The server room checklist we run before every handover

A rack that works on handover day and a rack that still works in three years are different objects. The difference is almost entirely documentation and labelling.

The categories

• Power — UPS runtime tested under real load, not rated load
• Cooling — measured at the top of the rack, not the room
• Cabling — every run labelled at both ends, no exceptions
• Access — who has keys, who has console, written down
• Recovery — restore tested, not just backup configured

An untested backup is a hope with a cron job attached.

We photograph the front and rear of every rack at handover and attach it to the documentation. When someone calls eighteen months later, that photograph answers half the questions before we get in the car.