Prepare an equipment manifest before shipping
Create a manifest for every device: manufacturer, model, serial number, rack units, chassis depth, weight, PSU count, expected power draw and management interface. Include rails, power cables, transceivers and any non-standard accessories. This document becomes the reference for receiving, rack installation and later hardware support, and it reduces ambiguity when the team operating the equipment is in another country.
Validate rack compatibility and power
A server may be 1U or 2U but still require unusual depth, rail clearance or power density. Confirm the physical dimensions, rail type, connector requirements and whether A/B power is needed. For GPU systems, storage arrays or dense appliances, provide measured or vendor-estimated normal and peak consumption. Power should be commercially confirmed before the equipment is dispatched.
Design production and management networks
Decide how production traffic, management traffic and any replication or backup traffic will be separated. If the server provides IPMI, iLO, iDRAC or another BMC, define how the remote engineering team will access it. Request the required IPv4 quantity and describe where traffic will go so local and international bandwidth can be reviewed before go-live.
Agree receiving and Rack and Stack scope
Do not assume that receiving a shipment automatically includes installation. The work order should state who receives the equipment, whether serial numbers and package condition must be recorded, which rack positions are reserved, how cabling should be labeled and what basic checks are required after power-on. If a specific cable map exists, include it before the installation window.
Define a safe handover
A useful handover records device position, power feeds, network ports, management access status and any exceptions discovered during installation. The remote team should confirm that the device is reachable and that the expected interfaces are visible before the physical deployment is considered complete. Application or data migration should be treated as a separate scope unless explicitly included.
Plan operations after day one
The deployment is not finished when the server first boots. Keep an inventory of spare drives, memory, PSUs and any proprietary components that may be difficult to source quickly. Define Remote Hands and Smart Hands authorization, normal maintenance windows and critical escalation contacts. This turns a one-time installation into an operating model that international teams can maintain remotely.