Home Server+ Prep
CompTIA Server+

The Server+ SK0-005 guide

Server+ is vendor-neutral and practical: racking and powering servers, RAID and storage, OS installs and virtualization, backups and disaster recovery, and a lot of troubleshooting. Questions are short scenarios where you pick the most likely cause or the next step. Know the RAID math, the backup types, the port numbers, and CompTIA's troubleshooting order.

01 · Format

What you're walking into

750passing score on a 100–900 scale
≤ 90questions, multiple choice and performance-based
90 minexam time
SK0-005current exam version
  • Performance-based questions (PBQs) come first: drag-and-drop or simulated tasks such as picking the right RAID level, matching ports, or ordering troubleshooting steps. They take longer. You can flag them and come back.
  • Multiple choice is single or multiple answer. Read for words like BEST, FIRST, and MOST likely.
  • CompTIA recommends A+ and about 2 years of server experience. There's no hard prerequisite.
  • This practice exam is 60 questions in 60 minutes, with a 5-question scenario set at the end. That's slightly faster than the real pace, which builds a margin.
02 · Weighting

Domain map

DomainWeightFocus
1.0 Server hardware installation and management18%Racks, power, cabling, RAID, storage types, memory, out-of-band management.
2.0 Server administration30%OS installs, partitions and file systems, networking, roles, virtualization, clustering, licensing, scripting, asset management.
3.0 Security and disaster recovery24%Physical and data security, hardening, media disposal, backups, DR sites, RPO/RTO.
4.0 Troubleshooting28%Methodology plus hardware, storage, OS, network, and security problems.
03 · Schedule

A two-week study plan

Day 1

Baseline

Take the full practice exam, timed. Note your weakest domains.

Days 2–5

Administration & troubleshooting

Together these are 58% of the exam. Memorize the troubleshooting steps and port table early.

Days 6–9

Security, DR & hardware

Backup types, RAID math, DR sites, media disposal. Practice each domain to 80%.

Days 10–13

Full exams

Full timed exams every other day. Flashcards daily, especially ports and RAID.

Day 14

Light review

Numbers, ports, and traps. Then rest.

04 · Hardware · 18%

Hardware cheat sheet

RAIDMin disksSurvivesUsable capacity (n disks of size s)Notes
02Nothingn × sStriping. Fast, no redundancy.
121 disksMirroring. Good for OS volumes.
531 disk(n − 1) × sDistributed parity. Write penalty. Long, risky rebuilds on big disks.
642 disks(n − 2) × sDual parity. Safer for large arrays.
1041 per mirror pair(n ÷ 2) × sStripe of mirrors. Best write performance with redundancy (databases).

Racks & form factors

  • 1U = 1.75 inches. A standard full rack is 42U. Put the heaviest gear (UPS, storage) at the bottom.
  • Tower (small offices), rack (datacenters), blade (a chassis shares power, cooling, and networking among many blades).
  • Hot aisle / cold aisle: server intakes face the cold aisle, exhausts face the hot aisle. Use blanking panels in empty U spaces.
  • Out-of-band management: iDRAC (Dell), iLO (HPE), IPMI/BMC. Power on, console, and hardware logs even when the OS is down. IP KVM does the same for console access.

Power

  • Redundant, hot-swappable PSUs connected to different PDUs on different circuits.
  • UPS types: standby (offline, cheapest, short switch time), line-interactive (regulates voltage), online / double-conversion (always on battery inverter, zero transfer time, best for servers).
  • Size a UPS in VA/watts for the load plus headroom. A generator covers long outages, and the UPS bridges the gap until it starts.
  • Calculate amperage: watts ÷ volts = amps. Don't exceed about 80% of a circuit's rating.

Cabling & network hardware

  • Copper: Cat 5e 1 Gbps at 100 m. Cat 6 10 Gbps up to 55 m. Cat 6a 10 Gbps at 100 m.
  • Fiber: single-mode (long distance, laser) vs multimode (short distance, cheaper). Connectors: LC (small, common in servers), SC, ST.
  • Transceivers: SFP (1G), SFP+ (10G), SFP28 (25G), QSFP (40/100G). Twinax DAC cables for short in-rack runs.
  • Label both ends of every cable and keep a cabling diagram.

Storage

  • Interfaces: SATA (cheap, single port), SAS (enterprise, dual-ported, higher RPM), NVMe (SSD over PCIe, lowest latency).
  • HDD speeds: 7.2k, 10k, 15k RPM. SSDs have no moving parts but finite write endurance.
  • DAS: directly attached. NAS: file-level over the network (SMB/NFS). SAN: block-level over iSCSI, Fibre Channel, or FCoE.
  • SAN security: zoning on the FC switch and LUN masking on the array control which hosts see which LUNs. HBAs connect servers to FC.
  • Hot spare: an idle disk that automatically replaces a failed one. RAID controller cache needs a battery or flash backup to protect write-back cache on power loss.

CPU, memory & firmware

  • ECC memory detects and corrects single-bit errors. Servers use registered (RDIMM) or load-reduced (LRDIMM) modules. Don't mix types.
  • Populate DIMMs across channels per the board manual for best performance.
  • Update firmware (BIOS/UEFI, BMC, RAID controller, NIC, drive) in the vendor's order, after a backup, during a maintenance window.
  • UEFI supports Secure Boot and GPT boot disks over 2 TB. Legacy BIOS uses MBR.
05 · Administration · 30%

Administration cheat sheet

OS installation

  • Methods: media, PXE network boot (needs DHCP options and a deployment server), image-based, unattended answer files, or a VM template.
  • MBR: up to 2 TB, 4 primary partitions. GPT: very large disks, 128 partitions in Windows, needs UEFI to boot.
  • File systems: NTFS, ReFS (Windows), ext4, XFS (Linux), ZFS (integrity and snapshots), VMFS (VMware datastores).
  • Server Core / minimal installs reduce attack surface and patching.

Networking

  • NIC teaming: active-active (load balancing, often LACP / 802.3ad with switch support) or active-passive (failover).
  • VLANs separate traffic. A trunk port carries several tagged VLANs (802.1Q).
  • Servers usually get static IPs or DHCP reservations. A 169.254.x.x address means DHCP failed.
  • Private ranges: 10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16.

Virtualization

  • Type 1 (bare metal: ESXi, Hyper-V, KVM) for servers. Type 2 (hosted on an OS: VirtualBox, Workstation) for desktops and labs.
  • Overcommitting CPU and memory raises density but risks contention. Watch host metrics.
  • Snapshots are short-term rollback points, not backups. They grow and slow I/O.
  • Templates and clones speed deployment. P2V converts physical servers, V2V moves between hypervisors.
  • Virtual switches: external, internal, private. VM NICs can tag VLANs.

Availability

  • Clustering: active-active (all nodes serve) or active-passive (standby takes over). Nodes use a heartbeat, and a quorum/witness prevents split-brain.
  • Load balancing: round robin, least connections, weighted. Health checks remove failed nodes.
  • Fault tolerance = no interruption (RAID, redundant PSUs). High availability = minimal downtime (failover).

Licensing & lifecycle

  • Models: per core, per socket, per instance, per user or per device (CALs), concurrent user, site, subscription. Virtualization can change license counts.
  • Asset management: asset tags, inventory database, serial numbers, warranty and support dates.
  • Lifecycle: procure → deploy → maintain → decommission (wipe or destroy data, update inventory, recycle per policy).
  • Documentation: network diagrams, rack diagrams, performance baselines, runbooks, change logs, SLAs.

Scripting & monitoring

  • Script types: Bash (.sh), PowerShell (.ps1), batch (.bat/.cmd), Python, VBScript. Know variables, loops, conditionals, and comments.
  • Use scripts for repeatable tasks: user creation, log rotation, backups, scheduled with cron or Task Scheduler.
  • Monitoring: SNMP (traps 162), syslog (514), event logs, thresholds and alerts, uptime checks. Compare to a baseline.
  • Patch management: test, schedule in a maintenance window, back up first, keep a rollback plan.
06 · Security & disaster recovery · 24%

Security & DR cheat sheet

Backup typeBacks upArchive bitRestore needs
FullEverythingClearedThat one backup
IncrementalChanges since the last backup of any typeClearedLast full + every incremental since. Fastest backup, slowest restore.
DifferentialChanges since the last fullNot clearedLast full + latest differential. Grows each day.
Synthetic fullBuilt from a full plus incrementals on the backup serverN/AOne backup, without a full read of the source
SnapshotPoint-in-time stateN/AUsually on the same storage, so not a real backup alone

Backup practice

  • 3-2-1 rule: 3 copies, on 2 types of media, 1 offsite (or immutable/offline).
  • GFS rotation: daily (son), weekly (father), monthly (grandfather) sets.
  • Test restores regularly. An untested backup isn't a backup.
  • Media: tape (LTO) for cheap long-term and offline copies, disk for fast restores, cloud for offsite.

DR sites & metrics

  • Hot site: running hardware and current data, failover in minutes. Most expensive.
  • Warm site: hardware and connectivity, data restored from backups. Hours to days.
  • Cold site: space, power, and cooling only. Weeks. Cheapest.
  • RPO: maximum acceptable data loss (time). RTO: maximum acceptable downtime.
  • MTBF: average time between failures. MTTR: average time to repair.
  • Replication: synchronous (no data loss, short distance) vs asynchronous (long distance, some data loss).

Physical & environmental

  • Layers: fences, guards, badges, access control vestibule (mantrap), biometrics, locked racks and cages, cameras, bollards.
  • Fire suppression for server rooms: clean agent (FM-200, Novec, inert gas). Avoid water sprinklers, or use pre-action dry pipe if required.
  • Temperature roughly 18–27 °C (64–81 °F). Control humidity: too low causes ESD, too high causes condensation.

Hardening

  • Disable unused services, ports, and accounts. Change default passwords. Remove unneeded software.
  • Host firewall, antivirus/EDR, patching, least privilege, separate admin accounts, MFA.
  • BIOS/UEFI password, Secure Boot, disable boot from USB, lock the chassis (intrusion detection).
  • Service accounts: unique, least privilege, non-interactive, rotated passwords.
  • Encrypt data at rest (BitLocker, LUKS, self-encrypting drives) and in transit (TLS, SSH, IPsec).

Media disposal

  • Overwriting/wiping: reuse the drive. Crypto-erase: destroy the encryption key (fast, works on SSDs).
  • Degaussing: destroys data on magnetic media (HDD, tape). Doesn't work on SSDs.
  • Physical destruction: shredding, crushing, incineration. Get a certificate of destruction from third parties.
  • Follow data retention policies before disposal. Legal holds override them.
07 · Troubleshooting · 28%

Troubleshooting cheat sheet

  1. Identify the problem and determine its scope. Question users, find recent changes, check logs, duplicate it, and back up before making changes.
  2. Establish a theory of probable cause. Question the obvious first.
  3. Test the theory. If it's wrong, form a new theory or escalate.
  4. Establish a plan of action and notify affected users (change management).
  5. Implement the solution or escalate.
  6. Verify full system functionality and implement preventive measures.
  7. Perform a root cause analysis.
  8. Document findings, actions, and outcomes.

Hardware symptoms

  • Beep codes or POST errors at boot → check the code against the vendor's list (often memory or CPU).
  • Random reboots or shutdowns → overheating (dust, failed fan), failing PSU.
  • Clock resets after power loss → CMOS battery.
  • Corrected memory errors in logs → a DIMM is failing. Replace it before it becomes uncorrectable.
  • Predictive failure / SMART warnings → replace the drive proactively.
  • Amber lights on the chassis → check the BMC (iDRAC/iLO) hardware log.

Storage symptoms

  • Degraded array: one disk failed. Replace it and let the rebuild finish. Performance drops during rebuild.
  • Server can't see a new LUN → zoning, LUN masking, or the initiator not rescanned.
  • Disk full → logs, temp files, or snapshots growing. Check with df -h / du or disk management.
  • File system corruption → chkdsk (Windows) or fsck (Linux, on an unmounted volume).
  • High disk latency → wrong RAID level for the workload, failing disk, or cache disabled (dead cache battery).

OS & software symptoms

  • Service won't start → dependency stopped, or the service account password expired or changed.
  • Memory use grows until a crash → memory leak. Restart and patch the app.
  • App broke after an update → roll back the patch, then test in staging.
  • Linux: systemctl status, journalctl, /var/log, top. Windows: Event Viewer, Task Manager, Resource Monitor, services.msc.
  • Permissions: share + NTFS combine, and the most restrictive wins. On Linux, check owner, group, mode bits, and SELinux context.

Network symptoms

  • 169.254.x.x → DHCP unreachable.
  • Ping by IP works but not by name → DNS.
  • Can reach local subnet but not others → default gateway or routing.
  • Intermittent connectivity → duplicate IP, speed/duplex mismatch, bad cable.
  • Port closed → host firewall or network ACL. Test with Test-NetConnection -Port, telnet, nc, or nmap.
  • Tools: ping, tracert/traceroute, nslookup/dig, ipconfig/ip addr, netstat/ss, arp, route, Wireshark/tcpdump.
08 · Recall

Ports to know

PortServicePortService
20/21FTP161/162SNMP / SNMP traps (UDP)
22SSH, SFTP, SCP389LDAP
23Telnet443HTTPS
25SMTP445SMB
53DNS514Syslog (UDP)
67/68DHCP (UDP)636LDAPS
69TFTP (UDP)1433Microsoft SQL Server
80HTTP3306MySQL
88Kerberos3389RDP
123NTP (UDP)623IPMI (UDP)
09 · Recall

Numbers to memorize

ItemValue
1 rack unit (U)1.75 in (44.45 mm)
Standard full rack42U
RAID 5 / 6 / 10 minimum disks3 / 4 / 4
Cat 6 at 10 Gbps / Cat 6a at 10 Gbps55 m / 100 m
MBR maximum disk size / primary partitions2 TB / 4
APIPA range169.254.0.0/16
Circuit load guideline≤ 80% of rating
Recommended inlet temperature (ASHRAE)18–27 °C
3-2-1 backup3 copies, 2 media, 1 offsite
10 · Traps

Classic traps

RAID isn't a backup. It doesn't protect against deletion, corruption, or ransomware.
Incremental = fastest backup, slowest restore. Differential = slower backups, restore needs only the full plus the latest differential.
Degaussing doesn't erase SSDs. Use crypto-erase, a secure erase command, or physical destruction.
Snapshots aren't backups. They live on the same storage.
"What should you do FIRST?" usually means gather information or back up, not replace hardware.
Ping by IP works, by name fails → DNS. 169.254 → DHCP.
Online/double-conversion UPS is the answer when the question says "no transfer time".
Hot site = fastest recovery, highest cost. Cold site = cheapest, slowest.
Share and NTFS permissions together → most restrictive wins.
Service fails after a password change → update the service account's stored password.
Heaviest equipment goes at the bottom of the rack.