Any noise it did not make before
Beep, click, grind, scrape, or a buzz with no spin. Mechanical failure grows with each rotation, therefore power off immediately rather than copy files. Data lost per second of time spent.
Power off
Symptoms appear in somewhat consistent order and by the time anyone notices they can be associated with a bad disk it may already have been deteriorating for several weeks.
Every minute that the drive runs, mechanical symptoms will continue to worsen. Symptoms related to performance and logic provide a window of opportunity for intervention. Identifying which category a symptom falls into is effectively the whole decision.
Knowing where in the process you are tells you the amount of time you have available.
It takes twenty seconds to open a file but opens immediately on opening again. This is the drive retrying a weak sector until it finally works and the file opened previously remains cached. The earliest reliable indication of problems, and the one most likely to be overlooked.
All activity freezes for thirty seconds while the disk light stays steady; after the freeze you can continue. The operating system is waiting for the end of a disk read operation. Virtually all users interpret this as "your computer is just old".
A file copy transfers at full speed until a specific percentage when it stops. Each time you try to run a copy it hits the exact same unreadable area. Unlike the earlier signs, this symptom is repeatable and predictable, so it is the easiest one to verify.
Your files show as corrupted; photos display grey areas instead of colour; applications refuse to start. The drive's retries are starting to fail instead of taking longer.
At this point your hard drive is unreliable and may be unmounted from the OS. This is the point where you have very little time left, if any, to pull your data off your drive yourself.
The difference between them will make the right answer different.
Beep, click, grind, scrape, or a buzz with no spin. Mechanical failure grows with each rotation, therefore power off immediately rather than copy files. Data lost per second of time spent.
A drive that goes missing while working and then comes back when plugged in again is unstable at the interface or stalling badly. With each drop there's an opportunity for partial writes.
It takes too long, freezes, throws errors but reads still? That means you are within a good time frame. Make copies to another physical drive based upon importance.
chkdsk, defrag and deep scanning all require hours of full-surface reading on hardware that may be too weak to survive the scan.
And anyone giving you a number is guessing.
There are some drives where they operate for months with an increasing "reallocated" count but still show no additional decline. On the other hand there are some drives that, after the first time they get stuck, go completely unreadable in under a week. There is no predictive tool or model that will give you the ability to determine how long your drive has left – as each cause is unique – a single bad part manufactured during production is vastly different from a head beginning to degrade.
The only thing we know about the failures is the direction of travel. The "reallocated", and "pending" sector count numbers will never decrease. So whatever window you have is narrowing, and when determining how much time you have left – the answer will always be the same – assume less time than you want to, and copy your critical information now.
Most of the costly recoveries that come to us were drives which had spent weeks in the early stages with someone deciding whether the laptop should be replaced.
Still reading, even badly. You are the best recovery tool available, and you cost nothing. Instead of copying files by alphabetical order, use importance when you copy the files to another hard drive, and stop using the machine for anything else during the file copy process. Do not image, do not clone, do not repair – simply copy what you need.
Making noise, or no longer presenting. Shut it down. At this stage every further attempt to recover data reduces what is left and the work requires equipment you don't have. Recovery for single disk drives in Manchester is from £300 +VAT after a 48-hour free assessment of your drive, and if we find you could have backed up those files yourself, we'll let you know.
Occasionally very slow openings of files which may then open instantly on a subsequent attempt, and occasionally freezing which stops itself within twenty or thirty seconds. These are both due to the drive trying to read weak areas of the disk while the operating system waits.
The signs of failure cannot be distinguished from those of software bloat (slow performance, freezes, intermittent failures). Genuine ageing typically occurs gradually over many years while a dying disk will show significant degradation over days to weeks.
No one can honestly tell you. There is no way to know how long it will last: some drives run for months, others fail within a week. One thing is certain: sector counts never go back down. Therefore, whatever window you have now is getting smaller instead of staying where it is.
You should copy all files that matter, in order of their importance as opposed to alphabetical. Full images read every single sector, including huge areas you do not want to read from, wasting the rest of your disk's life reading the wrong data.
Anytime it sounds different from the last time it made noise, or anytime it drops off and comes back. That is a mechanical or interface fault, and attempting to recover anything from the disk will just use up more of your chances at recovery.
Not in any way that makes it trustworthy again. The moment your disk starts reallocating bad sectors, it is eating away at its own reserve. It does not go backwards. Save your files and replace your disk.