Server Hardware Installation and Management
Racking and physical install
Servers are measured in rack units (U); 1U = 1.75 in (44.45 mm). Confirm rail-kit compatibility, install heavier units low, and leave space for airflow. Use cable management arms or hook-and-loop ties, label both ends of every cable, and never block intake fans. Front-to-back airflow means a cold aisle in front of the rack and a hot aisle behind it.
Power and cooling
Redundant power supplies protect against a single feed failure only if the two feeds are independent (separate circuits/PDUs). A UPS conditions power and bridges short outages; a PDU distributes it. Calculate the load in watts and confirm circuit capacity before adding hardware. Watch rack temperature and humidity; a server in a hot spot will throttle or shut down.
Internals
Know DIMM population rules (channels, matched pairs), ECC versus non-ECC, and RDIMM versus UDIMM. Mixing sizes or speeds can down-clock the whole bank or prevent POST. Storage choices matter: SATA, SAS, and NVMe differ in interface, throughput, and hot-swap support. RAID levels: 0 stripe (no redundancy), 1 mirror, 5 single parity, 6 dual parity, 10 stripe of mirrors.
Management
Out-of-band controllers (iDRAC, iLO, IPMI/BMC) provide console access and health telemetry even when the OS is down. Maintain a firmware baseline, asset tags, warranty records, and accurate documentation.
Traps: "hot spare" (idle standby disk) is not the same as "hot swap" (replace while running); RAID 6 usable capacity is easy to miscalculate; a second PSU on the same circuit is not true redundancy; and U height describes physical form factor, not performance.
Worked example: six 4 TB drives in RAID 6 give (6 - 2) x 4 TB = 16 TB usable and survive two simultaneous drive failures.