Recognising the warning signs
Drives rarely fail without some warning. Traditional spinning hard drives may click, grind or make repeated spin-up sounds, and the computer may become very slow, freeze when opening files or report disk errors. Solid state drives tend to fail more quietly, with sudden freezes, files that will not open, or the computer refusing to start. In servers and NAS devices, the management screen or email alerts may report a degraded array or a disk with increasing errors.
Do not ignore these signs. A drive that has started making noises or throwing errors is likely to get worse, and every hour of continued use risks further damage. If the machine is still working, the priority is to make sure its data is safely copied elsewhere, not to keep using it normally.
The first ten minutes
If you hear clicking or grinding, shut the computer down gracefully if you can and leave it off. Mechanical damage gets worse each time the drive spins. If the drive is still readable and there is important unbacked-up data, copy the most important files first to another drive or the cloud, starting with what you cannot recreate. Do not try to copy everything at once.
Next, check what is already protected. Is the data in OneDrive, SharePoint or Google Drive? Is the machine covered by a backup agent, and when did it last run? Is the failing drive part of a RAID array that is still running on the remaining disks? These answers decide whether this is a ten-minute inconvenience or a serious problem.
What not to do
Do not run disk repair or check-disk utilities on a drive that is physically failing. They can write changes across the disk and make professional recovery harder or impossible. Do not open the drive casing; even a speck of dust can destroy the platters. Do not put the drive in the freezer, tap it or try other home remedies found online. And do not keep restarting the computer hoping it will come good.
For RAID arrays in servers and NAS devices, do not pull disks out or rebuild the array without understanding exactly which disk has failed. Removing the wrong disk from a degraded array can take it offline completely. If in doubt, leave it running in its degraded state and get someone experienced to look before touching anything.
If the data is backed up
When the data exists in a good backup or in cloud storage, the drive is just hardware. Replace it, or replace the whole computer if it is old, then restore the files or the full system image. For a laptop with everything in OneDrive, this can be as simple as signing in on a new device and letting files sync down. For a server, restore the image to the new drive or replacement server, then check applications and permissions.
Before wiping or disposing of the failed drive, make sure the restore is complete and correct. Then destroy the old drive properly or have it securely wiped, because it may still contain readable personal or business information. Note the serial number and how it was disposed of, which helps if you ever need to show that client data was handled properly. If the device was leased, check the lease terms, since some require the drive to be returned or certified as destroyed.
If the data is not backed up
If the drive holds the only copy of important data, stop and think before acting. A professional data recovery lab can often retrieve data from drives with mechanical or electronic faults, but their chances drop with every attempt made by someone else first. Keep the drive powered off, note what happened and when, and get a quote and assessment from a reputable recovery service.
Professional recovery is not cheap and not always successful, so weigh up what the data is worth. Some files can be recreated from emails, client copies or paper records. Others, such as years of accounting data or irreplaceable project files, may justify the cost. Make the decision with the person who owns the data, not just the IT person, because they know what can and cannot be rebuilt from other sources.
Preventing the next one
Every drive will fail eventually; the only question is whether it matters when it does. Make sure work is stored in the cloud or on a backed-up server, not only on individual machines. Monitor drive health on servers and NAS devices, with alerts going to someone who acts on them. Replace servers and NAS units before they reach the age where failures become common, typically around five to six years.
After a failure, use it as a prompt to check that every device is covered. It is common to find that the failed laptop was the one machine missed when backups were set up, and that others are in the same position. A quick review of every computer's backup status after an incident usually costs very little and can stop the same problem happening again next month.
In short
- Stop using a failing drive and do not run repair tools on it.
- Check the cloud and backups before assuming data is lost.
- Leave degraded RAID arrays alone until someone experienced looks.
- Use each failure as a prompt to check every device is backed up.