CompTIA Server+ Course Notes

Server hardware decisions are driven by availability, workload, serviceability, and failure-domain design. Rack installation starts with weight distribution, rail compatibility, grounding, cable management, airflow, clearances, and physical access. Power design should account for actual and peak draw, redundant PSUs, PDUs, UPS capacity/runtime, separate circuits where required, generator coverage, and environmental monitoring. Cooling depends on unobstructed front-to-back airflow, hot/cold aisle practices, inlet temperature, humidity, dust control, and alerting. RAID is not backup: it improves availability/performance during drive failures, but it does not protect against deletion, corruption, ransomware, controller failure, or site loss. RAID 0 stripes without fault tolerance; RAID 1 mirrors; RAID 5 uses single parity; RAID 6 uses dual parity; RAID 10 stripes across mirrored sets. Shared storage choices such as NAS, iSCSI SAN, and Fibre Channel SAN differ in protocol, latency, throughput, cost, and operational complexity. Out-of-band management uses a separate management controller/path so administrators can power-cycle, access console, view hardware health, and mount remote media even when the host OS or production network is unavailable. Firmware and hot-swap work must follow vendor compatibility matrices, maintenance procedures, backups/configuration capture, and rollback/recovery planning.

• Treating redundant PSUs as redundant facility power without checking PDUs/circuits.

• For power redundancy, count independent failure domains—not PSU count.

Server Administration connects installation, networking, operating systems, high availability, virtualization, automation, and lifecycle control. Disk layout decisions include partitioning schemes, boot requirements, volume management, file-system features, permissions, resiliency, snapshots, quotas, and workload behavior. Avoid absolute rules such as 'large capacity requires NTFS or XFS'; NTFS, ReFS, ext4, XFS, and other server file systems have different limits/features and must be matched to the operating system and workload. Server networking requires correct IP addressing, subnet mask/prefix, default gateway, DNS, DHCP/reservations where appropriate, VLAN membership, bonding/teaming, routing, and firewall rules. Monitoring should correlate CPU, memory, disk IOPS/latency/queue, network throughput/errors, service state, logs, and application metrics rather than treating one utilization percentage as proof of root cause. Data migrations need inventory, capacity checks, permissions/ownership preservation, integrity validation, staged synchronization/cutover, and rollback. Clustering provides coordinated high availability, load balancing distributes requests/work, and failover transfers service when a node/path becomes unhealthy. Virtualization requires capacity planning at both host and guest layers; CPU, RAM, storage IOPS, network bandwidth, and NUMA/overcommit constraints can all create contention. Scripts should validate inputs, handle errors/exit codes, log changes, protect credentials, support safe reruns, and be version-controlled. Asset/lifecycle records should track ownership, serial/service tag, location, configuration, warranty/support, firmware/OS, dependencies, change state, and retirement.

• Using an absolute file-system rule without checking OS/workload/features.

• Partition/file-system questions start with boot mode, OS, workload, resiliency, and feature requirements.

Security and Disaster Recovery protect server confidentiality, integrity, availability, and recoverability across the full lifecycle. Data at rest can be protected with disk/file/database encryption and controlled key management; data in transit uses secure protocols such as TLS/HTTPS, SSH/SFTP, or IPsec rather than cleartext management/file-transfer protocols. Encryption does not replace authorization, key protection, retention, backups, or secure deletion. Physical controls include locks, badges, cameras, racks/cages, visitor controls, and media handling; environmental controls include UPS/generators, cooling, fire detection/suppression, water/leak detection, and monitoring. IAM should use unique identities, groups/roles, least privilege, MFA for privileged/remote access where appropriate, dedicated service accounts, password/credential lifecycle, and rapid deprovisioning. OS hardening includes patching, removing/disable unused services, secure baselines, host firewall, anti-malware/EDR where appropriate, logging/auditing, secure management protocols, least privilege, file permissions, and configuration control. Decommissioning requires retention review, backup/record preservation, credential/key/certificate removal, data sanitization using an approved method appropriate to the media/data, asset/inventory updates, and controlled disposal/reuse. DR begins with business requirements: RPO defines acceptable data loss; RTO defines acceptable recovery time. Backups, replication, alternate systems/sites, runbooks, communication, dependency order, and failover must be tested. A backup that has never been restored is unverified; HA and RAID are not substitutes for backup.

• Confusing encryption with authorization.

• Security questions often need layered controls: identity + encryption + hardening + logging.

Server+ troubleshooting is a structured evidence process: identify and scope the problem, gather symptoms and recent changes, form a theory, test the least disruptive hypothesis, plan and implement the fix, verify full functionality, and document. Hardware faults require separating facility power, PSU/PDU/UPS, cables, POST/diagnostic indicators, memory, CPU, system board, fans/thermals, firmware, and peripheral failures. Storage troubleshooting should trace the whole path: media health, cabling/backplane, HBA/controller/cache, RAID state/rebuild, multipathing, SAN/NAS connectivity, capacity, file-system health, permissions, and application access. OS/application issues are often revealed by logs, service state, resource saturation, recent patches/configuration changes, dependencies, permissions, certificates, DNS, and ports. Network latency/connectivity should be traced from physical/link state through addressing/VLAN, routing, security, DNS, transport ports, MTU, packet loss, and application response—starting at the layer suggested by evidence rather than blindly at Layer 1 every time. DR troubleshooting verifies backup integrity, restore procedures, dependency order, credentials/keys, network/DNS, alternate capacity, communications, and measured RPO/RTO. The goal is not merely to make the symptom disappear; it is to prove the root cause and confirm the repair.

• Replacing hardware before verifying power/cables/logs.

• Scope first: one server, one rack, one VLAN, one application, or everyone?