· 4 min read
A data center relocation is one of the riskiest projects an IT team can run. Servers, storage and network equipment have to be shut down, packed, transported, installed and brought back online, often within a tight maintenance window. The good news is that almost every problem on moving day can be prevented by decisions made weeks earlier.
Use this checklist to plan your move in five phases, from the first inventory to the final sign-off.
Phase 1: Discover and document everything
You cannot move what you have not listed. Build a complete inventory before any other planning starts.
- Hardware: every server, storage array, switch, firewall, KVM and PDU, with make, model, serial number and rack position.
- Connections: network ports, fibre links, storage paths and power feeds for each device. Label both ends of every cable.
- Applications: which business services run on which systems, and who owns each one.
- Dependencies: which systems must start before others. A database usually needs to be up before the applications that use it.
- Contracts: maintenance contracts, licences tied to hardware, and any provider that must be informed of the new address.
Note: Take clear photos of the front and back of every rack before anything is touched. They are the quickest way to check cabling during reinstallation.
Phase 2: Prepare the new site
The new room should be fully ready before the first device leaves the old one. Confirm each of these in writing:
- Power capacity per rack, redundant feeds and UPS runtime.
- Cooling capacity for the full heat load, including future growth.
- Rack layout, floor loading and clear access routes for equipment.
- Network uplinks, internet circuits and any links to other sites, tested end to end.
- Physical security, access control and fire suppression.
Circuits from telecom providers often have long lead times, so order them early.
Phase 3: Plan the move itself
Choose a move strategy
There are three common approaches, and many projects mix them:
| Strategy | How it works | Best for |
|---|---|---|
| Lift and shift | Equipment is shut down, moved and restarted | Smaller rooms and non-critical systems |
| Phased move | Systems move in groups over several windows | Larger environments with many dependencies |
| Parallel build | New hardware is installed first, then data is migrated | Critical services that cannot stop |
Write a run book
A run book is a step-by-step script for moving day: who shuts down what, in which order, at what time, and how each step is confirmed. Include contact numbers for every person and supplier involved.
Agree a rollback plan
Decide in advance what happens if a system does not come back as expected, and the latest time at which you will stop and roll back. This turns a stressful judgement call into a planned step.
Phase 4: Protect your data
- Take full backups of every system shortly before shutdown.
- Test a restore, not just the backup job. A backup that cannot be restored is not a backup.
- Keep at least one copy of critical data away from both the old and the new site during the move.
- Record configuration of network devices and storage arrays, and export them to files.
Phase 5: Move day and testing
- Announce the maintenance window to users and stakeholders.
- Shut down applications, then servers, then storage and network equipment, in the order set in the run book.
- Pack equipment with anti-static protection and shock-absorbing material, and transport it with insured, tracked vehicles.
- Install equipment in the new racks using the inventory, labels and photos.
- Power on in reverse order: network, storage, then servers and applications.
- Test each business service with its owner before you declare it live.
After the move, keep the team available for a few days to handle any issue quickly, and update the documentation with the new layout.
Common mistakes to avoid
- Underestimating the time needed to reconnect and test cabling.
- Forgetting licence keys or hardware-locked software.
- Not checking that firmware and spare parts are available in case a device fails after power-up. Older disks in particular can fail when restarted after a long run.
- Splitting responsibility between too many suppliers, so no one owns the whole result.
How ECS supports data center relocation
ECS plans and carries out data center relocation and installation for companies in Cairo, Alexandria and across Egypt. One team handles the inventory, dismantling, packing, transport, installation and testing, so responsibility stays in one place.
Because we also provide multi-vendor hardware support, our engineers can replace a failed part on the spot if a device does not come back after the move. If you are reviewing whether to keep older equipment at all, read our guide to third-party maintenance.
Planning a move? Tell us about your environment and we will help you build the plan.