mcAfee Secure Website
23

CompTIA XK0-006 Bundle

Exam Code: XK0-006

Exam Name CompTIA Linux+

Certification Provider: CompTIA

Corresponding Certification: CompTIA Linux+

XK0-006 Training Materials $19.99

Reliable & Actual Study Materials for XK0-006 Exam Success

The Latest XK0-006 Exam Questions as Experienced in the Actual Test!

  • 24
    Questions & Answers

    XK0-006 Questions & Answers

    160 Questions & Answers

    Includes questions types found on actual exam such as drag and drop, simulation, type in, and fill in the blank.

  • exam =30
    Study Guide

    XK0-006 Study Guide

    1297 PDF Pages

    Study Guide developed by industry experts who have written exams in the past. They are technology-specific IT certification researchers with at least a decade of experience at Fortune 500 companies.

exam =32

CompTIA Linux+ (XK0-006) Practice Tests

Guarantee

Satisfaction Guaranteed

Pass4sure has a remarkable CompTIA Candidate Success record. We're confident of our products and provide no hassle product exchange. That's how confident we are!

99.3% Pass Rate
Total Cost: $154.98
Bundle Price: $134.99

Purchase Individually

  • exam =34
    Questions & Answers

    Questions & Answers

    160 Questions

    $124.99
    exam =35
  • exam =36
    Study Guide

    Study Guide

    1297 PDF Pages

    $29.99

CompTIA XK0-006 Practice Test 68

Question 670

A Linux administrator evaluates: A Linux application times out only after an expired intermediate certificate breaks its secure API connection. Select TWO complementary answers: the most suitable corrective action AND the evidence proving it worked.

  1. Opening a port cannot help if the daemon only binds loopback or is stopped, and a listening socket alone cannot prove external access.
  2. The approved client authenticates to the verified host with a valid key or certificate and rejects an untrusted identity.
  3. Inspect free, vmstat, /proc/meminfo and kernel OOM events to distinguish reclaimable cache, active swapping and actual memory exhaustion.
  4. The intended domain resolves through the approved resolver and the client reaches the same service by its hostname.
  5. Differentiate SSH host key verification, user authorized-key permissions and TLS endpoint certificate errors using targeted client and server evidence.

Explanation

Correct Answers

B:

For the specific observation (A Linux application times out only after an expired intermediate certificate breaks its secure API connection), the acceptance evidence is: The approved client authenticates to the verified host with a valid key or certificate and rejects an untrusted identity. This verifies the action (Differentiate SSH host key verification, user authorized-key permissions and TLS endpoint certificate errors using targeted client and server evidence.) rather than merely proving that an unrelated Linux tool works. It also addresses the known risk: Disabling strict host checks or accepting unknown certificates during an incident can mask a man-in-the-middle attack.

E:

The observed Linux condition is: A Linux application times out only after an expired intermediate certificate breaks its secure API connection. The recommended direct administrative response is: Differentiate SSH host key verification, user authorized-key permissions and TLS endpoint certificate errors using targeted client and server evidence. The relevant failure or security risk is: Disabling strict host checks or accepting unknown certificates during an incident can mask a man-in-the-middle attack. Validate success with: The approved client authenticates to the verified host with a valid key or certificate and rejects an untrusted identity.

Incorrect Answers

A:

Listener and firewall differentiation on Linux (5.3) would demonstrate or implement: Opening a port cannot help if the daemon only binds loopback or is stopped, and a listening socket alone cannot prove external access. The case states: the most suitable corrective action AND the evidence proving it worked. The most direct paired action and test is: The approved client authenticates to the verified host with a valid key or certificate and rejects an untrusted identity.

C:

Memory pressure OOM killer and swapping (5.5) addresses another technical failure: Treating file cache as unavailable RAM or blindly disabling swap can obscure the application responsible for OOM kills. This case instead tests SSH key and certificate failure diagnosis; the choice measures or changes another part of the system.

D:

DNS versus connectivity diagnostic sequence (5.3) would demonstrate or implement: The intended domain resolves through the approved resolver and the client reaches the same service by its hostname. The required paired action and test concerns SSH key and certificate failure diagnosis; a passing result for this alternative could leave that requirement unresolved.

Question 671

A Linux administrator investigates: A four-core VM reports load average 24 while CPU utilization remains low and many tasks wait on disk. Which action BEST addresses the specific Linux technical requirement?

  1. Inspect ip route get, addressing, source routing and gateway reachability to isolate the incorrect forwarding path before changing firewall rules.
  2. Confirm service socket state with ss, test local and remote TCP reachability, and inspect nftables or firewalld rules to isolate the blocking layer.
  3. Use top, vmstat, pidstat and per-core utilization to distinguish CPU saturation, blocked tasks and load-average patterns before tuning.
  4. Differentiate SSH host key verification, user authorized-key permissions and TLS endpoint certificate errors using targeted client and server evidence.

Explanation

Correct Answer

C:

The observed Linux condition is: A four-core VM reports load average 24 while CPU utilization remains low and many tasks wait on disk. The recommended direct administrative response is: Use top, vmstat, pidstat and per-core utilization to distinguish CPU saturation, blocked tasks and load-average patterns before tuning. The relevant failure or security risk is: A high load average alone does not prove CPUs are busy, because tasks in uninterruptible I/O wait may contribute. Validate success with: Comparable workload samples isolate the actual runnable or blocked bottleneck and show improved response after a measured correction.

Incorrect Answers

A:

IP route and asymmetric next-hop analysis (5.3) addresses another technical failure: Incorrect default routes or overlapping prefixes can send traffic to an unreachable next hop even with healthy interfaces. The case states: A four-core VM reports load average 24 while CPU utilization remains low and many tasks wait on disk. The most direct diagnosis is: Use top, vmstat, pidstat and per-core utilization to distinguish CPU saturation, blocked tasks and load-average patterns before tuning.

B:

Listener and firewall differentiation on Linux (5.3) addresses another technical failure: Opening a port cannot help if the daemon only binds loopback or is stopped, and a listening socket alone cannot prove external access. This choice can be useful for that problem, but it does not establish the CPU saturation versus load-average analysis control being assessed here.

D:

SSH key and certificate failure diagnosis (5.4) addresses another technical failure: Disabling strict host checks or accepting unknown certificates during an incident can mask a man-in-the-middle attack. The required diagnosis concerns CPU saturation versus load-average analysis; a passing result for this alternative could leave that requirement unresolved.

Question 672

In the post-change validation, the team sees: A four-core VM reports load average 24 while CPU utilization remains low and many tasks wait on disk. Which evidence MOST directly confirms the appropriate correction succeeded?

  1. Inode utilization returns below the operational threshold and the application creates a new file successfully.
  2. The target service accesses its file with enforcing mode active and no matching denial appears in new audit events.
  3. Comparable workload samples isolate the actual runnable or blocked bottleneck and show improved response after a measured correction.
  4. A controlled service degradation crosses the documented threshold and produces an actionable alert with the relevant host and timestamp.

Explanation

Correct Answer

C:

For the specific observation (A four-core VM reports load average 24 while CPU utilization remains low and many tasks wait on disk), the acceptance evidence is: Comparable workload samples isolate the actual runnable or blocked bottleneck and show improved response after a measured correction. This verifies the action (Use top, vmstat, pidstat and per-core utilization to distinguish CPU saturation, blocked tasks and load-average patterns before tuning.) rather than merely proving that an unrelated Linux tool works. It also addresses the known risk: A high load average alone does not prove CPUs are busy, because tasks in uninterruptible I/O wait may contribute.

Incorrect Answers

A:

Inode exhaustion versus free space (5.2) would demonstrate or implement: Inode utilization returns below the operational threshold and the application creates a new file successfully. The case states: A four-core VM reports load average 24 while CPU utilization remains low and many tasks wait on disk. The most direct evidence is: Comparable workload samples isolate the actual runnable or blocked bottleneck and show improved response after a measured correction.

B:

SELinux denial versus traditional permission fault (5.4) would demonstrate or implement: The target service accesses its file with enforcing mode active and no matching denial appears in new audit events. This choice can be useful for that problem, but it does not establish the CPU saturation versus load-average analysis control being assessed here.

D:

SLO-based monitoring and actionable alert thresholds (5.1) would demonstrate or implement: A controlled service degradation crosses the documented threshold and produces an actionable alert with the relevant host and timestamp. The required evidence concerns CPU saturation versus load-average analysis; a passing result for this alternative could leave that requirement unresolved.

Question 673

A change review on a Linux server identifies: A batch process consumes nearly all allocated CPU and interactive web requests slow down. Which action BEST addresses the specific Linux technical requirement?

  1. Compare df -h and df -i, locate directories with excessive small-file creation and remediate safely according to retention policy.
  2. Correlate AVC records, file labels and discretionary file ownership before selecting a context-specific SELinux remediation.
  3. Use top, vmstat, pidstat and per-core utilization to distinguish CPU saturation, blocked tasks and load-average patterns before tuning.
  4. Define appropriate service-level indicators and objectives, collect representative metrics and logs, and alert on meaningful sustained symptoms.

Explanation

Correct Answer

C:

The observed Linux condition is: A batch process consumes nearly all allocated CPU and interactive web requests slow down. The recommended direct administrative response is: Use top, vmstat, pidstat and per-core utilization to distinguish CPU saturation, blocked tasks and load-average patterns before tuning. The relevant failure or security risk is: A high load average alone does not prove CPUs are busy, because tasks in uninterruptible I/O wait may contribute. Validate success with: Comparable workload samples isolate the actual runnable or blocked bottleneck and show improved response after a measured correction.

Incorrect Answers

A:

Inode exhaustion versus free space (5.2) addresses another technical failure: A filesystem can reject new files when its inode supply is exhausted even if many data blocks remain free. The case states: A batch process consumes nearly all allocated CPU and interactive web requests slow down. The most direct corrective action is: Use top, vmstat, pidstat and per-core utilization to distinguish CPU saturation, blocked tasks and load-average patterns before tuning.

B:

SELinux denial versus traditional permission fault (5.4) addresses another technical failure: Turning SELinux off or using chmod 777 cannot reliably solve a labeling issue and can weaken host security. This choice can be useful for that problem, but it does not establish the CPU saturation versus load-average analysis control being assessed here.

D:

SLO-based monitoring and actionable alert thresholds (5.1) addresses another technical failure: An alert based only on an arbitrary single CPU spike can create noise while missing a real service availability breach. The required corrective action concerns CPU saturation versus load-average analysis; a passing result for this alternative could leave that requirement unresolved.

Question 674

A Linux administrator reviews a reported issue: A batch process consumes nearly all allocated CPU and interactive web requests slow down. Which failure mode or risk is MOST relevant to this condition?

  1. Opening a port cannot help if the daemon only binds loopback or is stopped, and a listening socket alone cannot prove external access.
  2. An encapsulation overhead mismatch can blackhole larger packets even when small ICMP tests pass.
  3. A filesystem can reject new files when its inode supply is exhausted even if many data blocks remain free.
  4. A high load average alone does not prove CPUs are busy, because tasks in uninterruptible I/O wait may contribute.

Explanation

Correct Answer

D:

The reported situation (A batch process consumes nearly all allocated CPU and interactive web requests slow down) most directly creates this failure risk: A high load average alone does not prove CPUs are busy, because tasks in uninterruptible I/O wait may contribute. The corresponding safe response is: Use top, vmstat, pidstat and per-core utilization to distinguish CPU saturation, blocked tasks and load-average patterns before tuning. An appropriate outcome check is: Comparable workload samples isolate the actual runnable or blocked bottleneck and show improved response after a measured correction.

Incorrect Answers

A:

Listener and firewall differentiation on Linux (5.3) would demonstrate or implement: Opening a port cannot help if the daemon only binds loopback or is stopped, and a listening socket alone cannot prove external access. The case states: A batch process consumes nearly all allocated CPU and interactive web requests slow down. The most direct risk is: A high load average alone does not prove CPUs are busy, because tasks in uninterruptible I/O wait may contribute.

B:

MTU and path fragmentation troubleshooting (5.3) would demonstrate or implement: An encapsulation overhead mismatch can blackhole larger packets even when small ICMP tests pass. This choice can be useful for that problem, but it does not establish the CPU saturation versus load-average analysis control being assessed here.

C:

Inode exhaustion versus free space (5.2) would demonstrate or implement: A filesystem can reject new files when its inode supply is exhausted even if many data blocks remain free. This case instead tests CPU saturation versus load-average analysis; the choice measures or changes another part of the system.

Question 675

A production incident ticket states: An operations dashboard graphs system load but no one inspects the runnable queue or per-process CPU. Which action BEST addresses the specific Linux technical requirement?

  1. Confirm service socket state with ss, test local and remote TCP reachability, and inspect nftables or firewalld rules to isolate the blocking layer.
  2. Compare df -h and df -i, locate directories with excessive small-file creation and remediate safely according to retention policy.
  3. Use top, vmstat, pidstat and per-core utilization to distinguish CPU saturation, blocked tasks and load-average patterns before tuning.
  4. Measure effective path MTU with safe probes and interface counters, then adjust supported MTU or TCP MSS settings where warranted.

Explanation

Correct Answer

C:

The observed Linux condition is: An operations dashboard graphs system load but no one inspects the runnable queue or per-process CPU. The recommended direct administrative response is: Use top, vmstat, pidstat and per-core utilization to distinguish CPU saturation, blocked tasks and load-average patterns before tuning. The relevant failure or security risk is: A high load average alone does not prove CPUs are busy, because tasks in uninterruptible I/O wait may contribute. Validate success with: Comparable workload samples isolate the actual runnable or blocked bottleneck and show improved response after a measured correction.

Incorrect Answers

A:

Listener and firewall differentiation on Linux (5.3) addresses another technical failure: Opening a port cannot help if the daemon only binds loopback or is stopped, and a listening socket alone cannot prove external access. The case states: An operations dashboard graphs system load but no one inspects the runnable queue or per-process CPU. The most direct corrective action is: Use top, vmstat, pidstat and per-core utilization to distinguish CPU saturation, blocked tasks and load-average patterns before tuning.

B:

Inode exhaustion versus free space (5.2) addresses another technical failure: A filesystem can reject new files when its inode supply is exhausted even if many data blocks remain free. This choice can be useful for that problem, but it does not establish the CPU saturation versus load-average analysis control being assessed here.

D:

MTU and path fragmentation troubleshooting (5.3) addresses another technical failure: An encapsulation overhead mismatch can blackhole larger packets even when small ICMP tests pass. The required corrective action concerns CPU saturation versus load-average analysis; a passing result for this alternative could leave that requirement unresolved.

Question 676

An infrastructure engineer checks a specific situation: An operations dashboard graphs system load but no one inspects the runnable queue or per-process CPU. Which evidence MOST directly confirms the appropriate correction succeeded?

  1. An authorized process reaches the intended path, while unrelated users still fail read or execute tests.
  2. Representative large and small application payloads traverse the path without fragmentation-related loss or excessive retransmits.
  3. The intended domain resolves through the approved resolver and the client reaches the same service by its hostname.
  4. Comparable workload samples isolate the actual runnable or blocked bottleneck and show improved response after a measured correction.

Explanation

Correct Answer

D:

For the specific observation (An operations dashboard graphs system load but no one inspects the runnable queue or per-process CPU), the acceptance evidence is: Comparable workload samples isolate the actual runnable or blocked bottleneck and show improved response after a measured correction. This verifies the action (Use top, vmstat, pidstat and per-core utilization to distinguish CPU saturation, blocked tasks and load-average patterns before tuning.) rather than merely proving that an unrelated Linux tool works. It also addresses the known risk: A high load average alone does not prove CPUs are busy, because tasks in uninterruptible I/O wait may contribute.

Incorrect Answers

A:

ACL mask and directory traversal permission failures (5.4) would demonstrate or implement: An authorized process reaches the intended path, while unrelated users still fail read or execute tests. The case states: An operations dashboard graphs system load but no one inspects the runnable queue or per-process CPU. The most direct evidence is: Comparable workload samples isolate the actual runnable or blocked bottleneck and show improved response after a measured correction.

B:

MTU and path fragmentation troubleshooting (5.3) would demonstrate or implement: Representative large and small application payloads traverse the path without fragmentation-related loss or excessive retransmits. This choice can be useful for that problem, but it does not establish the CPU saturation versus load-average analysis control being assessed here.

C:

DNS versus connectivity diagnostic sequence (5.3) would demonstrate or implement: The intended domain resolves through the approved resolver and the client reaches the same service by its hostname. This case instead tests CPU saturation versus load-average analysis; the choice measures or changes another part of the system.

Question 677

An authorized Linux operations assessment observes: A sudden context-switch increase coincides with a server reporting high latency during peak activity. Which action BEST addresses the specific Linux technical requirement?

  1. Test numeric-IP reachability, local resolver configuration and DNS answers separately with getent, dig and suitable connection tools.
  2. Use top, vmstat, pidstat and per-core utilization to distinguish CPU saturation, blocked tasks and load-average patterns before tuning.
  3. Measure effective path MTU with safe probes and interface counters, then adjust supported MTU or TCP MSS settings where warranted.
  4. Inspect id, namei -l, stat and getfacl to diagnose effective file access, including parent directory traversal and ACL masks.

Explanation

Correct Answer

B:

The observed Linux condition is: A sudden context-switch increase coincides with a server reporting high latency during peak activity. The recommended direct administrative response is: Use top, vmstat, pidstat and per-core utilization to distinguish CPU saturation, blocked tasks and load-average patterns before tuning. The relevant failure or security risk is: A high load average alone does not prove CPUs are busy, because tasks in uninterruptible I/O wait may contribute. Validate success with: Comparable workload samples isolate the actual runnable or blocked bottleneck and show improved response after a measured correction.

Incorrect Answers

A:

DNS versus connectivity diagnostic sequence (5.3) addresses another technical failure: A working route to a server cannot compensate for a stale hosts entry or unresponsive name server. The case states: A sudden context-switch increase coincides with a server reporting high latency during peak activity. The most direct corrective action is: Use top, vmstat, pidstat and per-core utilization to distinguish CPU saturation, blocked tasks and load-average patterns before tuning.

C:

MTU and path fragmentation troubleshooting (5.3) addresses another technical failure: An encapsulation overhead mismatch can blackhole larger packets even when small ICMP tests pass. This case instead tests CPU saturation versus load-average analysis; the choice measures or changes another part of the system.

D:

ACL mask and directory traversal permission failures (5.4) addresses another technical failure: Granting world access to the final file may not fix a missing execute permission on a parent directory and exposes unnecessary data. The required corrective action concerns CPU saturation versus load-average analysis; a passing result for this alternative could leave that requirement unresolved.

Question 678

In a controlled system validation, the observation is: A sudden context-switch increase coincides with a server reporting high latency during peak activity. Which failure mode or risk is MOST relevant to this condition?

  1. Opening a port cannot help if the daemon only binds loopback or is stopped, and a listening socket alone cannot prove external access.
  2. Disabling strict host checks or accepting unknown certificates during an incident can mask a man-in-the-middle attack.
  3. A filesystem can reject new files when its inode supply is exhausted even if many data blocks remain free.
  4. A high load average alone does not prove CPUs are busy, because tasks in uninterruptible I/O wait may contribute.

Explanation

Correct Answer

D:

The reported situation (A sudden context-switch increase coincides with a server reporting high latency during peak activity) most directly creates this failure risk: A high load average alone does not prove CPUs are busy, because tasks in uninterruptible I/O wait may contribute. The corresponding safe response is: Use top, vmstat, pidstat and per-core utilization to distinguish CPU saturation, blocked tasks and load-average patterns before tuning. An appropriate outcome check is: Comparable workload samples isolate the actual runnable or blocked bottleneck and show improved response after a measured correction.

Incorrect Answers

A:

Listener and firewall differentiation on Linux (5.3) would demonstrate or implement: Opening a port cannot help if the daemon only binds loopback or is stopped, and a listening socket alone cannot prove external access. The case states: A sudden context-switch increase coincides with a server reporting high latency during peak activity. The most direct evidence is: A high load average alone does not prove CPUs are busy, because tasks in uninterruptible I/O wait may contribute.

B:

SSH key and certificate failure diagnosis (5.4) would demonstrate or implement: Disabling strict host checks or accepting unknown certificates during an incident can mask a man-in-the-middle attack. This choice can be useful for that problem, but it does not establish the CPU saturation versus load-average analysis control being assessed here.

C:

Inode exhaustion versus free space (5.2) would demonstrate or implement: A filesystem can reject new files when its inode supply is exhausted even if many data blocks remain free. This case instead tests CPU saturation versus load-average analysis; the choice measures or changes another part of the system.

Question 679

During a planned service review, engineers find: A support engineer assumes adding RAM will fix a CPU-bound compression task without collecting process metrics. Which action BEST addresses the specific Linux technical requirement?

  1. Confirm service socket state with ss, test local and remote TCP reachability, and inspect nftables or firewalld rules to isolate the blocking layer.
  2. Use top, vmstat, pidstat and per-core utilization to distinguish CPU saturation, blocked tasks and load-average patterns before tuning.
  3. Differentiate SSH host key verification, user authorized-key permissions and TLS endpoint certificate errors using targeted client and server evidence.
  4. Compare df -h and df -i, locate directories with excessive small-file creation and remediate safely according to retention policy.

Explanation

Correct Answer

B:

The observed Linux condition is: A support engineer assumes adding RAM will fix a CPU-bound compression task without collecting process metrics. The recommended direct administrative response is: Use top, vmstat, pidstat and per-core utilization to distinguish CPU saturation, blocked tasks and load-average patterns before tuning. The relevant failure or security risk is: A high load average alone does not prove CPUs are busy, because tasks in uninterruptible I/O wait may contribute. Validate success with: Comparable workload samples isolate the actual runnable or blocked bottleneck and show improved response after a measured correction.

Incorrect Answers

A:

Listener and firewall differentiation on Linux (5.3) addresses another technical failure: Opening a port cannot help if the daemon only binds loopback or is stopped, and a listening socket alone cannot prove external access. The case states: A support engineer assumes adding RAM will fix a CPU-bound compression task without collecting process metrics. The most direct corrective action is: Use top, vmstat, pidstat and per-core utilization to distinguish CPU saturation, blocked tasks and load-average patterns before tuning.

C:

SSH key and certificate failure diagnosis (5.4) addresses another technical failure: Disabling strict host checks or accepting unknown certificates during an incident can mask a man-in-the-middle attack. This case instead tests CPU saturation versus load-average analysis; the choice measures or changes another part of the system.

D:

Inode exhaustion versus free space (5.2) addresses another technical failure: A filesystem can reject new files when its inode supply is exhausted even if many data blocks remain free. The required corrective action concerns CPU saturation versus load-average analysis; a passing result for this alternative could leave that requirement unresolved.

Question 680

A Linux administrator evaluates: A support engineer assumes adding RAM will fix a CPU-bound compression task without collecting process metrics. Select TWO complementary answers: the most suitable corrective action AND the evidence proving it worked.

  1. An encapsulation overhead mismatch can blackhole larger packets even when small ICMP tests pass.
  2. The approved client authenticates to the verified host with a valid key or certificate and rejects an untrusted identity.
  3. Comparable workload samples isolate the actual runnable or blocked bottleneck and show improved response after a measured correction.
  4. Use the bootloader, emergency-mode logs and journalctl -b or prior-boot records to isolate a failed startup dependency before a controlled correction.
  5. Use top, vmstat, pidstat and per-core utilization to distinguish CPU saturation, blocked tasks and load-average patterns before tuning.

Explanation

Correct Answers

C:

For the specific observation (A support engineer assumes adding RAM will fix a CPU-bound compression task without collecting process metrics), the acceptance evidence is: Comparable workload samples isolate the actual runnable or blocked bottleneck and show improved response after a measured correction. This verifies the action (Use top, vmstat, pidstat and per-core utilization to distinguish CPU saturation, blocked tasks and load-average patterns before tuning.) rather than merely proving that an unrelated Linux tool works. It also addresses the known risk: A high load average alone does not prove CPUs are busy, because tasks in uninterruptible I/O wait may contribute.

E:

The observed Linux condition is: A support engineer assumes adding RAM will fix a CPU-bound compression task without collecting process metrics. The recommended direct administrative response is: Use top, vmstat, pidstat and per-core utilization to distinguish CPU saturation, blocked tasks and load-average patterns before tuning. The relevant failure or security risk is: A high load average alone does not prove CPUs are busy, because tasks in uninterruptible I/O wait may contribute. Validate success with: Comparable workload samples isolate the actual runnable or blocked bottleneck and show improved response after a measured correction.

Incorrect Answers

A:

MTU and path fragmentation troubleshooting (5.3) would demonstrate or implement: An encapsulation overhead mismatch can blackhole larger packets even when small ICMP tests pass. The case states: the most suitable corrective action AND the evidence proving it worked. The most direct paired action and test is: Comparable workload samples isolate the actual runnable or blocked bottleneck and show improved response after a measured correction.

B:

SSH key and certificate failure diagnosis (5.4) would demonstrate or implement: The approved client authenticates to the verified host with a valid key or certificate and rejects an untrusted identity. This choice can be useful for that problem, but it does not establish the CPU saturation versus load-average analysis control being assessed here.

D:

Linux boot failure recovery using journal and GRUB evidence (5.2) addresses another technical failure: Blindly reinstalling the kernel or disabling essential units can destroy the evidence and leave the actual root-device fault unresolved. The required paired action and test concerns CPU saturation versus load-average analysis; a passing result for this alternative could leave that requirement unresolved.