IT and Server Room Relocation: Packing, Labelling and Reconnection

The move is the easy part

Carrying a rack across Dubai is a two-hour job. Getting the systems inside it to come back up in the right order, on the right VLANs, with the right cabling, is what takes the weekend. Plan accordingly: the documentation phase deserves more time than the transport.

Before anything is unplugged

Photograph everything. Front and back of every rack, wide and close. Cable management arms, patch panel faces, power distribution units. You will refer to these photographs more than any spreadsheet.

Build an asset register. Every device: make, model, serial, asset tag, rack position (U number), power draw, and the ports it uses.

Map the dependencies. Which systems must be up before others can start? Domain controllers and DNS before application servers. Storage before hypervisors. Write the shutdown order and the reverse startup order down, on paper, and give a copy to each engineer.

Take a backup and verify it restores. Not "a backup exists" — a tested restore. This is the only insurance that matters.

Labelling

Label both ends of every cable. A label at one end tells you where a cable starts and leaves you guessing where it goes.

A workable scheme: RACK-U-PORT → RACK-U-PORT. So R3-U14-P1 → R1-U42-P17. It is verbose and it is unambiguous at 2am.

Colour-code by function if you have the tape — production, management, storage, out-of-band. It saves more time than it costs.

Preparing the destination first

The new room should be finished before a single rack leaves the old one:

• Power: correct circuits, correct phases, sufficient capacity, UPS tested under load.
• Cooling: running and at temperature. A sealed room in a UAE summer reaches damaging temperatures within the hour if the cooling is not live.
• Connectivity: the circuit from the ISP tested end to end. Etisalat and du lead times for a new business line run to weeks, not days. This is the single most common cause of a failed cutover date.
• Structured cabling: patch panels terminated and tested.
• Floor: load rating checked, and a clear path from the loading bay with no steps and no lift that is too small.

Powering down

Shut down cleanly and in dependency order. Never pull power on a running database.

Then:

• Remove every device from the rack. Racks travel empty. A populated 42U rack is top-heavy, exceeds most ramp ratings, and will tip.
• Bag and label rails, screws and cage nuts per device. A missing rail kit on a Sunday morning stops the whole rebuild.
• Remove cable management arms and pack them with their rack.

Packing hardware

• Servers and switches: original boxes are ideal. Failing that, anti-static bags inside foam-lined cases. Not bubble wrap against bare boards.
• Drives: if they come out, they come out into anti-static packaging, labelled with their array and slot. If an array is moving assembled, note the drive order — some controllers care.
• Optics and SFPs: small, expensive, and lost easily. One labelled box.
• Tapes and media: climate-sensitive. They travel in the cab, not the container.

The transport itself

Air-ride suspension where you can get it. Racks strapped upright, never laid down. Straps on the frame, not on equipment. And keep the vehicle temperature controlled — the back of an unrefrigerated box truck in Dubai in July exceeds the storage rating of most enterprise hardware.

For short inter-emirate moves, the risk is handling and ramps, not distance. Most damage happens in the last ten metres.

Bringing it back up

Reverse the shutdown order. Bring up one layer, verify it, then the next. Resist the temptation to power everything on at once — you lose the ability to tell what failed.

Check after each layer:

• Link lights and negotiated speed on every port
• Storage paths and multipath status
• Backup jobs running
• Monitoring and alerting actually reporting
• Out-of-band management reachable, which is what saves you on the next outage

Timing it

Most UAE businesses cut over Friday evening to Sunday morning, which gives roughly 36 usable hours. Build in a rollback point: a time by which, if the new room is not up, you restore service at the old site. Decide that time before you start, not at 4am when judgement is worst.

And keep the old room's power and connectivity live for at least a week after cutover. It costs very little and it is the difference between an inconvenience and an outage.