photographs/artworks etc. Out of the blue. Hubby swore up and down to me this would never happen, but he did not ask me and put those things on this machine anyway. Now all is poofed, I'm crying my eyes out (but attempting to move on) and he's trying to find a way to restore what may be restorable.

I have no idea what I've typed here, I'm a luddite when it comes to computers, but he's at the 'grasp at straws' point. :rolleyes:

Hubby asks: "Should I use --re/add ? "
Click to expand...

From Hubby:
HI, this is "Hubby". It was a Synology NAS. It was 4, 1TB Hard drives configures as a RAID 5 array. Problem is, and I didn't know this, three of the drives were the infamous Seagate SD15 firmware drives. And the SD15 firmware bug decided to manifest itself on two of the drives simultaneously after a power outage.
A RAID5 can recover with the failure of one drive, but not two. I think we are screwed. The only gleam of hope is that I upgraded the firmware in the drives, and they might be intact. But, I have to figure out how to get them re-joined to the RAID 5 Volume.
I am reaching out to the forum because I believe that some members have a high degree of IT skill, and may be able to help.
Click to expand...

I have no idea what any of the above means. :headbash: :nuts:
 
I recently recovered a failed RAID 0 array with ReclaiMe RAID Recovery Software _http://www.freeraidrecovery.com/ . It runs on Windows, though. After the successful recovery I also bought a subscription to BackBlaze to backup my important data online. A RAID 5 array is not a good backup solution, it's a good "availability" solution for servers.

So, I don't know about recovery softwares or methods under Linux, maybe someone else on the forum will. But if you have a spare hard drive on which you can install any version of Windows, you might be able to run the ReclaiMe software.

Good luck Gimpy :hug:
 
I can't help with the restoration of the disk station, as I don't understand much in terms of RAID architecture. Maybe this may not be possible, if two drives have failed.

The options that you have as far as I can see is:


If you cannot access the drives anymore buy a SATA to USB adapter (around 30 USD) and hook each failed drive directly to a Linux machine and mount it. The reason for using Linux is that you can tweak the dead disks much more than with eiither Wndows or Mac OS. Then you extract the files from the dead disk with some rescue software unto another disk. If the files are large this might not work in a RAID5 environment, but if all else fails it might be worth a try.


There are many powerful and free Linux HD recovery tools and many ressources on the net. Here are just a few, you might want to google to find others:
_http://www.cgsecurity.org/wiki/TestDisk
_http://www.linuxforu.com/2012/05/how-to-recover-deleted-files-linux-1/
_http://www.debianadmin.com/recover-data-from-a-dead-hard-drive-using-ddrescue.html
_http://www.gnu.org/software/ddrescue/ddrescue.html


There are also a few free and commercial application able to recover RAIDs. You may have to mount each of the four drives separately onto one machine:
_http://raid-recovery-guide.com/raid-recovery-comparison.aspx
_http://z-a-recovery.com/download.htm


The other thinig is to hire a professional. I did it once after having tried for two weeks in vain to fix a problem, the guy did it in one and a half hours - money well spent.

You may also want to contact Synology which has a recovery service, I believe.

Not sure i was of any help ... Good luck!
 
Thanks guys! Any and all help would be great. :flowers:

Not getting my hopes up, though. Its that whole cautious realist thing....
 
Oh, and I am now probably preaching to the converted ... NEVER EVER backup valuable data on only one device (I have 4, and one is off-site) ...
 
I don't know of a good answer for this one--having two drives completely fail at the same time is generally pretty rare (depending on the circumstances) and I've never tried to rebuild a RAID array before, but here are a few links that seem promising:

_http://serverfault.com/questions/183183/can-i-recover-a-raid-5-array-if-two-drives-have-failed
_http://unix.stackexchange.com/questions/8861/how-to-recover-a-crashed-linux-md-raid5-array
_http://blog.al4.co.nz/2011/03/recovering-a-raid5-mdadm-array-with-two-failed-devices/

The last one seems to suggest that you might be able to recover some data with:

mdadm –assemble –force
Click to expand...

however, please don't run that command until you read through all the links as it could put your data in an even worse position than it is now. Do you have space on another drive to write recovery data to? You might need enough storage to exactly copy (probably using ddrescue) the RAID array to another disk.

Unfortunately, I think the best case scenario you're looking at is at least some data loss, if not total data loss. But hopefully I'm wrong and good luck in the data salvaging! :flowers:
 
First thing I'll do is going to the Synology.com homepage and look for support.

General rules dictates that all the drives on any RAID must be exactly the same: model, size, buffer, brand, firmware, etc... exactly the same, and there must be list of supported drives too somewhere, given by the producer.

This one may help a bit too:

_http://forum.synology.com/wiki/index.php/How_to_retrieve_data_from_RAID_Volumes_on_Linux
 
Running the -force attribute is very dangerous - only do that if you a) know what you are doing and b) there is no other choice left. It could permanently destroy the RAID architecture ...
 
With your machine down perhaps you have not had much time to do much checking on the various internet tech forums. It looks like they were talking about this in 2009 . There seems to be some helpful discussion here:
 
sorry, the "Quick Reply" page posted itself somehow :-(

A discusion here: _forums.seagate.com/t5/Barracuda-XT-Barracuda-Barracuda/ST31000340AS-firmware-sd15/m-p/20128
offers some hope I think. Oh no, further on the suggested firmware upgrade gets pulled by Seagate (I wouldn't have posted this extra noise if the previous post hadn't sent itself off)
If your disks are spinning up and not making horrible noises then the data is still there on them, if that is any consolation. It then becomes a matter of how much money you can afford to get it off again.
Don't try using those `force' options though, not while your data is lost.
 
The -force option is certainly a last resort, but if the two options are have no data due to a broken RAID set or potentially further break the data that you're unable to access (but with the potential for partial recovery), then I think -force is worth a try before throwing the drives in the garbage. Depending on how much you want that data, it may be worth taking it to someone with expertise in the area (recovering the data, not necessarily your locality), but that may be expensive.
 
Ohhh, not fun at all...

In my experience...

If you have very valuable data you want to recover, I would
try a data recovery service but be expected to pay a lot of
money for it, with no guarantee of success. This may or may
not be an option for you depending on the circumstances.

Chances are probably very remote for linux OSes. If you decide
to so this, then obviously do not DO ANYTHING to your system
in attempts to "repair" or "recover" the data yourself!

Review and research on data-recovery companies, and inquire to
your particular setup (hardware & software) to make sure that
they can do it for you and make sure that you talk to at least two
or more independent companies so that you have an idea what it
will cost you and your time. Make sure their reputation proceeds
them.

I have lost many drives myself, due to various reasons (entropy,
stupidity, brownouts, "acts of God", bad vendors, bugs, ...), so
I never hold data backups on a single drive but on multiple drives
locally, remote, and offline. I guess I have a thing with triplets ;)
But seriously, one (offline) drive is NEVER used except when the
other two drives are updated successfully, working for some time,
before attempting the update the third drive and then offline it.

So, yeah, paranoid perhaps, but I do not underestimate entropy.
To do so is... (Taa Daaaa!) Wishful thinking (or Ignorance). :/

Keep in mind that physical media (HD/CD/DVD/Tapes/...) and
systems, all deteriorate over time, some faster than others
depending on the situation.

Oh, yes, backups! I cannot strongly enough stress the importance
of backups! Had you done backups properly, we wouldn't be having
this discussion now, would we?

I use Acronis and for the most part is inexpensive for simple home
OSes, but once one moves into RAID, Server systems, Databases
and so on, Acronis can get to be expensive. But all in all it is a great
product.
 
Yes, I looked around some more. The problem is not with the raid so leave it alone.

You need to get a firmware upgrade into a broken disk but it is stuck in a permanent busy state. You can't see it from the bios and so you can't get at it to do the firmware upgrade. Is that the sort of thing you are seeing?

The fix(!) is to get at the drive through a serial connection and clear the BSY state, then do the firmware upgrade, then put your drive back into the raid. (Then get the data off, ... then junk the disks ). Making the serial connection involves a bit of messing around with wires so you might want to pay someone to do it if you are really not confident.
The data should not have been affected by these operations (until you junk the unwanted disks that is) so all should be well and your photos will be back.

There were some posts on the net that seemed to say seagate would do this for you but I guess that was for disks under warranty, and the posts were all dated 2009/10

Remember you have two bust drives but you only need one to get the raid back up so you can even completely trash one drive while fiddling about with it and still recover. You see, that is good news :-)

If you check some of these links I think you will see what I mean.
Good luck.

_http://hackaday.com/2012/07/30/recovering-from-a-seagate-hdd-firmware-bug/
_http://www.msfn.org/board/topic/128807-the-solution-for-seagate-720011-hdds/
_http://laptoprepair.com.ph/seagate-barracuda-7200-11-high-failure-rate-solved
_http://www.geekzone.co.nz/forums.asp?forumid=50&topicid=60713

Seagates firmware upgrades are linked here (they are very dependant on serial numbers)
_http://knowledge.seagate.com/articles/en_US/FAQ/207931en/?language=en_US&fs=Search&pn=1
 
OK, a bit more. Re-reading your post, I realised I missed the point that you have in fact put the new firmware on, and now can't join the array. So it is no longer the SD15 bug but something to do with the array and mdadm.
Googling "mdadm raid 5 recovery" should definitely help you over this. 3 or 4 promising discussions pop up:
_http://blog.al4.co.nz/2011/03/recovering-a-raid5-mdadm-array-with-two-failed-devices/
_http://kevin.deldycke.com/2008/07/heroic-journey-to-raid-5-data-recovery/
_http://www.mysqlperformanceblog.com/2011/08/27/recovering-linux-software-raid-raid5-array/

I've had to run quit a lot of disks under software raid before now, but I have never used mdadm so can't help with the precise details. If the disks are really intact then your data is not going anywhere and it is worth getting clear about the principles of what you need to do and what the software is trying to do, then you can be more confident about running those scary --force commands.
There is a lot of know-how out there and a lot of information on the tech sites even though you can not accept it all at face value. Posters don't always know as much as they like to think they do, or they answer the wrong question (ahem).
I am sure you are will be all right. Don't despair now :-)
 
You must log in or register to reply here.

Trending content

Back
Top Bottom