Exam objective 3.1 asks you to troubleshoot hardware and storage issues – the most hands-on part of the Server+ exam. This matters because servers are built from physical components that fail: power supplies die, hard drives crash, memory sticks go bad. Without a methodical approach, you could spend hours swapping parts that aren’t faulty, or worse, miss a failing drive until it takes down the entire array.
Jump to a section
A simple way to picture Troubleshooting Hardware and Storage Issues
A copier is a complex machine with many interdependent subsystems. When it jams, you don't immediately call a technician. You first check the paper trays (storage), look at the error display (system logs), listen for odd noises (spindle whine), and test the power button (system power).
Imagine your coworker says the copier won’t print. You open the front cover and see a crumpled sheet blocking the rollers. That’s a drive failure in a RAID array – a single physical problem that stops the whole workflow. You remove the jammed paper (replace the failed drive) and close the cover. The copier automatically attempts to resume the print job (RAID rebuild). But if the copier lacks an internal buffer (cache memory), the lost pages may never recover. Similarly, if the power cord is loose (faulty PSU), the copier may intermittently turn off mid-job. You trace the power cord to a wall outlet that’s also plugged into a noisy space heater (overloaded circuit). You move the heater to another socket (rebalance load) and the copier runs smoothly. Every step mirrors server troubleshooting: isolate the subsystem, test the power source, check storage media, and verify the memory buffer before declaring success.
Troubleshooting hardware and storage issues means figuring out why a physical part of a server isn’t working, then fixing it. Let’s break down the common culprits.
Power Issues A server’s power supply unit (PSU) converts mains AC electricity into the DC voltages the motherboard, drives, and fans need. A failing PSU can cause random shutdowns, failure to boot, or even component damage. Symptoms include a burning smell, no LEDs, or beep codes from the motherboard’s POST (Power-On Self Test). Redundant power supplies (two PSUs in one server) help, but if one fails you still need to replace it promptly.
Drive Issues Hard disk drives (HDDs) and solid-state drives (SSDs) store data. HDDs have spinning platters and read/write heads – they fail mechanically (head crash, motor seizure). SSDs have no moving parts but wear out from write cycles. Common failure signs:
Clicking or grinding noises (HDD)
Slow read/write speeds
S.M.A.R.T. (Self-Monitoring, Analysis and Reporting Technology) warnings – a built-in diagnostic that predicts failure
Bad sectors (areas of the disk that can no longer store data reliably)
Memory Issues RAM (Random Access Memory) holds data the CPU is currently using. Server memory often uses ECC (Error-Correcting Code) which can detect and fix single-bit errors. Symptoms of bad RAM:
System crashes (blue screens, kernel panics)
Random application failures
Beep codes at boot (e.g., one long, two short means memory error)
Memory parity errors in logs
RAID (Redundant Array of Independent Disks) RAID combines multiple physical drives into one logical unit for performance, redundancy, or both. Common levels:
RAID 0: Striping – data is split across drives. High speed, zero redundancy. One drive fails, all data is lost.
RAID 1: Mirroring – data is copied identically to two drives. If one fails, the other continues working.
RAID 5: Striping with parity – data and parity information are distributed across three or more drives. Can survive one drive failure.
RAID 6: Similar to RAID 5 but uses two parity blocks, surviving two simultaneous failures.
RAID 10: A combination of mirroring and striping (minimum four drives). Excellent performance and redundancy.
When a drive fails in a redundant RAID, the array is still online but ‘degraded’. A rebuild process reads data from surviving drives and reconstructs the missing data onto a new replacement drive. Rebuilds stress the remaining drives – a second failure during rebuild can cause total data loss.
Disk Failures and Recovery A disk can fail completely (won’t spin up) or partially (bad sectors). For partial failure, S.M.A.R.T. values like Reallocated Sector Count are warnings. Enterprise drives often have a ‘TLER’ (Time-Limited Error Recovery) feature to avoid being kicked out of a RAID array. Tools like fsck (file system check) or manufacturer diagnostics can isolate the issue.
Why This Exists Before RAID, a single disk failure meant restoring from tape backup – slow and risky. RAID provides hot-swap capability (replace a drive while the server is running) and keeps uptime high. ECC memory prevents silent data corruption in databases and virtualisation hosts. Redundant power supplies let you replace a PSU without a shutdown. All these features exist to maximise availability and data integrity for business-critical systems.
Identify the Symptom
Listen to the user’s description and observe the server: is it unresponsive, making noises, showing error LEDs, or logging critical events? This narrows the component type (power, storage, memory).
Check Power and Physical Connections
Verify the server is plugged in, the PSU switch is on, and cables are secure. If the server has redundant PSUs, check that both are lit. A loose power cord is one of the most common causes of intermittent failure.
Use Built-in Diagnostics
Run POST or view the system logs. Check for beep codes or error codes on the front LCD panel. Enter the BIOS/UEFI and check hardware health (voltages, temperatures, fan speeds).
Isolate the Component
For disk failures, note which bay LED is amber or red. For memory errors, note the DIMM slot reported in logs. If the server boots but crashes, try reseating RAM or testing with one stick at a time.
Replace or Rebuild
If a drive fails in a redundant array, hot-swap it with a compatible replacement and initiate the rebuild. For a dead PSU, replace it with an identical spare. For faulty RAM, replace the stick and verify the slot works.
Verify the Fix
After replacement, confirm the server boots cleanly, the RAID array status shows 'Optimal', and the workload runs without errors. Run a stress test if possible (e.g., disk benchmark, memory test). Document the serial numbers and date.
Imagine you’re the only IT person at a growing e-commerce company. Your boss calls: “The inventory database server is unresponsive. Customers can’t search products.” You walk to the server room, hearing a faint clicking from one of the tower servers.
Step 1 – Observe and Identify The server’s front panel shows a solid amber LED on drive bay 3. The operating system is still running (you can SSH in from your laptop). You check the RAID controller software and see the array is ‘Degraded’ – one drive has failed. The system logs show multiple S.M.A.R.T. errors on that drive over the past week. This is a classic disk failure in a RAID 5 with four drives.
Step 2 – Prepare for Replacement You grab a compatible replacement drive from your spares shelf. It should be the same type (SAS or SATA, same RPM and capacity, ideally same firmware version). You don’t need to power down the server – the drive caddy is hot-swappable.
Step 3 – Replace the Drive You press the release button on the faulty drive caddy, slide it out, and immediately insert the new drive. The sled clicks in. The RAID controller detects the new drive and starts a rebuild automatically (or you initiate it via software). The rebuild LED flashes.
Step 4 – Monitor Rebuild The rebuild takes about 2 hours for a 1TB drive. During this time, the array operates in degraded mode – the remaining three drives handle all I/O and parity calculations. You keep an eye on the rebuild progress and watch the server’s temperature; a rebuild generates heat. You also ensure no other drives fail (unlikely but possible if they were all same age).
Step 5 – Verify and Document Once the rebuild finishes, the array status returns to ‘Optimal’. You run a file system check and test a few reads. The database is accessible again. You update your asset tracking with the failed drive’s serial number and replacement date. Then you create a ticket to RMA the failed drive.
Parallel Scenario – Power Issue Another week, the same server starts rebooting every few hours at random times. You check the UPS logs – no power events. You open the server chassis and notice the LED on one PSU is dark. The server is still running on the secondary PSU, which is now overloaded. You replace the failed PSU with a compatible spare, and the random reboots stop. The lesson: always check redundant components before blaming the motherboard or CPU.
The SK0-005 exam loves to test your ability to troubleshoot methodically, not just memorise part numbers. Here’s what to expect.
Question Types - Multiple-choice single answer: e.g., “Which RAID level provides the best fault tolerance?” - Multiple-response: “Select two symptoms of a failing power supply.” - Drag-and-drop: Order the steps to replace a failed drive in a hot-swap RAID array. - Performance-based: You might be given a simulated scenario where you choose the next diagnostic step.
Concepts They Love to Test - RAID levels – know data capacity and fault tolerance for RAID 0, 1, 5, 6, 10. Example trap: RAID 5 can survive one failure but rebuild stress can cause a second failure – they ask “what is a risk during a RAID 5 rebuild?” - Memory error symptoms – beep codes, parity errors, system hangs. They may give a POST beep pattern and ask what component is faulty. - Power supply diagnostics – PSU failure symptoms vs. motherboard failure vs. cable issues. If the server has no power at all, you must check the wall outlet, PSU switch, power cord, and PSU itself in that order. - Hot-swap vs. cold-swap – they’ll test when you can remove a component without powering down. Hot-swap: drives, PSUs, fans. Cold-swap: motherboard, CPU, RAM (in most servers). - S.M.A.R.T. attributes – know what S.M.A.R.T. is (self-monitoring) and that it predicts failures, not guarantees them. - Troubleshooting order – the CompTIA troubleshooting methodology: identify symptom, determine scope, establish theory of cause, test theory, plan action, implement, verify, document.
Traps to Avoid - Assuming RAID is a backup – it’s not. Data corruption, accidental deletion, or ransomware will still affect all drives. - Confusing SATA (simpler, cheaper) with SAS (enterprise, dual-port, higher queue depth) – exam will ask which is used in servers (SAS for critical, SATA for large capacity). - Thinking ‘degraded’ means the array has failed – it means one drive is out, but the array is still functional until a second failure. - Ignoring the role of the RAID controller battery – if the battery is dead, write-back caching may be disabled, causing performance drop.
Key Definitions to Memorise - POST – Power-On Self Test, runs when server starts. - Beep codes – audible patterns indicating hardware faults (e.g., 1 long, 2 short = memory). - MTBF – Mean Time Between Failures, a reliability metric. - UPS – Uninterruptible Power Supply, provides battery backup and surge protection. - RAID rebuild – process of recreating data from parity/mirror onto a new drive.
RAID provides fault tolerance against drive failure, not data corruption, accidental deletion, or ransomware.
A server with a failed PSU can still run on a redundant second unit, but random reboots may occur if the load exceeds the remaining PSU's capacity.
POST beep codes are a quick diagnostic tool – learn common patterns for memory, video, and CPU failures.
S.M.A.R.T. data predicts drive failures; a reallocated sector count increasing over time is a strong indicator to replace the drive.
During a RAID rebuild, the array is vulnerable; never assume it's safe until the rebuild completes and the status shows 'Optimal'.
Hot-swappable components (drives, PSUs, fans) can be replaced without powering down; cold-swap components (motherboard, CPU) require a shutdown.
The first step in hardware troubleshooting is always to physically inspect the server – listen for noises, look for LEDs, and smell for burning.
These come up on the exam all the time. Here's how to tell them apart.
RAID 0
Data is striped across all drives for maximum read/write speed
No fault tolerance – one drive failure loses all data
Usable capacity equals total of all drives
RAID 1
Data is mirrored identically to two or more drives
Survives one drive failure (if two drives mirrored)
Usable capacity equals smallest drive times one (half total)
HDD
Uses spinning magnetic platters and moving read/write heads
Susceptible to mechanical failure and vibration damage
Lower cost per gigabyte, but slower access times
SSD
Uses flash memory chips with no moving parts
Resistant to shock and less prone to sudden failure
Faster read/write speeds, higher cost per gigabyte
ECC Memory
Detects and corrects single-bit memory errors automatically
Used in servers and mission-critical workstations
Slightly higher latency due to error-checking logic
Non-ECC Memory
Cannot detect or correct memory errors
Common in consumer PCs and gaming systems
Lower cost and slightly faster in raw benchmarks
Hot-Swap Drive
Can be replaced while the server is running and powered on
Requires a backplane or controller that supports hot-plugging
Common in enterprise servers with RAID arrays
Cold-Swap Drive
Requires powering down the server before removal
Standard desktop internal drives are cold-swap only
If removed while powered, may cause data corruption or power surge
UPS (Uninterruptible Power Supply)
Provides battery backup so servers keep running during short power outages
Regulates voltage (brownouts, surges) to protect electronics
Requires battery replacement every 3-5 years
Surge Protector
Only diverts dangerous voltage spikes to ground – no battery backup
Cheaper but does not keep servers running if power is lost
Must be replaced if it takes a major surge (absorbed energy degrades protection)
Mistake
RAID 5 is safe because you can lose one drive and everything is fine until the next failure.
Correct
While RAID 5 survives one drive failure, the array is vulnerable during rebuild – the remaining drives are stressed, and a second failure can cause total data loss.
New IT staff often think redundancy means 'no problem', not understanding that rebuilds are high-risk windows.
Mistake
If a server is overheating, you should immediately replace the CPU.
Correct
Overheating is usually caused by failed fans, blocked vents, or dust – not the CPU itself. Check the cooling subsystem first.
Learners equate heat with the hottest component, but CPUs have thermal protection. Cooling failure is far more common.
Mistake
RAID 1 (mirroring) provides the same read performance as RAID 0.
Correct
RAID 1 can read from both disks simultaneously, so reads are faster than a single disk, but writes are slower because both disks must be written to. RAID 0 is faster for both reads and writes.
People assume 'copying everything twice' doubles work, not realising parallel read benefits.
Mistake
ECC memory is only needed for servers running databases – regular desktops don’t benefit.
Correct
ECC memory detects and corrects single-bit errors in any computer where data integrity matters. Desktops doing scientific computing or financial analysis also benefit.
The myth persists because consumer PCs rarely use ECC, leading to an assumption it's not important.
Mistake
A hard drive with bad sectors must be replaced immediately.
Correct
A few bad sectors are acceptable if S.M.A.R.T. values are stable and the drive can reallocate spare sectors. Replacement is needed when the reallocation count climbs rapidly or if the bad sectors affect critical data.
Beginners think any error is catastrophic; they don't understand that drives are designed to handle a limited number of defects.
Reveal each answer, then mark whether you got it right. Score 60%+ to unlock the next chapter.
RAID combines multiple hard drives into one logical unit to improve performance, provide redundancy, or both. You need it in a server to keep the system running if a single hard drive fails, avoiding downtime.
Listen for clicking or grinding noises, check the S.M.A.R.T. status in the BIOS or using a tool like smartctl, and look for bad sector counts increasing. Also, the RAID controller may show a degraded array.
Yes, but the array will use the smallest drive’s capacity for each member. For example, mixing a 1TB and a 2TB drive in RAID 1 gives you only 1TB of usable space. It’s best practice to use identical drives.
A continuous single beep usually indicates a memory failure or that no RAM is detected. Check that the DIMMs are seated properly and try swapping sticks to isolate the faulty one.
Random restarts can be caused by a failing power supply, overheating, or memory errors. Check PSU LEDs and system logs for temperature thresholds or memory parity errors. A UPS with voltage regulation can also reveal power issues.
Hot-swap means you can remove and replace a component (like a hard drive or PSU) while the server is running. Cold-swap requires powering down the server before replacing components like the motherboard or CPU.
You've finished Troubleshooting Hardware and Storage Issues. Continue through the SK0-005 study guide to build a complete picture of the exam.
Done with this chapter?