
You have moved your files and email to Microsoft 365, SharePoint or another cloud service, and the old server in the comms cupboard is finally switched off. Before it goes in a skip, back to a leasing company or to a reseller, it needs a proper ending. A small office server often holds years of the company's history, and in most cases nobody has a full list of what is on it or how many drives it contains.
Confirm the migration really is finished
The first risk with an old server is destroying data you still need, not leaking it. Before anything is erased, check with whoever ran the migration that every share, mailbox and database has been moved, and that someone has opened a sample of files in the new system to confirm they arrived intact. Look out for the things that often get left behind: a finance or practice-management database, an old accounts package, scanned documents in a folder nobody uses any more, and personal drives belonging to staff who have left.
It is also worth asking whether anything on the server has to be kept for a set period, such as accounting records, HR files or material covered by a client contract or a legal hold. Under UK GDPR you should not keep personal data for longer than you need it, but you also must not destroy records you are obliged to retain. If the retention position is unclear, settle it before the server is wiped, not afterwards.
Many businesses leave the old server powered down but intact for a few weeks after the move, so that anything missed can still be retrieved. That is sensible, as long as the server stays somewhere secure and has a firm end date in the diary.
Find every drive, not just the obvious ones
A server rarely has a single disk. When you prepare it for disposal, count and record each of these:
- Data drives in the front bays, usually several disks working together as a RAID set.
- Hot spares: drives sitting in a bay unused, waiting to take over if another fails. They are easy to overlook, and may hold data if they were ever brought into service.
- Boot drives: many servers start up from a separate small SSD, an M.2 card, an internal USB stick or an SD card on the motherboard, holding the operating system along with its logs and settings.
- Failed drives removed in the past: when a disk failed and was replaced, the old one often went into a drawer or a box in the cupboard. It still holds data.
- Backup kit: USB drives used for rotation backups, an attached backup NAS or a tape drive with cartridges still in it.
Write down the serial number of each drive as you go. That list becomes the basis for checking the certificate of destruction later.
Why a RAID set doesn't make single disks safe
A common assumption is that a drive taken out of a RAID array is useless on its own, because the data is spread across several disks. That is not a safe basis for disposal. Depending on the RAID level, a single disk may hold a complete copy of everything (mirroring), and even in striped arrays each disk holds readable fragments: whole emails, parts of documents, names and addresses. Treat every drive from the server as if it holds the full contents.
Deleting the RAID volume in the controller settings, or reinstalling the operating system, is not erasure either. It removes the structure that tells the server where data lives, but most of the data itself stays on the disks.
Don't forget the settings stored in the server itself
Servers from the main manufacturers have a built-in management controller (Dell calls it iDRAC, HPE calls it iLO) that keeps its own network settings, user accounts and sometimes alert email details. The RAID controller and BIOS also hold configuration. Before the chassis leaves your control, reset these to factory defaults and remove any accounts and passwords you set up, particularly if the hardware is going back to a leasing company or to a reseller.
Change or revoke anything that the old server used to reach other systems, too: service account passwords, VPN details and any API keys or certificates installed on it. Shutting the machine down does not cancel those credentials.
Erase or destroy?
The right method depends on what happens to the hardware next:
- Leased or to be resold: the drives need to work afterwards, so they should be erased with software that overwrites or purges every part of the disk and produces a report for each serial number. Check your lease agreement, because some leasing companies require drives to be returned, while others let you keep or destroy them for a fee.
- End of life: for older servers with no resale value, physical destruction of the drives is simpler and leaves no doubt. Drives that have failed or won't complete an erase should always be destroyed.
- SSDs: solid state drives store data differently from spinning disks, and standard overwriting doesn't always reach every part of them. Use a purge method designed for SSDs, or have them physically destroyed. Our guide to wiping, degaussing and shredding explains the differences.
The empty server chassis, rails and power supplies can then go for WEEE recycling with the drives removed. See our hard drive disposal and recycling service for how the remains are handled.
Keep the paperwork
Once the drives have been collected, you should receive a certificate of destruction listing each drive by serial number and the method used. Check it against the list you made when you removed them, and chase any drive that is missing. Our article on what a certificate of destruction should include covers what to look for.
Finally, update your asset register to show the server as disposed of, with the date and a reference to the certificate. If a regulator, auditor or client ever asks what became of the data on that machine, that record is your answer.
Booking a collection
Data Destruction London collects server drives, loose disks and backup media from businesses across London, using government-approved methods and issuing a certificate of destruction for every job. Call 020 7064 4996 or book a collection online.
