The Data Center Migration Checklist for Canadian Enterprises

A data center migration fails in the planning phase, not on the truck. Below is the exact checklist we use for enterprise migrations between Canadian facilities — refined across colocation, enterprise and carrier environments.

Phase 1 — Discovery (weeks -8 to -6)

• Full physical audit: rack elevations, PDU load per phase, cable inventory with source and destination ports
• Logical audit: VLAN map, IP address inventory, external DNS records, certificate expiry dates
• Dependency map: every service that talks to every service, with protocol and port
• Change freeze calendar cross-referenced against business events

Phase 2 — Design (weeks -6 to -4)

• Target rack elevations approved by the destination facility
• New cross-connects ordered (LOA/CFA lead time is 10-15 business days in most Canadian facilities)
• IP re-addressing plan, or L2 stretch design if you cannot re-address
• Rollback plan with a hard decision point and a named owner

Phase 3 — Staging (weeks -4 to -1)

• Cabling installed and tested at the destination before any equipment arrives
• Spare parts kit shipped to destination: optics, patch cables, PSUs, rails
• Runbook rehearsed with the migration team
• Communication plan distributed: customer notice, vendor bridges, on-call rotation

Phase 4 — Cutover weekend

• Snapshot every stateful system before shutdown
• Physical move follows the rack elevation, not the packing order
• Power-on sequence follows dependency order — storage first, then compute, then app tier
• Smoke test each service against a documented pass/fail criteria, not gut feel
• Decision point at T+4 hours: continue or roll back

Phase 5 — Decommissioning (weeks +1 to +2)

• Chain-of-custody documentation for every asset removed
• NIST 800-88 wipe or physical destruction for every drive
• Cross-connect disconnects filed with the source facility (billing continues until you file)
• As-built diagrams delivered and archived

The three items that kill migrations

1. Cross-connect lead time — teams forget the paperwork and lose two weeks
2. Undocumented dependencies — a legacy monitoring server nobody remembered
3. No rollback authority — the bridge cannot decide whether to continue at 3 AM

Fix these three and the migration becomes a paperwork exercise. Miss any of them and the cutover weekend becomes a 60-hour bridge call.

StackTrue plans and executes migrations across all major Canadian markets — from discovery through decommissioning — with signed chain-of-custody and Fluke-certified as-builts on delivery.