DCAT common mistakes
DCAT · Understand the traps and how to avoid them.
Treating single-corded vs dual-corded server power feeds as interchangeable when planning redundancy, and assuming any two feeds provide fault tolerance.
Candidates memorize 'A/B power' as a slogan and never learn that both cords on one PDU or both PDUs on one upstream circuit defeat redundancy; the exam tests whether you trace the actual path.
What to do instead
Trace each cord end-to-end to its power source before answering. Redundancy exists only when the two paths share no PDU, UPS, or upstream circuit. Practice sketching feed topologies and identifying the single point of failure.
Confusing hot-swap, hot-plug, and redundant components, and stating that a failed non-redundant component can be replaced without downtime.
The terms sound similar and candidates learn them as synonyms; the exam hinges on whether the system keeps running during removal.
What to do instead
Define each precisely: hot-swap means the system stays operational during replacement, which requires redundancy of the affected subsystem. Ask 'what else is carrying the load?' before claiming no downtime.
Mixing up single-mode and multi-mode fiber limits, connector types, and transceiver compatibility, then selecting a cable/optics pair that cannot work together.
Candidates learn cable names without the distance, wavelength, and optic-mating constraints, and the exam presents plausible-sounding but incompatible combinations.
What to do instead
Build a small reference table: multimode vs single-mode distances, common connector types, and the rule that both ends must match transceiver type and media. Verify reach against the actual link distance before choosing.
Applying RAID levels by name without reasoning about what each protects against, e.g., claiming RAID 5 tolerates two simultaneous drive failures or that RAID is a backup.
RAID 0/1/5/6/10 get memorized as numbers rather than as tradeoffs in capacity, performance, and fault tolerance.
What to do instead
For each level state: minimum drives, drives that can fail, usable capacity, and rebuild behavior. Explicitly note RAID is not backup and does not protect against controller, enclosure, or site loss.
Answering cooling questions using only temperature while ignoring airflow direction, containment, and hot/cold aisle layout.
Candidates focus on setpoints and forget that recirculation and bypass airflow cause most thermal problems in real halls.
What to do instead
Reason in terms of cold-aisle supply to intake, hot-aisle return, and blanking/sealing to prevent recirculation. When a scenario shows a hot spot, check airflow path and containment before adjusting setpoints.
Skipping the documented pre-replacement steps (verifying the replacement part, confirming redundancy, recording serials, ESD precautions) and jumping straight to the physical swap.
Practical experience encourages shortcuts, and candidates answer as they would act in a lab rather than as a procedure requires.
What to do instead
Memorize a standard replacement sequence: identify and verify part, confirm redundancy/notification, ESD and safety prep, remove, install, verify link/health, update records. Answer with the earliest required step when asked 'what first.'
Misreading layered troubleshooting scenarios by jumping to the most dramatic cause instead of following a bottom-up or top-down diagnostic order.
Candidates pattern-match on symptoms they have seen and skip basic checks like link lights, cabling, and power state.
What to do instead
Apply a consistent model (physical through application) and eliminate the cheapest, most common causes first. In scenario questions, ask which layer the evidence actually points to before selecting a fix.
Treating physical security and environmental safety as separate trivia rather than reasoning about access control, audit trails, and lockout/tagout in operational scenarios.
Candidates study compliance as a list of standards to memorize instead of as behaviors applied during maintenance.
What to do instead
For each scenario ask: who is authorized, what is logged, and what de-energization or PPE step is required before work. Tie each answer back to the control being satisfied.
Confusing network addressing and segmentation concepts, such as management vs data plane separation, VLAN purpose, and where a device sits in the topology.
Candidates learn device names and speeds but not the function of each network segment or why separation matters.
What to do instead
Map each network (management, storage, data, out-of-band) to its purpose and traffic type, and practice placing devices on the correct segment. Ask what breaks if the segments are merged.
Assuming a component's rated capacity or throughput equals what the workload receives, ignoring oversubscription, derating, and environmental limits.
Spec sheets invite candidates to quote nameplate numbers, while the exam tests effective capacity under real conditions.
What to do instead
Always distinguish nameplate from usable: apply derating for power (e.g., circuit loading limits), oversubscription for network uplinks, and reserve capacity for storage. State the assumption explicitly when calculating.