DIY Computers

Home >> Data Recovery >> Data Recovery after a Hard Disk Failure

Data Recovery from a Failed Hard Disk

Types of Hard Drive Failure

We can broadly classify hard drive failures into two categories. The type of hard disk failure that almost always springs to mind first is a mechanical or electronic failure of the disk, where the disk itself is physically damaged seemingly beyond repair through mistreatment or wear and tear. Usually this type of failure is attributed to defective components, specifically the hard drives platters, motor, read/write heads or the printed circuit board (controller).

The other type of damage is logical damage, referring to damage of the file system itself. Logical damage is most commonly found after power failures or similar events that results in the filesystem being inconsistant or not completely written. Problems with drivers, RAID controllers and system crashes can all also contribute to logical damage.

Recovering Data from a drive that has suffered Logical Damage

Most logical damage can be repaired by using a program that checks the disk for consistancy and makes repairs based on the inconsistancies it finds. Chkdsk for Windows, Disk First Aid and Disk utility for Mac and fsck for unix and linux are examples of programs that check disk consistancy and repair file systems. Most operating systems have some sort of disk consistancy checking program built in. Both fsck and Chkdsk will run automatically on boot after the computer has been shutdown abruptly such as the result of a power failure.

Care must be taken when using consistancy checking programs as they can occasionally delete files and folders that they erroneously feel are out of place or unexplainable. These files more often than not turn out to be user files rather than system files and it is often a good idea to use a disk imaging program such as Acronis True Image or Nortons Ghost to take a block by block image of the disk before checking it for consistancy. Data loss as a result of running a consistancy checking program is certainly a rare thing and more of an indicator of extensive file system problems. I have used chkdsk and fsck extensively over the years and have not lost any files yet as a result of running disk consistancy checks however the possibility of data loss exists.

If the file system is extensively damaged you have the option of contacting a professional data recovery service that performs Data Carving, a technique used to extract lost data from a disk that has no file allocation information relating to the file by identifying sectors and clusters that the file is stored on.

Recovering Data from a drive that has suffered Physical Damage

A hard disk drive that has been damaged physically by either mistreatment or wear and tear is far more challenging to recover data from. Unfortunately there are very few mechanical or electronic failures that a hard drive can suffer that can be repaired by the end user. I strongly urge anyone who has critical data that must be recovered from a drive to contact a data recovery service and have them recover the data.

When a hard disk drive or other types of storage media are manufactured it is in a clean and dust free environment and for good reason. The read/write head for example misses the disk platter by only a few thousandths of an inch during operation. A single speck of dust can do some serious damage with such fine clearances, particularly when you consider that the platter may be spinning anywhere between 5200rpm and 20,000rpm. If you attempt to replace the read/write head, unless you have a clean room, you are unlikely to assemble the disk without getting dust into it. Dust is absolutely everywhere.

So without a cleanroom we cannot perform surgery on the insides of our hard disk, in reality if we did we would probably loose more data, if not all of the data that remains. That really only leaves the PCB that contains the controller and cache for the hard drive. If we have an exactly identical hard drive the PCB can be swapped between them, this may actually resolve your problems if the computer does not recognise that the hard disk is attached to it when it boots (especially if you hear the hard drive spin up), it may well be the controller that has failed in this instance. Well worth a try if failure to recover the data is an option.

Often the damage on the drive is limited to a few bad data blocks. This may be enough to stop the operating system from functioning correctly or the machine might boot with bad media errors when running Chkdsk or fsck. A drive with only a few bad blocks on it can usually be mounted in a second computer so that the data may be recovered from it. I recently had a linux mail server become unresponsive and when it was brought into the workshop it reported several bad blocks on boot and many other dire warnings of impending total disk failure. The drive was mounted into a Windows XP computer as a secondary drive and an attempt to make a disk image was made with Acronis True Image. The imaging process was slow as the host computer had to make several attempts to read the bad blocks but only reported one as completely unreadable. The error message reporting the unreadable block asked if I wanted to continue and I clicked Yes to allow the image to be completed. The completed disk image was then put onto a brand new disk and installed in the server. Upon booting off the new disk the machine ran fsck, repaired the missing block and the server then booted perfectly. 20 minutes later the customers were collecting their emails again.

This story serves to illustrate how data and indeed entire hard disks can be recovered off disks that are close to the point of complete failure. If you suspect a disk is defective act swiftly to back up your data and effect repair wherever practical. I have recovered many entire operating systems using disk imaging software, good timing and perhaps a little luck too.

Often when a hard drive completely fails you may hear a ticking noise coming from the drive at regular intervals and this is usually the read/write head hitting the platter or perhaps other components inside of the hard disk. I must remark that a failed controller can cause this problem and replacing the controller with an identical one from the same model of disk can resolve this problem without the following drastic measure. If the BIOS of the computer sees the disk present and it doesnt spin up or spins up intermittantly and makes ticking noises, wrap the disk in a plastic bag or two and place it in the coldest freezer you have access to for about two days. After two days remove it from the freezer and plug it immediately into another computer as a secondary drive. If the machine can use the disk successfully it will probably only be for a short time so recover your data or image the disk as quickly as you can.