Guides with no-nonsense advice on bad drives, missing files and what you should do before you mess up, created by the engineers who fix these problems in-house all of the time.
If your drive is making clicking sounds or is failing, the very best thing to do right now is stop using it as quickly as possible and call us. The reason most people turn recoverable data into unrecoverable data is that they continue to use a failing drive.
Clicks, goes away, dies, is dead: the most common drive symptoms. What do you do about them?
Because the heads are unable to locate their servo data, the actuator is retracted and then tries again. The number of times the drive attempts to access information once it starts to click will determine whether or not the information can be recovered, more than anything we do.
Before your drive stops completely, it will give you warnings for weeks. The five warning signs in order of occurrence; the three SMART fields that you care about; and the only utility tool that you shouldn't be running.
We explain four diagnostic tests that don't cost anything and are performed progressively further out from your drive. We see that most enquiries made regarding this problem resolve at either the cable, port or the power supply.
Often it was your boot configuration that went wrong instead of your drive. How to figure out if it's the boot configuration or the drive itself, and why repair tools write to a disk you might still want.
Sometimes a bad connection or physical problems cause the checksum in the sectors that the drive reads to be incorrect. Chkdsk will make this problem worse.
The operating system (Windows) cannot read the file system. This does not mean the disk is completely empty; there are 4 reasons for this problem, and all are recoverable if you decline the prompt from Windows to format the disk.
This NTFS message means the Master File Table is corrupted; however, your files are not. It is generally a bad idea to let chkdsk fix this problem.
The index contains information about each file on the disk. Even with a damaged index, the content of those files often remains intact. However, chkdsk optimises for consistency across the entire disk volume rather than optimising for recovery of your individual files.
There are four types of faults with four different prospects and price tags: silent, whining, spinning-but-undetected and detected-but-unreadable.
Three SMART parameters provide useful information regarding potential issues related to a failing drive, and the overall health status isn't one of them. Most other parameters provide meaningless 'noise'.
When you can safely perform some DIY fixes, and how to retrieve lost/deleted files without exacerbating the damage.
When you delete something from your computer, all you've done is remove the reference/pointer that points to that location on the disk. If you're using a hard drive, then the deleted data is still sitting on the disk. If you're using an SSD, then the deleted data gets erased by the TRIM command within minutes. What not to install and where not to install it.
Commonly from the card in a camera; uncommonly from a phone. The reason why the source device decides the outcome; the options for checking for free prior to spending money on someone else.
Card failures occur while writing. This can happen when formatting or using a bad reader. Card failure does not occur while at rest. That is why the pictures are usually still there, but some prompts will destroy them.
There are three distinct sections within an external hard drive enclosure. Two of these are not the hard drive itself. If you “shuck” a Western Digital My Passport hard drive (remove it from its housing), you lose access to the data contained within it.
No-one fixes a broken stick and gives it back working. What you're paying for is a single read, and how that works depends on whether the build is discrete or monolithic.
The 48-digit key will almost certainly be found in one of four places; the first of these is your Microsoft account. If none of these have it, then we need to tell you the truth about where that leaves you.
Yes, on modern phones; yes, permanently. No, on many older mechanical drives. What this means if you are either selling or trying to rectify a past error.
No. Apple replaces defective components and does not return the original drive. Therefore, if the data matters, you need to recover your data first, then repair the defective component(s).
What is the difference? Documentation, not method. Write-blocked acquisition, hash verification and custody are all methods you cannot apply later.
Costs of recovering data on a single hard drive range from £250-£600 and costs of recovering data on an array of hard drives (multiple) will be £500-£1,500. There are four important questions you should ask any laboratory with whom you have dealings, including this one.
Higher-stakes recovery: failed arrays, failed solid-state drives (SSDs). Explained simply.
You were almost certainly observing the second failure. Why a rebuild is the most risky course of action you may take, and what you should send as an alternative.
A stripe does not include parity or mirror; therefore, when one disk fails, it removes a portion of each file. What can be recovered and why no one is able to rebuild from parity.
Each member of a mirror contains a duplicate copy of all information and often mounts in an ordinary desktop computer. Try that prior to sending money to anyone, and never rebuild in the wrong direction.
If multiple faults occur and fall into different pairs, then the array survives. But if both parts of a pair fail, then everything dies. This is why the pair, not the array, is the unit of risk.
A failed rebuild indicates that the surviving disks cannot supply a clean read. Every time a rebuild (retry) occurs, the system performs a full scan on those disks which are already failing.
There is only one data fault in the group of four. Most systems will run ordinary Linux mdadm underneath. In most cases, there is no lost volume associated with a single dead box.
Even though the array itself may be functioning properly, a volume may be unable to mount if the mapping between the logical block numbers and physical block numbers has been corrupted. Which QTS prompts should you reject?
Yes, usually. It is rare for the flash memory to fail; however, it is common for the controller to fail and the data generally remains intact. How is it determined if the data will come back?
SMART will report on flash wear but provides little information about controller health, and it's the controller that typically fails. What to take a look at regardless.
We see Seagate, WD, Toshiba, Samsung and SanDisk drives most often. If your drive fails, what should you do?
Why does the 7200.11 firmware fault report 0 GB when there is no physical problem? What makes an SMR Barracuda so slow to image, and how do donor disks get matched?
Colour indicates what it was built for. And why your My Passport and My Book have to be sent in their cases (the key's on the bridge board).
Works until it doesn't. Controller failure without a SMART warning, how important the thermal history is, and how encryption limits options.
Worked Monday, didn't come back Tuesday (flash fine). Why a firmware upgrade makes things worse and the two ways around a bad controller.
Primarily laptop drives and primarily drop damage. The majority of the outcome depends on whether the drive was operating at the time it was dropped.
Built better than most; therefore, they arrive worn rather than crashed, often from equipment that nobody has even thought about for a decade.
A Seagate drive in a better case. The bumper actually does work. The Thunderbolt bridge is what causes a healthy disk to appear dead.
XFS on mdadm locked. Why you cannot mount these drives in a computer and why the firmware recovery program can reset your array.
2023 Extreme Portable failures; differences between discrete and monolithic design; spotting counterfeits that report greater than actual storage.
Hardware disappeared. Formats have not. Zip cartridges, StorCenter arrays, and eGo disk drives can still be read.
There has been no new product manufactured since 2006. All units are considered “legacy” and generally sourced from an old industrial machine. We are limited by donor availability, not recovery method.
Reading will only take you so far. Provide a few details about what happened to your drive, or call. We will provide a truthful assessment and a complimentary diagnostic. In most cases you will only pay if we recover your data.