Question 670
A service must be pinned to specific available CPU cores for an isolated performance test. Which command can inspect or set affinity?
- Single-thread CPU bottleneck
- Load includes tasks waiting in uninterruptible sleep
- taskset
- Raise the batch process nice value
Explanation
Correct Answer
C:taskset controls the CPU affinity mask of a process. Affinity limits scheduling placement but does not reserve exclusive cores.
Incorrect Answers
A:One fully saturated thread can limit application throughput despite aggregate idle CPU capacity; it will not provide restricting process placement to selected CPUs under these conditions.
B:Linux load average counts runnable tasks and certain uninterruptible waits; using it would leave restricting process placement to selected CPUs unresolved.
D:Higher niceness generally lowers scheduling priority for normal CFS tasks; it does not deliver restricting process placement to selected CPUs, which is decisive in this scenario.
Question 671
An on-call engineer suspects a kernel hot path after network throughput dropped. Which profiling utility can sample CPU hotspots?
- pidstat -rud -p PID 1
- iostat -xz 1
- perf record and perf report
- Correlate application timing with block-device latency metrics
Explanation
Correct Answer
C:Linux perf can sample hardware and software performance events to identify costly call paths. Sampling needs suitable kernel permissions and symbol availability.
Incorrect Answers
A:pidstat can sample process CPU, memory and I/O activity by PID over time; the task involves sampling kernel and userspace CPU hotspots, whereas this addresses another subsystem.
B:Extended iostat reports device throughput, utilization and latency-related statistics by interval; its scope differs from sampling kernel and userspace CPU hotspots, so it cannot meet the stated requirement.
D:Matching request latency with I/O latency, throughput and queue depth helps identify storage stalls; the administrator instead needs sampling kernel and userspace CPU hotspots; this option serves a different purpose.
Question 672
A storage workload issues many small random writes and throughput is low despite good sequential benchmarks. Which workload property explains the gap?
- Compare df filesystem fullness with du directory sizes
- Small synchronous write latency
- IOPS and access pattern limit
- Disk I/O service capacity
Explanation
Correct Answer
C:Random small operations stress per-request latency and IOPS rather than maximum sequential bandwidth. Match benchmarks to the real access pattern.
Incorrect Answers
A:df shows allocated filesystem capacity while du attributes visible path consumption; that changes a different part of the system; the requested outcome is comparing random IOPS with sequential throughput.
B:Sequential megabytes per second do not characterize durability-sensitive small writes; this does not achieve comparing random IOPS with sequential throughput on the system described.
D:A busy storage device can accumulate latency as outstanding operations wait for service; using it would leave comparing random IOPS with sequential throughput unresolved.
Question 673
The OS keeps evicting useful cached pages during heavy scanning of large files. What operating-system mechanism is responsible?
- Firmware-reported memory and online memory blocks
- Memory leak
- Page cache reclaim under memory pressure
- Container cgroup memory limit
Explanation
Correct Answer
C:Linux may reclaim page-cache pages to satisfy memory demands. Repeated re-reads can then increase storage I/O.
Incorrect Answers
A:The OS can only use memory that firmware and kernel successfully enumerate and online; it does not deliver explaining cache churn under memory pressure, which is decisive in this scenario.
B:A leak retains memory that is no longer needed, gradually increasing working-set pressure; the evidence calls for explaining cache churn under memory pressure, not merely this separate operation.
D:A memory cgroup can enforce an allocation ceiling independent of host capacity; its scope differs from explaining cache churn under memory pressure, so it cannot meet the stated requirement.
Question 674
A process uses 100% of one core on an eight-core host, while overall CPU usage is modest. What does this imply?
- Load includes tasks waiting in uninterruptible sleep
- Context-switch rate from vmstat or pidstat
- Raise the batch process nice value
- Single-thread CPU bottleneck
- taskset
Explanation
Correct Answer
D:One fully saturated thread can limit application throughput despite aggregate idle CPU capacity. Per-core metrics and process threads reveal the constraint.
Incorrect Answers
A:Linux load average counts runnable tasks and certain uninterruptible waits; the administrator instead needs finding a saturated single CPU core; this option serves a different purpose.
B:Very high context-switch rates can indicate scheduling overhead or contention; it answers a neighboring question rather than finding a saturated single CPU core.
C:Higher niceness generally lowers scheduling priority for normal CFS tasks; that changes a different part of the system; the requested outcome is finding a saturated single CPU core.
E:taskset controls the CPU affinity mask of a process; it will not provide finding a saturated single CPU core under these conditions.
Question 675
A container is OOM-killed while the host still has free RAM. Which scope may explain the behavior?
- Container cgroup memory limit
- Resident set size (RSS)
- Memory pressure causing disk-backed paging
- /proc/PID/smaps_rollup
Explanation
Correct Answer
A:A memory cgroup can enforce an allocation ceiling independent of host capacity. Inspect the container's limit and OOM events.
Incorrect Answers
B:RSS measures physical memory currently resident for a process, though shared pages complicate accounting; the requirement is diagnosing memory failure within a constrained cgroup, outside the effect of this choice.
C:Swapping moves memory pages between RAM and backing storage, adding latency when working sets exceed available memory; it does not deliver diagnosing memory failure within a constrained cgroup, which is decisive in this scenario.
D:smaps_rollup provides summarized memory mappings and residency information for a process; the evidence calls for diagnosing memory failure within a constrained cgroup, not merely this separate operation.
Question 676
A server has a high load average but low CPU utilization and many tasks in D state. What is the strongest inference?
- Single-thread CPU bottleneck
- CPU frequency scaling and throttling
- User versus system CPU time
- Raise the batch process nice value
- Load includes tasks waiting in uninterruptible sleep
- taskset
Explanation
Correct Answer
E:Linux load average counts runnable tasks and certain uninterruptible waits. High load does not always mean CPU saturation.
Incorrect Answers
A:One fully saturated thread can limit application throughput despite aggregate idle CPU capacity; its scope differs from interpreting high load with low busy CPU, so it cannot meet the stated requirement.
B:Frequency governors and thermal or power limits change effective CPU capacity; that is not equivalent to interpreting high load with low busy CPU.
C:System CPU accounts for kernel execution on behalf of tasks; the administrator instead needs interpreting high load with low busy CPU; this option serves a different purpose.
D:Higher niceness generally lowers scheduling priority for normal CFS tasks; it answers a neighboring question rather than interpreting high load with low busy CPU.
F:taskset controls the CPU affinity mask of a process; this does not achieve interpreting high load with low busy CPU on the system described.
Question 677
The system repeatedly kills a Java process and dmesg identifies it as an OOM victim. What should be inspected?
- Resident set size (RSS)
- Memory allocation and OOM logs
- Container cgroup memory limit
- Firmware-reported memory and online memory blocks
Explanation
Correct Answer
B:OOM messages show allocation context and selected processes. Container limits, memory leaks and total RAM pressure are plausible causes.
Incorrect Answers
A:RSS measures physical memory currently resident for a process, though shared pages complicate accounting; this does not achieve tracing a process killed by memory exhaustion on the system described.
C:A memory cgroup can enforce an allocation ceiling independent of host capacity; using it would leave tracing a process killed by memory exhaustion unresolved.
D:The OS can only use memory that firmware and kernel successfully enumerate and online; the requirement is tracing a process killed by memory exhaustion, outside the effect of this choice.
Question 678
A host runs slower after power-policy changes although process counts are unchanged. Which hardware behavior should be checked?
- CPU frequency scaling and throttling
- Raise the batch process nice value
- taskset
- Context-switch rate from vmstat or pidstat
Explanation
Correct Answer
A:Frequency governors and thermal or power limits change effective CPU capacity. Clock-rate evidence can explain latency without high load percentages.
Incorrect Answers
B:Higher niceness generally lowers scheduling priority for normal CFS tasks; the task involves detecting reduced CPU clock performance, whereas this addresses another subsystem.
C:taskset controls the CPU affinity mask of a process; its scope differs from detecting reduced CPU clock performance, so it cannot meet the stated requirement.
D:Very high context-switch rates can indicate scheduling overhead or contention; that is not equivalent to detecting reduced CPU clock performance.
Question 679
An engineer suspects excessive context switching from thousands of tiny tasks. Which metric helps validate it?
- Raise the batch process nice value
- Context-switch rate from vmstat or pidstat
- CPU frequency scaling and throttling
- User versus system CPU time
Explanation
Correct Answer
B:Very high context-switch rates can indicate scheduling overhead or contention. Interpret the rate with workload and CPU utilization.
Incorrect Answers
A:Higher niceness generally lowers scheduling priority for normal CFS tasks; it answers a neighboring question rather than evaluating context-switch overhead.
C:Frequency governors and thermal or power limits change effective CPU capacity; this does not achieve evaluating context-switch overhead on the system described.
D:System CPU accounts for kernel execution on behalf of tasks; it will not provide evaluating context-switch overhead under these conditions.
Question 680
A Linux server begins swapping heavily and user response time degrades even though CPU is not fully used. Which bottleneck is likely?
- /proc/PID/smaps_rollup
- Memory allocation and OOM logs
- Memory leak
- Memory pressure causing disk-backed paging
Explanation
Correct Answer
D:Swapping moves memory pages between RAM and backing storage, adding latency when working sets exceed available memory. Identify competing processes and working-set demands.
Incorrect Answers
A:smaps_rollup provides summarized memory mappings and residency information for a process; the requirement is diagnosing slowdown from memory paging, outside the effect of this choice.
B:OOM messages show allocation context and selected processes; it does not deliver diagnosing slowdown from memory paging, which is decisive in this scenario.
C:A leak retains memory that is no longer needed, gradually increasing working-set pressure; the evidence calls for diagnosing slowdown from memory paging, not merely this separate operation.