- RAID
- Redundant Array of Independent Disks — combining multiple drives for performance, redundancy, or both.
- Rack unit (U)
- A unit of vertical rack space equal to 1.75 inches (44.45 mm). A 2U server occupies two of these slots in a standard 19-inch rack.
- Tower server
- A standalone server in an upright chassis like a desktop PC; good for small offices, but consumes floor space and doesn't rack-mount efficiently.
- Rack-mount server
- A server (e.g., 1U, 2U, 4U) that bolts into a standard 19-inch rack to maximize density in a data center.
- Blade server
- A stripped-down server module that slides into a shared blade enclosure, which supplies common power, cooling, and networking to many blades.
- Blade enclosure (chassis)
- The shared housing for blade servers; provides redundant power supplies, cooling, and backplane/network connectivity for all installed blades.
- Form factor
- The physical size and shape of a server: tower, rack-mount (1U/2U/4U), or blade. Chosen for density, scalability, and the environment.
- CPU socket
- The connector on a motherboard that holds the processor; servers often have multiple sockets for multiprocessing.
- Multiprocessing
- Using two or more physical CPUs (multi-socket) so a server can execute more threads in parallel for heavy workloads.
- ECC memory
- Error-Correcting Code RAM that detects and corrects single-bit memory errors; standard in servers for data integrity.
- Registered (buffered) memory
- RDIMM memory with a register between DRAM and the controller to improve stability with large amounts of RAM; common in servers.
- NIC teaming
- Combining multiple network interface cards into one logical interface for higher bandwidth and/or failover redundancy.
- HBA
- Host Bus Adapter — a card that connects a server to storage (e.g., Fibre Channel or SAS), offloading I/O from the CPU.
- RAID 0
- Striping with no parity. Data is split across drives for speed and full capacity, but ANY drive failure loses all data (no fault tolerance).
- RAID 1
- Mirroring — identical copies on two drives. Survives one drive failure; usable capacity is half the raw total.
- RAID 5
- Striping with distributed single parity across 3+ drives. Survives one drive failure; capacity = (n − 1) drives.
- RAID 6
- Striping with double distributed parity across 4+ drives. Survives two simultaneous drive failures; capacity = (n − 2) drives.
- RAID 10 (1+0)
- A stripe of mirrors — combines RAID 1 redundancy with RAID 0 speed. Needs 4+ drives; high performance and fault tolerance.
- Hot spare
- A standby drive that automatically rebuilds a RAID array when a member drive fails, minimizing exposure time.
- Hot-swappable
- A component (drive, power supply, fan) that can be replaced while the server is powered on, with no downtime.
- JBOD
- Just a Bunch Of Disks — drives presented individually with no RAID; no striping, mirroring, or parity.
- SAS
- Serial Attached SCSI — a fast, reliable enterprise drive/interface common in servers; supports dual-porting for redundancy.
- SATA
- Serial ATA — a lower-cost drive interface common for capacity-oriented or less performance-critical server storage.
- NVMe
- Non-Volatile Memory Express — a high-speed protocol for SSDs over PCIe, offering far lower latency than SAS/SATA.
- DAS
- Direct-Attached Storage — storage connected directly to one server (internal disks or an external SAS enclosure).
- NAS
- Network-Attached Storage — file-level storage shared over the LAN using protocols like SMB/CIFS and NFS.
- SAN
- Storage Area Network — block-level storage over a dedicated high-speed network (Fibre Channel or iSCSI).
- iSCSI
- Internet Small Computer System Interface — block storage encapsulated in TCP/IP, letting a SAN run over standard Ethernet.
- Redundant power supply
- A second PSU so the server keeps running if one supply fails; often hot-swappable and on separate circuits.
- PDU
- Power Distribution Unit — distributes power to rack equipment; managed PDUs allow remote monitoring and per-outlet control.
- UPS
- Uninterruptible Power Supply — battery backup that bridges power loss and enables a graceful shutdown.
- BMC / IPMI
- Baseboard Management Controller using IPMI — out-of-band hardware management for remote power, console, and sensor monitoring.
- Out-of-band management
- Managing a server over a dedicated channel (BMC/iLO/iDRAC) independent of the OS, even when the server is off or unresponsive.
- Cable management
- Organizing power and data cables in a rack to maintain airflow, simplify service, and prevent accidental disconnects.
- Hot aisle / cold aisle
- A data-center layout that alternates equipment rows so cold intake air and hot exhaust air are separated for efficient cooling.
- KVM
- Keyboard, Video, Mouse switch — lets one console control many servers; KVM-over-IP adds remote access.
- Hypervisor
- Software that creates and runs virtual machines by abstracting physical hardware. Type 1 runs on bare metal; Type 2 runs on a host OS.
- Type 1 hypervisor
- A bare-metal hypervisor that runs directly on hardware (e.g., ESXi, Hyper-V) for maximum performance; standard in production servers.
- Type 2 hypervisor
- A hosted hypervisor that runs as an application on top of an OS (e.g., VirtualBox); used mainly for testing and labs.
- Virtual machine (VM)
- A software-based computer with its own virtual CPU, memory, disk, and NIC, running on a hypervisor.
- Container
- A lightweight, isolated package of an application and its dependencies that shares the host OS kernel; faster and smaller than a VM.
- Snapshot
- A point-in-time capture of a VM's state so you can roll back after a change; not a substitute for a real backup.
- P2V
- Physical-to-Virtual migration — converting a physical server into a virtual machine.
- Server role
- A primary function a server provides, such as file, print, web, database, DNS, DHCP, or directory services.
- DHCP
- Dynamic Host Configuration Protocol — automatically assigns IP address, subnet mask, gateway, and DNS to clients.
- DNS
- Domain Name System — resolves hostnames to IP addresses; a core server role for name resolution.
- Directory services
- A centralized identity store (e.g., Active Directory, LDAP) for authentication, authorization, and resource management.
- Active Directory
- Microsoft's directory service for managing users, computers, and policies via domains, organizational units, and Group Policy.
- Group Policy
- A Windows feature that centrally enforces configuration and security settings on users and computers in a domain.
- Patch management
- The process of testing and applying OS and firmware updates on a schedule to fix bugs and close security holes.
- Firmware update
- Updating low-level device code (BIOS/UEFI, RAID controller, NIC) to fix issues or add support; test before applying in production.
- Baseline
- A documented record of normal performance (CPU, memory, disk, network) used to detect abnormal behavior later.
- Performance monitoring
- Tracking CPU, memory, disk I/O, and network utilization over time to find bottlenecks and plan capacity.
- Resource pool
- Aggregated CPU and memory from hosts that can be allocated to VMs, often with reservations, limits, and shares.
- Memory overcommit
- Allocating more virtual RAM to VMs than the host physically has, relying on the fact that not all VMs peak at once.
- Thin provisioning
- Allocating storage on demand as data is written, so a volume can be presented larger than the space currently consumed.
- Thick provisioning
- Pre-allocating the full storage capacity up front, guaranteeing space at the cost of immediate consumption.
- Logical volume
- A flexible storage volume (e.g., via LVM) that can span disks and be resized without repartitioning physical drives.
- Mount point
- The directory location where a filesystem or volume is attached and made accessible within the OS.
- Local user account
- An account that exists only on a single server, used when domain authentication is unavailable or not appropriate.
- Service account
- A dedicated, often non-interactive account that runs an application or service; should use least privilege.
- Scripting / automation
- Using shell, PowerShell, or Bash scripts to automate repetitive administration consistently and at scale.
- Documentation
- Maintaining diagrams, asset inventories, configuration records, and runbooks so the environment is understood and recoverable.
- Change management
- A formal process to request, review, approve, schedule, and document changes to reduce risk and outages.
- Asset management
- Tracking hardware and software inventory, including warranties, licenses, and lifecycle, for the entire server estate.
- Licensing
- Tracking and complying with software license terms (per-core, per-socket, per-user, or subscription) for servers and apps.
- NTP
- Network Time Protocol — synchronizes server clocks; accurate time is essential for logging, authentication, and certificates.
- Remote administration
- Managing a server from elsewhere via SSH (Linux), RDP (Windows), or web/console tools; secure the channel.
- SSH
- Secure Shell — an encrypted protocol (TCP 22) for remote command-line administration and secure file transfer.
- RDP
- Remote Desktop Protocol — Microsoft's protocol (TCP 3389) for graphical remote administration of Windows servers.
- High availability (HA)
- Designing systems with redundancy and failover so services keep running despite component failures.
- Clustering
- Linking multiple servers to act as one system for high availability or load balancing; a failed node's load shifts to others.
- Load balancing
- Distributing client requests across multiple servers to improve performance and avoid overloading any single node.
- VM migration (live)
- Moving a running VM between hosts with little or no downtime (e.g., vMotion/Live Migration) for maintenance or balancing.
- CIA triad
- Confidentiality, Integrity, and Availability — the three core goals that guide all information security decisions.
- Least privilege
- Granting users and services only the minimum access required to do their job, limiting damage from compromise or error.
- Role-based access control (RBAC)
- Assigning permissions to roles rather than individuals, then placing users into roles for consistent, manageable access.
- Multifactor authentication (MFA)
- Requiring two or more factors — something you know, have, or are — to authenticate, strongly reducing credential theft risk.
- Hardening
- Reducing a server's attack surface by removing unused services/roles, closing ports, patching, and applying secure baselines.
- Host-based firewall
- Software on the server itself that filters inbound/outbound traffic by rules, complementing network firewalls.
- Data at rest encryption
- Encrypting stored data (full-disk or volume encryption) so drives are unreadable if stolen or improperly disposed.
- Data in transit encryption
- Protecting data moving over the network with TLS, IPsec, or SSH so it can't be read if intercepted.
- HSM
- Hardware Security Module — a tamper-resistant device that securely generates, stores, and uses cryptographic keys, and accelerates crypto.
- TPM
- Trusted Platform Module — a chip that stores keys and measurements to support secure boot and full-disk encryption (e.g., BitLocker).
- Secure boot
- A UEFI feature that verifies the bootloader and OS signatures so unauthorized or tampered code cannot load at startup.
- Physical security
- Controlling physical access to servers with locked racks, badge entry, cameras, and mantraps to prevent tampering and theft.
- Data destruction
- Securely disposing of media by wiping, degaussing, or physical destruction (shredding) so data cannot be recovered.
- Backup
- A copy of data kept separately so it can be restored after loss, corruption, ransomware, or disaster.
- Full backup
- A complete copy of all selected data; slowest to create and largest, but fastest and simplest to restore.
- Incremental backup
- Backs up only data changed since the LAST backup (full or incremental); fast/small but restore needs the full plus every increment in order.
- Differential backup
- Backs up data changed since the last FULL backup; larger than incremental but restore needs only the full plus the latest differential.
- Synthetic full backup
- A full backup assembled from a prior full plus subsequent incrementals, without re-reading all source data.
- 3-2-1 backup rule
- Keep 3 copies of data, on 2 different media types, with 1 copy offsite, to survive most failure and disaster scenarios.
- Backup rotation (GFS)
- Grandfather-Father-Son — a media rotation scheme using daily, weekly, and monthly sets for layered retention.
- Snapshot vs backup
- A snapshot captures point-in-time state on the same system (quick rollback); a true backup is a separate, restorable copy offsite.
- RTO
- Recovery Time Objective — the maximum acceptable time to restore a service after an outage.
- RPO
- Recovery Point Objective — the maximum acceptable amount of data loss, measured as time back to the last good backup.
- MTTR
- Mean Time To Repair — the average time to fix a failed component and restore service.
- MTBF
- Mean Time Between Failures — the average operating time between failures; a measure of reliability.
- Disaster recovery (DR) plan
- A documented, tested process for restoring IT services after a major disruption, defining roles, steps, RTO, and RPO.
- Business continuity plan (BCP)
- A broader plan to keep the whole business operating during and after a disruption; DR is the IT subset of BCP.
- Hot site
- A fully equipped, continuously updated alternate site that can take over almost immediately — the fastest, most expensive option.
- Warm site
- A partially equipped alternate site with hardware and connectivity ready, needing data restore and some setup before takeover.
- Cold site
- A basic facility (space, power, cooling) with no ready equipment; cheapest but slowest to bring online after a disaster.
- Failover
- Automatically switching to a redundant system or site when the primary fails, to maintain availability.
- Failback
- Returning operations to the primary system or site once it is repaired and verified after a failover.
- Replication
- Copying data to another system or site in real time or near-real time so a current copy is always available.
- Backup testing
- Periodically performing test restores to verify backups are complete, valid, and recoverable — an untested backup may not work.
- Air-gapped backup
- A backup kept physically or logically disconnected from the network so ransomware and attackers cannot reach it.
- Patching for security
- Applying security updates promptly to close known vulnerabilities before they can be exploited.
- Troubleshooting methodology
- CompTIA's structured process: identify the problem, theorize a cause, test it, plan, implement/escalate, verify, and document.
- Identify the problem
- The first step: gather information, question users, identify symptoms, and determine whether anything recently changed.
- Establish a theory
- Step two: form a probable cause, questioning the obvious and considering multiple possibilities.
- Test the theory
- Step three: confirm or refute the suspected cause; if not confirmed, form a new theory or escalate.
- Verify functionality
- After fixing, confirm the full system works and implement preventive measures before closing the issue.
- Document the outcome
- The final step: record the cause, the solution, and lessons learned for future reference.
- POST
- Power-On Self-Test — firmware diagnostics at boot; beep codes or POST codes indicate hardware faults like bad RAM or no video.
- Beep codes
- Audible POST signals whose pattern maps to a hardware fault (e.g., memory or video) per the vendor's reference.
- No POST
- A server that won't complete POST — check power, reseat RAM and cards, and look for shorts or a failed component.
- Boot failure
- The OS won't load — possible causes include a bad boot order, failed boot drive, corrupt bootloader, or degraded RAID array.
- Blue Screen / kernel panic
- A fatal OS crash (Windows BSOD or Linux kernel panic) often caused by bad drivers, hardware faults, or corruption.
- Event logs
- OS logs (Windows Event Viewer; Linux /var/log, journalctl) that record errors and warnings to diagnose problems.
- Degraded array
- A RAID array running with a failed member; still operational but unprotected — replace the drive and let it rebuild promptly.
- Array rebuild
- The process of reconstructing data onto a replacement drive from parity or a mirror after a RAID member fails.
- Disk failure signs
- SMART errors, bad sectors, unusual noise, rising latency, or array degradation indicate a failing drive.
- SMART
- Self-Monitoring, Analysis and Reporting Technology — drive self-diagnostics that predict failures before they happen.
- Overheating
- Excessive temperature from failed fans, blocked airflow, or dust; can cause throttling, shutdowns, or hardware damage.
- Thermal shutdown
- An automatic power-off triggered when temperature exceeds safe limits, protecting components from damage.
- Memory error
- Faulty RAM causing crashes or corruption; ECC logs and memory diagnostics help isolate the bad module.
- Network connectivity issue
- No or poor network access — check cabling, link lights, NIC config, IP settings, VLAN, and switch port.
- ping
- Tests reachability and round-trip time using ICMP; a basic first step for network connectivity problems.
- traceroute / tracert
- Maps the per-hop path to a destination to find where traffic stops along the route.
- ipconfig / ifconfig / ip
- Shows a server's IP address, subnet mask, and gateway; reveals misconfiguration or a missing DHCP lease.
- nslookup / dig
- Diagnoses DNS resolution; if you can ping by IP but not by name, suspect DNS.
- Storage capacity issue
- A full disk or LUN causing failures; identify large files/logs, expand the volume, or archive data.
- High CPU utilization
- Sustained high processor use causing slowness — find the offending process, then tune, scale, or schedule the workload.
- Memory leak
- A process that fails to release RAM over time, eventually exhausting memory and degrading or crashing the server.
- Service won't start
- A failed service — check dependencies, configuration, permissions, ports in use, and the related event/log entries.
- Performance bottleneck
- The single resource (CPU, memory, disk I/O, or network) limiting throughput; identified by comparing usage to the baseline.
- Reseat component
- Removing and firmly reinstalling a module (RAM, card, cable) to fix faults caused by a poor connection.
- Swap / page file thrashing
- Excessive paging to disk when RAM is exhausted, causing severe slowdowns; add memory or reduce load.
- Authentication failure
- Users can't log in — check the directory service, time sync (Kerberos), account lockouts, and password/policy issues.
- Cable / connector fault
- A bad or loose data/power cable causing intermittent or no connectivity; test or replace and verify link.
- Change-related issue
- A problem that started right after a change — review and, if needed, roll back the recent change first.
- Escalation
- Handing an issue to someone with more access or expertise when a fix is beyond your scope or authority.
- SFP / SFP+
- Small Form-factor Pluggable transceivers that connect a switch/server port to fiber or copper; SFP+ supports 10 Gbps.
- QSFP
- Quad Small Form-factor Pluggable transceiver for high-speed links (40/100 Gbps) using multiple lanes.
- Single-mode fiber
- Fiber with a tiny core for long-distance, high-bandwidth runs using laser light; uses LC/SC connectors.
- Multimode fiber
- Fiber with a larger core for shorter, high-speed runs; cheaper optics than single-mode.
- RJ45
- The 8-pin modular connector used for twisted-pair Ethernet (Cat 5e/6/6a) copper cabling.
- PCIe
- Peripheral Component Interconnect Express — the high-speed expansion bus for NICs, HBAs, GPUs, and NVMe SSDs.
- Registered (buffered) RAM
- RDIMM memory with a register that improves stability with large RAM capacities; common in servers.
- DIMM
- Dual In-line Memory Module — the physical RAM stick; servers favor ECC and registered variants.
- RAID controller
- Hardware (or firmware) that manages a RAID array, handling striping, mirroring, parity, and rebuilds.
- Rail kit
- Mounting hardware that lets a rack server slide in and out of the rack for installation and service.
- KVM switch
- Keyboard, Video, Mouse switch — lets one console control many servers; KVM-over-IP adds remote access.
- Virtual switch
- A software switch inside a hypervisor that connects VMs to each other and to physical networks.
- vNIC
- Virtual network interface card presented to a VM by the hypervisor.
- VM cloning
- Creating a copy of a VM as a template or for rapid deployment of identical servers.
- Live migration
- Moving a running VM between hosts with little or no downtime (vMotion / Live Migration).
- LVM / logical volume
- A flexible volume that can span disks and be resized without repartitioning physical drives.
- LDAP
- Lightweight Directory Access Protocol — the standard protocol for querying directory services.
- Licensing models
- Per-core, per-socket, per-user (CAL), or subscription terms that servers and apps must comply with.
- Failover cluster
- A cluster where a surviving node takes over a failed node's workload to maintain availability.
- Synchronous replication
- Replication that confirms a write on both sites before acknowledging — zero data loss, higher latency.
- Asynchronous replication
- Replication that acknowledges locally and copies later — faster, with a small potential data lag.
- Capacity planning
- Forecasting future resource needs from trends so the server is scaled before it runs out.
- Server roles
- Primary functions a server provides: file, print, web, database, DNS, DHCP, and directory services.
- RBAC
- Role-Based Access Control — assigning permissions to roles, then placing users into roles.
- Cryptographic erase
- Destroying a drive's encryption key so the encrypted data becomes unrecoverable — fast sanitization.
- Degaussing
- Erasing magnetic media (HDD/tape) with a strong magnetic field; does not work on SSDs/flash.
- GFS rotation
- Grandfather-Father-Son — a media rotation using daily, weekly, and monthly retention sets.
- Business continuity plan
- A plan to keep the whole business operating during a disruption; DR is its IT subset.
- DR plan
- A documented, tested process to restore IT services after a disaster, defining roles, RTO, and RPO.
- MFA
- Multifactor authentication — requiring two or more factors to log in, blocking most credential theft.
- Data sanitization (clear/purge/destroy)
- NIST SP 800-88 methods of increasing assurance for removing data before media reuse or disposal.
- Patch for security
- Applying security updates promptly to close known vulnerabilities before exploitation.
- Hardening baseline
- A documented secure configuration that defines and maintains a server's expected state.
- Replication for DR
- Copying data to another site in near-real time so a current recoverable copy always exists.
- POST codes
- Numeric/LED codes a server displays during the Power-On Self-Test to indicate the failing stage.
- BSOD / kernel panic
- A fatal OS crash (Windows BSOD or Linux kernel panic) usually from drivers, hardware faults, or corruption.
- Event Viewer
- The Windows tool for reviewing system, application, and security logs to diagnose problems.
- syslog / journalctl
- Linux logging — /var/log files and journalctl — used to investigate errors and service failures.
- Swap / page-file thrashing
- Excessive paging to disk when RAM is exhausted, causing severe slowdowns; add RAM or cut load.
- ipconfig / ip
- Shows a server's IP, mask, and gateway; reveals misconfiguration or a missing DHCP lease.
- APIPA address
- A 169.254.x.x self-assigned address indicating the host couldn't reach a DHCP server.
- Duplex mismatch
- When two link ends disagree on half/full duplex, causing collisions, errors, and poor throughput.