I85PREP
⌘ K
EXIN CDCE study materials
EXIN CDCE study resources

EXIN CDCE: Data centre operations and maintenance

EXIN CDCE · Work through your exam, one topic at a time.

All guide sections

Data centre operations and maintenance

Domain 2 - Data centre operations and maintenance. Here the exam tests the operational discipline that preserves the availability the design promised: documented procedures, controlled change, planned and predictive maintenance, monitoring, capacity management, incident and emergency preparedness.

The document set

  • SOPs - steady-state procedures such as start-up, switchover and rounds.
  • MOPs - one-off changes with numbered steps, risk assessment, approvals, rollback and post-change validation.
  • EOPs - emergency scenarios (utility loss, cooling failure, fire, water ingress), written and drilled.
  • Supporting controls: permit to work, lock-out/tag-out, change advisory review, spares register, as-built documentation.

Maintenance strategy

  • Preventive (calendar or run-hours), predictive (IR thermography, vibration analysis, battery impedance, oil analysis) and corrective maintenance.
  • Commissioning levels 1–5 establish the baseline; retro-commissioning and measurement & verification prove it still exists.
  • Availability = MTBF / (MTBF + MTTR). Reducing MTTR through spares, training and remote diagnostics is usually cheaper than raising MTBF.

Worked example

A UPS block has MTBF 40,000 h and MTTR 6 h.

  • Availability = 40,000 / 40,006 = 99.985%, which is about 79 minutes of unplanned downtime per year (0.00015 × 8,760 h).
  • The client requires 99.99% = 52.6 min/year. The block alone cannot meet it, and planned maintenance outages make the operational figure worse.
  • Two remedies: make the block 2N so planned work and single faults never touch the load, and cut MTTR with on-site spares and a drilled response team. Both must be evidenced by test records, not by intent.

Traps that cost marks

  • Treating maintenance as optional while the load runs fine; deferred maintenance is a hidden risk acceptance.
  • Changes executed without a MOP, rollback plan or change advisory approval.
  • EOPs written but never exercised; generators or UPS never tested under load.
  • Capacity managed from nameplate data instead of metered kW per rack.
  • Key-person dependency: one engineer who alone knows the switchover sequence.
  • Documentation that is not as-built, so troubleshooting starts from a wrong diagram.