OPS HubOPS Academy · free for everyone Sign in

Data centre operations

Keep OPS Data's halls running: replace failed parts, keep power and cooling up, keep racks online and add capacity.

Overview

A data hall is racks of servers with a UPS (on mains or a generator), CRAC cooling and network uplinks. Every running server makes heat. If cooling can't keep up, the hall warms, servers degrade at 32 °C and shut down at 40 °C. On battery, the UPS runs out. A rack needs a ToR switch and CAT6 to an online router. OPS Cloud customers' servers run on these hosts: when hosts fail, their servers restart elsewhere or go down, and OPS Web, DNS and mail can depend on them too.

Jobs this covers

dc_fault, dc_cooling, dc_power, dc_network, dc_capacity — certificate: Data centre operations (required)

The equipment — what everything does

42U rack

Holds servers, storage, switch and firewall by rack unit.

In game: /towers → OPS Data. [E] on the rack → Install kit / the unit → Repair, Power, Role.

Compute / storage servers

Halcyon C1/C2 run cloud VMs or OPS services; S4 is storage.

In game: Installed into the rack from [E] → Install kit.

ToR switch

Puts the rack's servers on the network. No switch, no network.

In game: Install kit → Meridian TOR-48X.

UPS & generator

Keep racks powered through power cuts.

In game: /towers → OPS Data → Racks & power. The UPS needs a socket within 3 m or a generator in the hall.

CRAC unit

Removes 30 kW of heat each.

In game: /towers → OPS Data → Cooling. Needs a socket within 3 m.

Tools & PPE

Engineer laptop & console cable

Configures routers, NVRs and servers, runs speed tests and opens OPS Hub.

Torque driver

Tightens electrical terminals and rack rails to the correct torque — not too loose, not stripped.

Thermal camera

Shows hot spots: overloaded cables, failing fans, blocked server airflow.

RJ45 crimper & cable tester

Fits RJ45 plugs onto CAT6 and proves all eight cores are connected in the right order.

ESD wrist strap PPE

Grounds you so static electricity doesn't destroy server and network boards.

Standalone servers: tools are in your van (OPS Work → My van & tools). Items servers: collect them — and the job’s parts — at an OPS depot.

Step by step

1
Read the alert

OPS Hub → Data & Cloud shows which rack, unit and problem.

Engineer laptop & console cable OPS Hub → Data & Cloud type
2
Go to the rack

Set GPS to the job address (the data hall).

The data hall phone
3
Diagnose

[E] on the rack: power source, online, hall temperature and each unit's status and fault.

Thermal camera [E] on the rack inspect
4
Fix

Fault: the unit → Repair. Cooling: add or plug in CRAC units. Power: give the UPS mains or a generator. Network: fix the switch / CAT6 uplink. Capacity: install new compute servers.

Torque driver In the hall crimp
5
Confirm

Check the hall is healthy and the unit shows Running and online.

Engineer laptop & console cable [E] on the rack type
6
Complete

Press Check on the job.

OPS Work → the job phone

Health & safety

Static electricity

Risk: Static destroys server components.

Control: Wear an ESD wrist strap when handling boards and drives.

Manual handling

Risk: Heavy kit such as racks, UPS cabinets and cable drums can injure your back.

Control: Lift with your legs, keep loads close, and use two people or a trolley for anything over 25 kg.

Electricity

Risk: Mains voltage can give a fatal shock and cause burns.

Control: Never open powered kit. Isolate and prove dead with a voltage tester before touching terminals.

Heat & noise

Risk: Overheated halls and loud fans.

Control: Limit time in hot aisles, use hearing protection.

In game you confirm these in a risk assessment before starting. Skipping them on risky work can cause a fall or a shock — and it’s logged as a safety incident.

When it’s not done correctly

Leaving the hall overheating

Servers degrade at 32 °C and shut down at 40 °C, cloud servers go down, and the cooling check needs cooling ≥ the full demand.

UPS without mains in a power cut

The battery drains, every rack goes dark, and customers' servers and OPS services fail.

Rack without a working switch

Everything in it is unreachable — the rack offline job stays open.

Installing compute servers with the wrong role

Only 'cloud' role servers add capacity; the capacity check counts those.