I recently read a brilliant article about the guy that wrote the original Prince of Persia game for the Apple II in the 80’s. He had long since lost the original source code, until an old box of floppy disks was uncovered in his father’s apartment.
Sensibly, he enlisted some experts to help with the data extraction, and after a day of collaboration was able to release the source code online.
This got me thinking. Although I am meticulous with my backups of current data, I still have boxes of old software on floppy disks, which are happily degrading as we speak. John even found an old 212.6MB hard drive with some vintage data recovery software on it. Now this stuff isn’t always useful, but occasionally a really old drive comes down to us, and it is only this old software that can do the job.
As a result, John and I have started a project to get all of our old data recovery software from floppy disks and hard drives, and back it up to our file server. The 212.6 MB hard drive in the picture had 128MB of old DOS recovery software, which would easily fit on my mobile phone. Who knows when we might need it, but we now have it available when the need arises.
The Register has today posted two articles about the ongoing battle to expand hard drive capacities.
First is an actual device for sale, a 2TB Western Digital portable drive. This drive has a fancy new case and USB3 connection. It contains backup software and also the option to encrypt the data with a password. I wonder if it encrypts the data by default like some of their previous portables. (A bad thing!)
Second is a futuristic announcement from Seagate about their new HAMR technology. This new tech uses a laser to heat part of the disk before magnetising it. This apparently allows for much higher densities, theoretically paving the way for 60TB hard drives. There doesn’t appear to be any products using this technology at the moment.
60TB drives will be fantastic for backups, but horrible to backup without a new, faster form of connection. These would take almost forever (exaggeration) to fill up by SATA.
This news helps prove that hard drives are far from dead. It will take a long time until SSDs can cope with such massive capacities, at a similar cost to these beasts.
The Register today reported of expected supply problems for Western Digital, due to the severe flooding in Thailand at the moment. We are already having problems getting hold of certain hard drives, and this is sure to make the situation worse.
Today’s large SATA drives shouldn’t be used in 4 drive RAID 5 arrays due to the high likelihood of a read error after a drive failure, which will abort the RAID rebuild.
It is a common misconception that if you run a RAID system then you can avoid keeping backups. Although fault tolerant to a point, there are plenty of issues with RAIDs that can at best cause lengthy downtime and at worst prevent any recovery at all.
The Register reports that Hitachi have beaten Seagate to market with 1TB per platter drives. These will apparently pave the way for 4TB drives, however are currently limited to single platter options up to 1TB. I hope nobody tries to make a RAID 5 with a bunch of 4TB DRIVES. That would be asking for trouble.
A new 3TB drive is looming on the horizon, yet this may not be the breakthrough it seems. There appears to be major problems in the way older versions of Windows handle drives above 2.1TB. Windows XP and below will be unable to make use of these drives in any meaningful way, with some reports suggesting that only 990MB of usable capacity would be available to these vintage operating systems. It sounds like one more nail in XP’s coffin. If you need this mountain of storage it seems like it’s time to ditch XP already!
I will start by saying that we can, and have, recovered data from these drives. As with all RAID recovery we carry out, we never use the original controller or drives to access the data. We image the drives on an individual basis and then work on these drives to rebuild the RAID using a form of RAID emulation. The most time consuming part of this type of recovery is discovering the RAID settings that the manufacturer has used.
Due to the complex nature of these NAS devices, we always recommend they are backed up to another form of media. Although recovery is possible, it can be expensive and relatively time consuming. Also if the unit is powered on and RAID rebuilds are attempted then the recovery can be made more complicated or even impossible.
A Little Background
The LaCie 5Big NAS device contains five hard drives and allows for a number of different configurations. One of these configurations is RAID 6, which works in much the same way as RAID 5, but with an additional parity stripe. This additional parity stripe uses an algorithm which requires a relatively high processing overhead, so RAID 6 has rarely been seen in consumer level devices. The low cost of processing means that RAID 6 is now a viable option for embedded NAS devices such as the LaCie 5Big.
The advantage of RAID 6 over RAID 5 is that with 5 disks it can theoretically cope with two disk failures. This gets round an increasingly common problem with large capacity RAID 5 arrays where if a second disk fails whilst rebuilding a failed disk, there should be enough parity information to continue the rebuild successfully.
It is important to remember that due to the complex nature of RAID arrays, it is crucial to make backups of the data to a different type of storage.
If you have a failed or broken LaCie 5Big then use the contact details on the right to contact us. Alternatively you can leave a comment here and we will get back to you.
We have recently started to see a peak in the number of 500GB WD drives sent to us from Western Digital MyBook World edition NAS Servers. These NAS devices use two WD5000AAKS drives in a RAID configuration.
The worrying thing for users is that even drives used in RAID 1 mirror mode are having problems, where either both drives are failing at the same time or where one drive has failed in the past and the other following suit some time later.
It is an important reminder that RAID does not equal backup. If there is important data on these types of devices, it also needs to be copied to another device for peace of mind.
We are currently able to recover the data from these devices using a combination of firmware repairs and other recovery methods.
An interesting note is that we have also seen a small increase in other WD drives at the moment, such as the 320GB WD3200AAJS drives from iMacs and WD5000AACS drives from external enclosures.
Over the years we have seen stacks of LaCie Big Disks (and quite a few LaCie Bigger Disks too). With their designer looks and an abundance of different interfaces, it is no surprise that they are so popular. There is however one main reason that people send us their drives, and that is because they are not working.
Those of you familiar with hard drives will be aware that these LaCie drives are substantially larger in physical size and storage capacity when compared to a standalone hard disk drive. The reason being is that the LaCie Big Disk contains two hard disks (and the LaCie Bigger Disk contains four hard drives). Of course when you attach the drive to your computer you only see one volume. This is due to a RAID controller inside the LaCie drive which allows for multiple disks to appear as one large, usable disk. The main advantage to this setup is that read and write speeds can be very fast, as the reads and writes are spread over multiple disks. Another bonus is that the capacity of the volume is as large as the two drives. So two 500GB drives will give you a 1TB volume. There is however a massive downside to all of this clever RAID business which boils down to some simple mathematics.
The larger the number of hard disk drives used in this particular striped (RAID 0) setup, the more chance that one of them will fail, therefore a higher chance that you will lose all of the data stored on these drives. The way a RAID 0 stripe works is to distribute the data across the disks at block level. The amount of blocks used for each stripe is determined by the RAID controller and varies between different manufacturers. What this means, from a data recovery point of view is that in order to extract data from a LaCie Big Disk, you need to figure out the block size, and read the data from both drives in sequence, in order to extract usable data. This sounds a lot more complicated than it actually is and is often explained with the analogy of a filing cabinet.
Imagine the LaCie Big Disk is a filing cabinet. The cabinet has two drawers (disks) with an index in the first drawer. Now imagine that when you save a file into the cabinet, all of the odd numbered pages are put into the first drawer, and all of the even numbered pages are put into the second drawer. Once saved, the files location is stored in the index. In order to read back the saved file, it must first be collected from the relevant drawers, one page from each drawer at a time, and arranged into the correct order.
This is all great until a problem occurs. What happens when one of the drawers becomes damaged and can no longer be opened. Sure you can access all of the pages in the other drawer but having every other page is not much use to anybody! So until you can access both drawers, the documents are worthless.
Special Hard Drives?
The hard drives in a LaCie Big Disk are much the same as any other hard drive on the market. This means they are just as likely to have the same failures as a standalone drive. They suffer from electronic problems on the PCB, firmware corruption and also internal component failure and head crashes. These are problems which need to be overcome before any attempt at a recovery is even possible.
Aside from the usual hard drive problems, we have also seen other problems such as failed power supplies or damaged circuit boards within the LaCie Big Disk, which prevent access to the stored data.
This is why we have spent a lot of time researching the process of recovering data from RAID systems such as the LaCie Big Disk. We follow the same precautions with RAID as we do with single volume drives.
We make binary images of all of the individual hard drives on a read-only basis to protect against drive failure.
We don’t use the original hardware to read the RAID data, as this may be part of the problem.
We never write the recovered data back to the LaCie Big Disk, as this would prevent any further recovery process if it was required.
As with all data recovery, the most important thing to remember is that any attempts to access the data without following strict precautions could result in the data being either lost forever, or extremely expensive to get back.
The best advice is to keep these drives backed up as regularly as possible to avoid future headaches.
It’s something we should all be doing but never seems important until it’s too late. I’m not talking about taking the dog for a walk or feeding the cat, I’m talking about backing up your PC. In the words of Joni Mitchell “You don’t know what you’ve got ’till it’s gone.” How would you feel if you never saw your data again. Family photos, years worth of e-mails, documents, music and videos all gone in the blink of an eye. This is usually where we come in with our data recovery process. But there is an alternative. Lifehacker has an excellent guide on using free software to backup your PC. The only prerequisite is that you purchase an external hard disk of sufficient storage capacity.
For Mac users there is a totally different process. If you are running Leopard (10.5) then take a look at Time Machine. (More on this in a future post)