EX442 Red Hat Certified Specialist in Performance Tuning exam Practice Questions
Prepare for EX442 with more than an answer.
- Exam fee
- $500 USD
- Time limit
- 240 minutes
- Questions on the exam
- N/A - Hands-on tasks, not fixed question count
- Passing score
- 210 out of 300 (70%) (scale 0-300)
- Level
- Specialist
- Valid for
- 3 years
Domains covered on the exam 7
- Use Utilities to Analyze System Behavior20%
- Monitor and Alter Kernel Behavior15%
- Analyze System and Application Performance20%
- Tune Running Systems20%
- Tune Memory Utilization10%
- Configure Disk and File Subsystems10%
- Tune Network Performance5%
- 1
A high-frequency trading firm is designing a monitoring strategy for their RHEL 8 clusters. They require:
- Sub-second granularity for real-time analysis.
- Low overhead on the monitored systems.
- Historical data retention for 30 days.
- Centralized analysis capability.
Which toolset combination best meets these requirements?
Show answer details
Correct answer: A
PCP is designed exactly for this scenario. It supports high-resolution (sub-second) metrics, has a distributed architecture (PMDAs on nodes, pmlogger for archiving), and is efficient.
sysstat(sar) typically runs at 1-10 minute intervals by default and is less suited for sub-second real-time aggregation across a cluster.graph LR Node1[Trading Node 1 pmcd] -->|metrics| Central[Central Monitor pmlogger] Node2[Trading Node 2 pmcd] -->|metrics| Central Node3[Trading Node 3 pmcd] -->|metrics| Central Central -->|archives| Analysis[Analysis Tool pmchart/Grafana] - 2
You need to modify the
vm.swappinesskernel parameter to10to improve performance for a database workload. This change must persist across reboots. Which of the following is the correct procedure to achieve this?Show answer details
Correct answer: C
The correct method for persistent kernel tuning in RHEL 8 is to place a configuration file in
/etc/sysctl.d/.sysctl -wis only for runtime.~/.bashrcis user-specific and late in the boot process. Directly editing/procis not persistent. - 3
You are configuring a specialized network driver that requires the parameter
debug=1to be set when the module is loaded. You want this to apply automatically whenever the module is loaded. Which file location is the appropriate standard for this configuration?Show answer details
Correct answer: B
The
/etc/modprobe.d/directory is the standard location for configuration files that specify options formodprobe. You would create a file (e.g.,driver-debug.conf) containingoptions debug=1. - 4
A system administrator runs
sysctl -w net.ipv4.tcp_keepalive_time=300. Five minutes later, the server is rebooted. After the reboot, the administrator checks the value and finds it is back to7200. Why did this happen?Show answer details
Correct answer: C
sysctl -wmodifies the kernel parameter in memory immediately but does not write it to any configuration file. Upon reboot, the kernel loads default values or values from/etc/sysctl.d/. To persist, the setting must be added to a file in/etc/sysctl.d/. - 5
A financial institution is running a latency-sensitive trading application on Red Hat Enterprise Linux 8. The application performance degrades intermittently. You capture the following
iostatoutput during a degradation event:avg-cpu: %user %nice %system %iowait %steal %idle 12.40 0.00 4.10 48.50 0.00 35.00 Device tps kB_read/s kB_wrtn/s kB_read kB_wrtn nvme0n1 850.00 4500.0 125000.0 ... ...Based on these metrics, which subsystem is the primary bottleneck and what is the most appropriate next step to analyze the root cause?
Show answer details
Correct answer: B
The
iostatoutput shows a high%iowait(48.50%) and significant write throughput (125MB/s) relative to the read throughput, while the CPU idle time is 35%. This indicates the system is waiting on I/O completion rather than being CPU bound. The next logical step is to identify which process is generating this I/O using tools likeiotoporpidstat -d. Checking CPU saturation would be incorrect as the CPU is largely idle waiting for I/O.graph TD A[Analyze iostat] --> B{High %iowait?} B -- Yes --> C{High Throughput?} B -- No --> D[Check CPU/Memory] C -- Yes --> E[Identify Process (iotop/pidstat)] C -- No --> F[Check Disk Health/Latency] E --> G[Tune Application/Filesystem] - 6
You are analyzing a database server using
vmstat 1and observe the following output consistently:procs -----------memory---------- ---swap-- -----io---- -system-- ------cpu----- r b swpd free buff cache si so bi bo in cs us sy id wa st 15 0 0 250000 50000 800000 0 0 5 10 4000 8500 85 14 1 0 0The server has 8 vCPUs. Which conclusion is most accurate based on this data?
Show answer details
Correct answer: C
The
rcolumn (runnable processes) is 15, which is nearly double the number of vCPUs (8). Theus(user) +sy(system) CPU usage is 99%, and idle is 1%. This indicates clear CPU saturation. The run queue depth (r) exceeding the core count confirms that processes are waiting for CPU time. Swap (si/so) and I/O (bi/bo) are negligible.
