Seeking advice on a RAID failure

[neildawson]

The situation in short:

I had three 500GB hard drives in a 2008 Mac Pro in RAID0, one now fails to mount but sometimes shows in in disk utility. Very little was backed up.



The question:

Is it possible to simply repair or clone the failed hard drive myself?



The situation in full:

The hard drives are all 500GB Seagate Barracuda 7200.11, ST3500320AS. Bought refurbished from Dabs.com.

Each of the hard drives was was divided into a small and large partition, the small partitions were combined in Disk Utility (software RAID0) to create a 'scratch' volume for current work, the large partitions were combined to create a large storage volume.

The computer started making unusual noises, and these two volumes would occasionally unmount and disappear. I managed to backup my iPhoto library before it stopped mounting completely. One of the hard drives did not show up in Disk Utility. The other two are still recognised as having two RAID slices. At this point I stopped using the computer for many months, unsure what to do.

Early this year I tried again, the volumes would still not mount, but at least all three disks showed up in Disk Utility. I removed all three hard drives from the computer and stored them in the padded bags they were delivered in, to prevent further damage until I could figure out what to do.



The sob story:

I've called a data recovery agency and their price for this job starts at £975+VAT. I recently graduated from university, am doing unpaid work for my portfolio, and really can't afford that much money. The RAID has all of my university work since 2007 on it, as well as audio recordings of my Granpa which are very precious and irreplaceable. I unthinkingly deleted my iPhoto backup last week (believing it to be backed up somewhere else) taking with it all of my photos since 2007. I've learned my lesson about getting greedy for speed and not backing up, but I still hope to be able to get this stuff back.

 

[philipma1957]

Not good. I have never had a lot of luck when your problem has happened to me. I hope others on site can help you.

 

[mjsmke]

If the drives show up in disk utility. Maybe you could create an image from each drive, then restore to new drives.

 

[DeeEss]

OK, so failing to mount at all but showing up in Utility. Were the noises loud clicks/crunches or constant but maybe loud search and seek sounds?

it could be that the directory has corrupted in which case something like Disk Warrior can do a pretty good job of rebuilding it.

Although if it mounted and unmounted several times before it's likely to be hardware.

Have you tried swapping bays and seeing if it's a sata connection issue with your computer? if you have a spare drive put it in the effected bay and see what it does. Limit the use of the effected drive as much you can obvs.

You could try put the drive in an enclosure. I've had that work before too.

 

[philipma1957]

HE NEEDS them to stay in raid0. if he has a pci e card with an esata port like this.


http://eshop.macsales.com/item/Newer Technology/MXPCIE6GRS/

and a four drive external like this


http://www.newegg.com/Product/Product.aspx?Item=N82E16816111148&cm_re=4_bay-_-16-111-148-_-Product

he could put the three raid0 hdds in it. his mac pro may be able to "see" and read the raid0 set he could then try to clone it to a big 2tb hdd in his mac. the problem is he is in the uk if he was near me in The usa we could give it a try.
One more thing he needs to keep the osx in his mac pro somewhat close to the one he made the raid0 set on. I wonder if he had snow and upgraded to snow leopard on the mac pro osx. worse yet if he went from tiger to snow or snow leopard.

 

[Honumaui]

just a thought for setup next time

doing a partition on a drive and using that for scratch and the other for data is actually slower in PS and LR not sure about audio or video aps
but just so ya know better to just point it at the same drive unless you can dedicate a drive to scratch/cache only


sorry to hear but Backup backup backup

 

[neildawson]

Were the noises loud clicks/crunches or constant but maybe loud search and seek sounds?

If I remember right it was constant sound (at first I thought it was the fan) rather than clicks or crunches, and it would stop when the disk would unmount. This happened last summer so it's difficult to remember, if need be I can stick the drives back in and see how it goes.

Have you tried swapping bays and seeing if it's a sata connection issue with your computer? if you have a spare drive put it in the effected bay and see what it does. Limit the use of the effected drive as much you can obvs.

I hadn't thought of this, I'll try this with one of the unaffected drives and see if they show in Disk Utility. For future reference do you know if it matters which bay the drives are inserted in, i.e. in what order? I labelled them when I took them out just in case.

One more thing he needs to keep the osx in his mac pro somewhat close to the one he made the raid0 set on. I wonder if he had snow and upgraded to snow leopard on the mac pro osx. worse yet if he went from tiger to snow or snow leopard.

I keep the OS up to date, so I was definitely running Snow Leopard, though I'm not sure exactly which version. It was in in the early summer months of last year. The machine is now running 10.6.6.

just a thought for setup next time

doing a partition on a drive and using that for scratch and the other for data is actually slower in PS and LR not sure about audio or video aps
but just so ya know better to just point it at the same drive unless you can dedicate a drive to scratch/cache only

sorry to hear but Backup backup backup

That's good to know, there was a definite performance boost in Final Cut and iPhoto but I didn't notice anything in Photoshop.



I've searched a little bit on home repairs to hard drives. Could this issue be attributed to a faulty logic board? Replacing one seems to be a fairly simple, if risky matter, provided I can get my hands on a matching hard drive. At the time I was convinced by the noises it was making that something more fundamental and mechanical was faulty.

The machine has been in for repair twice at authorised service centres. In one case the engineer commented on the extra drives, saying I had not set up the S.M.A.R.T. system correctly.



Thank you all for your replies by the way. The first thing I will do if I recover this stuff is back it up to my external HD, and burn the really precious stuff to a CD!

 

[nanofrog]

I had three 500GB hard drives in a 2008 Mac Pro in RAID0, one now fails to mount but sometimes shows in in disk utility. Very little was backed up.

The first thing you need to understand, is how RAID 0 works (aka stripe set). Unlike a single disk, which contains the entire file, a stripe set splits files amongst all the disks in the set. So each disk only contails a pieces of files. All the disks have to be mounted and operational in order for the file to be strung together.

So if any disk is lost, so is the array, and all the data it contains. This is why a backup system is even more critical with RAID systems, a stripe set most of all.

Is it possible to simply repair or clone the failed hard drive myself?

If the disk controllers are fine, maybe, via software like DiskWarrior (already mentioned). It has the ability to recover data off of bad sectors, but is not perfect or fool-proof (there's a good chance that even if this is the case, you won't get it all back).

But if the controller's shot, NO. You'd need to send it into a Data Recovery service, as they'd disassemble the disk and either replace the controller, or even do a deep surface scan of the platters to get the data back. And as you've discovered, this isn't cheap.

The hard drives are all 500GB Seagate Barracuda 7200.11, ST3500320AS. Bought refurbished from Dabs.com.

These were bad disks to begin with (all sorts of information on the 7200.11 series that they were junk), so buying them refurbished was a really bad idea, I'm sorry to say (lesson = do not buy refurbished disks, no matter the brand or model number; data is too important to risk like this).

Combine this with the noises you mentioned, this lends me to think the disk is truly shot (nearly certain to me at this point).

Each of the hard drives was was divided into a small and large partition, the small partitions were combined in Disk Utility (software RAID0) to create a 'scratch' volume for current work, the large partitions were combined to create a large storage volume.

As mentioned, this is not a good idea.

The reason, is simultaneous access. That is, with multiple arrays on the same disks, particularly those that will be accessed at the same time, will actually slow you down. There's only one spindle, servo, and armature containing the heads, so depending on which read/write request got there first (queue), anything else has to wait it's turn before it can be executed.

 

[initialsBB]

Wirelessly posted (iPhone: Mozilla/5.0 (iPhone; U; CPU iPhone OS 4_2_1 like Mac OS X; en-us) AppleWebKit/533.17.9 (KHTML, like Gecko) Version/5.0.2 Mobile/8C148 Safari/6533.18.5)

This is a sad story. Hope you get back your data but as nanofrog says, you should first kiss that data goodbye, then see what may eventually be doable.

I had a friend who managed to recuperate a sizeable portion of data off a bad drive by popping it into a watertight pack and leaving it in a freezer for a few hours... It worked but it was last ditch attempt. You can use google translate to check this French forum for instructions : http://forum.macbidouille.com/index.php?showtopic=163946&st=0

I'd definitely start by trying the three drives in an external esata enclosure first.

 

[dknightd]

First, I hope you have learned from this experience and now have at least 2 backups of any data you think is important.

A constant noise to me suggests that the disk motor, or its bearings went bad.
A data recovery place can disassemble your disk, put the platters in a different mechanism, and might be able to recover everything from the disks.

The fact that the disk shows up in disk utility suggests the logic board (or at least part of it) is OK. The fact that the raid volume does not get mounted suggests the disk can not be read (the information about raid configuration is stored on the disk - if disk utility cannot read the disk, it cannot read the raid info so does not know what to do with it. BTW, yes, you can put the disks in any slot, and if all the disks are readable osx will recognize the raid volume. You can also put them in a different mac.

Because the disk made noise before it died it is not likely the sata port, or the controller, but it could be I suppose. Noise to me suggests a mechanical problem. A more or less constant noise suggests bearings or motor - this is good - it means the data on the platters might be intact (as opposed to a head crashing in the platter - which generally destroys part of the platter, and hence data).
You could try one of the various tricks (freezing, heating, tapping). But before you do, I'd think about how important the data is to you. You might have lost 3 years of your computer data. You've been without it for some time. You have many more years ahead of you. If you really really want that data back, I'd put the disks back in bags and start saving up for data recovery - and hope it works. If you want the data back sooner than that, and are wiling to risk loosing it, then try one of the tricks. $1000 might seem like alot of money now, but in 5 years it might be affordable if you really want that data. Of course by then you might not even miss the data. Only you can decide.

You have learned a painful lesson. It is better to learn this lesson while you are young. Too bad you had to learn it the hard way.

All hard disks die. It is much cheaper to keep backups, than it is to pay for data recovery (which is not always successful).

Good luck.

 

[ActionableMango]

Probably the data is fine just sitting there on the platters and either hardware or the controller board is bad.

The only thing I can think of on the cheap, with no guarantee of success, is to buy an *exactly* identical hard drive and try swapping parts.

The controller board will be the easiest to try first. Although there are mechanical sounds, perhaps a bad controller is causing the motor or some other mechanism to go haywire.

If that doesn't fix it, you can try swapping the platters.

Many times out of curiosity I've pulled apart hard drives, disassembled them, and even run them without cases. It won't be terribly hard to swap the platters, but you'll need all manner of strange torx bits, allen bits, and screwdriver bits, in all sorts of small sizes. It will take an afternoon or two and maybe a trip to the hardware store for tools.

One more thing, if you are lazy and run it without the cover, wear eye protection. We once had a part fling off at what seemed like 100mph and embed itself into the drywall.

Good luck.

 

[dime21]

you had valuable data and you chose not to back it up?? AND on top of that, you used RAID0? not to be a dick, but good grief man, you were just begging for this to happen. 3 member RAID0 and no backups is like spilling gasoline all over AND smoking a cigarette - the outcome is very predictable and should not be a surprise to anyone.

Your only chance of recovering any data is to pay the professionals. Their price is very reasonable - if you cannot afford it now, take all three members of your RAID0 and put them in a box!! Do not ever touch them until you have saved enough money to pay the professionals!! If you attempt anything yourself, you're likely to screw it up even worse, such that even the professionals cannot recover it.

backup backup backup backup backup your data! regularly! hard drives are so stupid cheap these days, the only reason to not backup your data, is that you don't care about losing it! hard drives are consumables - they always always always fail!! It's not a matter of 'if', it's a matter of 'when'. Some last longer than others, but ALL of them fail!!

 

[nanofrog]

Probably the data is fine just sitting there on the platters and either hardware or the controller board is bad.

Assuming the sectors are fine (or heads didn't nail the platters), it's possible the data still exists.

The only thing I can think of on the cheap, with no guarantee of success, is to buy an *exactly* identical hard drive and try swapping parts.

The controller board will be the easiest to try first. Although there are mechanical sounds, perhaps a bad controller is causing the motor or some other mechanism to go haywire.

If that doesn't fix it, you can try swapping the platters.

I've the impression the OP isn't familiar with this, so I didn't even attempt to mention this due to the risk. The controller board is doable, but if any of the mechanical parts are shot, it's really difficult. Specifically, the platter assembly cannot be disassembled (you cannot shift the platters from their original positions relative to one another). If you do, you'll loose the start point created by the original low level format, and won't be able to read it (it's possible to get it working again, but it would require another low level format, which will wipe any existing information).

Definitely risky, so if the data's that important, I'd say save up the funds, and let a professional Data Recovery service take a crack at it.

Good note about the safety issues BTW.

 

[mjsmke]

You said the drives show up in disk utility. So in disk utility click on the restore tab. select your drive as the source and a new drive of at least the same size as your data in the destination box.

Do this for all 3 drives (may take several hours) then try accessing the RAID volume.

 

[ActionableMango]

Specifically, the platter assembly cannot be disassembled (you cannot shift the platters from their original positions relative to one another). If you do, you'll loose the start point created by the original low level format, and won't be able to read it (it's possible to get it working again, but it would require another low level format, which will wipe any existing information).

That's a really good that I didn't know about. Whenever we messed around with opening drives and used them afterward, we were not at all concerned with data recovery. In this instance I suppose the best you could do is carefully swap entire assemblies.

A data recovery specialist would definitely be best. Personally, I've never had data worth the price, and I have twice lost a non-backed-up hard drive.