EPI Data Centre Foundation common mistakes
EPI Data Centre Foundation · Understand the traps and how to avoid them.
When a question quotes a temperature, humidity, or redundancy figure, the candidate answers from general IT intuition (for example 'about 18–21 °C is comfortable for people') instead of applying the envelope defined for IT equipment inlet conditions, and picks the option that describes human comfort rather than equipment tolerance.
Candidates carry over everyday office-environment assumptions and have not internalised that data centre environmental limits are specified at the equipment air inlet, not at room level, and that recommended versus allowable ranges are distinct concepts tested separately.
What to do instead
Before answering any environmental question, restate to yourself whether it asks for recommended or allowable conditions and where the measurement point is. Build a one-page revision sheet listing the recommended and allowable temperature and humidity envelopes with the measurement location, and practise classifying ten sample scenarios as 'within recommended', 'within allowable only', or 'outside allowable'.
The candidate treats all redundancy configurations as equivalent, selecting 'N+1' or '2N' interchangeably when the question describes a specific failure scenario, or assuming a fault-tolerant topology also removes the need for concurrent maintainability.
The labels are memorised as slogans without a mental model of what each topology tolerates. Candidates also conflate component-level redundancy (an extra UPS module) with system-level topology (two independent paths), which are assessed differently.
What to do instead
Draw the four common topologies from memory (N, N+1, 2N, 2N+1) and for each write down which single failures it survives and whether maintenance can be performed without load loss. Then answer topology questions by sketching the described system and striking out the failed component before choosing an option.
On cooling and airflow questions the candidate answers from the perspective of the cooling unit's supply temperature or the room's average temperature, rather than reasoning about hot spots, bypass airflow, and recirculation at the rack face.
Cooling is studied as a plant-capacity topic (chillers, CRAHs) rather than as an airflow-management topic, so candidates miss that adequate total capacity can coexist with local overheating.
What to do instead
Reframe every cooling question as 'what is happening to the air at the inlet of the IT equipment?'. Revise containment, blanking panels, hot/cold aisle arrangement, and bypass/recirculation as airflow pathologies, and for each practice question state explicitly whether the issue is capacity, distribution, or control.
The candidate selects an answer describing a corrective action (dispatch a technician, replace a part, raise a ticket) when the question stem is asking about the first step in an operational process, which is usually detection, verification, or escalation according to procedure.
Exam items often present a plausible technical fix that would indeed solve the problem eventually, and candidates optimise for 'fixes the fault' rather than 'is the correct next step in the documented process'. Operational discipline is being tested, not troubleshooting skill.
What to do instead
For every scenario question, identify the process stage named in the stem (detect, verify, escalate, isolate, restore, record, review) before reading the options. Eliminate any option that jumps stages. Practise by writing the correct sequence for alarm response, change execution, and incident escalation from memory.
The candidate treats maintenance windows and change management as administrative overhead, choosing answers in which a technician performs work immediately to resolve an issue, without authorisation, permit, method statement, or rollback plan.
Hands-on candidates identify with the engineer who acts quickly, and underestimate how heavily the operations domain rewards controlled, authorised, reversible change.
What to do instead
Adopt the rule that in this domain the 'correct' answer is the one that preserves control: authorisation before action, documented procedure, risk assessment, rollback, and post-change verification. Review the change lifecycle and, for each practice item, check whether the chosen option includes approval and rollback; if not, look again.
The candidate answers capacity and availability questions using instantaneous readings or nameplate ratings, ignoring trend data, growth forecasting, and the difference between design capacity, installed capacity, and usable capacity after redundancy is accounted for.
Capacity is studied as a specification rather than as a management process, so candidates do not distinguish what is installed from what is available with redundancy maintained, and forget that utilisation trends drive forecasting decisions.
What to do instead
Learn the capacity chain explicitly: design capacity, installed capacity, capacity available with redundancy intact, and consumed capacity. For each practice question, state which of these the stem is describing and whether the decision horizon is immediate (headroom) or forward-looking (forecast).
The candidate mixes up the roles of standards bodies and frameworks, for example attributing data centre facility design guidance to an IT service management framework, or citing a certification scheme as if it were a standard, and answers 'which document specifies X' questions by association rather than by scope.
Standards are memorised as a list of acronyms without their scope, so candidates cannot discriminate between a design guidance document, a management system standard, a certification scheme, and a professional code of conduct.
What to do instead
Build a scope table with one row per document or framework and columns for: what it covers, what it does not cover, who it is aimed at, and whether it supports certification or guidance only. Test yourself by writing the scope of each from memory before checking, and answer 'which document' questions only after matching the stem's subject to a scope row.
When a question asks what an audit or assessment should examine, the candidate chooses the answer describing whether the facility looks well run or whether staff are competent, instead of evidence-based verification against defined criteria, documented records, and traceability.
Candidates conflate audit with inspection or with operational review, and are drawn to options that sound like good practice rather than options that describe evidence, criteria, sampling, and findings.
What to do instead
Anchor on the audit definition: independent, documented, evidence-based evaluation against agreed criteria. For each option ask 'is this evidence against a criterion, or is this an opinion?'. Revise the audit cycle (scope, criteria, evidence gathering, findings, corrective action, follow-up) and practise classifying sample activities as audit, inspection, or operational check.
The candidate answers governance questions by picking the technically optimal solution, ignoring that governance concerns decision rights, accountability, risk appetite, and oversight structures rather than engineering choices.
Governance is abstract compared with cabling and cooling, so candidates fall back on technical reasoning, and they miss that the correct answer often describes who decides, who is accountable, and how risk is accepted and recorded.
What to do instead
For governance items, map the stem to one of four questions: who is accountable, what is the risk appetite, what decision rights exist, and how is performance or compliance overseen. Eliminate options that are purely technical or that assign accountability to a committee without a named owner. Revise the distinction between governance, management, and operations with one concrete example each.
The candidate treats risk, business continuity, and professional competence as separate memorised topics, so when a scenario mixes them (for example a single-person dependency on a critical procedure) the candidate picks an answer addressing only the technical fault and overlooks the competence, succession, or ethical dimension.
Revision is organised by syllabus heading, but exam scenarios are deliberately cross-cutting; candidates stop analysing once they recognise a familiar technical cue.
What to do instead
For scenario items, run a fixed checklist before choosing: technical impact, process control, risk and continuity, people and competence, and ethical or reporting obligations. Revise the concepts of single points of failure in staffing, competence records, and the duty to escalate or refuse unsafe work, then practise applying the checklist to multi-domain scenarios.