Install and Use Sysbench on RHEL 9 | CentOS 9

 

Introduction

Sysbench is a versatile, modular, and open-source benchmarking tool widely used to evaluate system performance in various environments. It can test CPU, memory, I/O, file I/O, and database performance, making it indispensable for systems administrators and database engineers. In this post, we’ll walk through how to install Sysbench on RHEL 9 and CentOS 9, and explore some CLI examples to help you get started with benchmarking on your setup.

Prerequisites

To begin, ensure that you have:

• Root or sudo access to your RHEL 9 or CentOS 9 machine.
• A working internet connection to download and install Sysbench from the official repositories.

Prerequisite	Description
OS Version	RHEL 9 or CentOS 9
User Access	Root or sudo privileges
Network	Active internet connection

Let’s dive in!

Install and Use Sysbench on RHEL 9 | CentOS 9

To install Sysbench on RHEL 9 or CentOS 9, we’ll need the EPEL (Extra Packages for Enterprise Linux) repository, as Sysbench isn’t available in the default repository for these distributions.

Add the EPEL Repository

sudo dnf install -y epel-release

This command installs the EPEL repository, which provides additional packages for RHEL-based distributions, including Sysbench.

Install Sysbench

sudo dnf install -y sysbench

Photo by admingeek from Infotechys

Once completed, verify the installation by checking the Sysbench version:

sysbench --version

You should see output similar to:

sysbench 1.0.20

Using Sysbench for Benchmarking

Sysbench allows you to perform several types of benchmarks, such as CPU, memory, and I/O tests. Here’s a look at some of the most common benchmarking commands you can use.

CPU Benchmarking

Test your CPU performance by running:

sysbench cpu --cpu-max-prime=20000 run

The --cpu-max-prime parameter defines the upper limit for calculating prime numbers, with higher values providing more stress on the CPU.

sysbench 1.0.20 (using system LuaJIT 2.1.0-beta3)

Running the test with following options:
Number of threads: 1
Initializing random number generator from current time


Prime numbers limit: 20000

Initializing worker threads...

Threads started!

CPU speed:
    events per second:   333.66

General statistics:
    total time:                          10.0023s
    total number of events:              3338

Latency (ms):
         min:                                    2.86
         avg:                                    3.00
         max:                                    5.70
         95th percentile:                        3.30
         sum:                                 9998.99

Threads fairness:
    events (avg/stddev):           3338.0000/0.00
    execution time (avg/stddev):   9.9990/0.00

This output provides an idea of how well your CPU can handle high workloads.

Memory Benchmarking

Memory performance testing involves writing to and reading from memory. Run the following command to test memory read and write speed:

sysbench memory --memory-block-size=1M --memory-total-size=4G run

• --memory-block-size: Sets the block size for each read/write operation.
• --memory-total-size: Total data size to transfer.

Memory Benchmark Result Table:

Metric	Value
Transfer Rate	4096.00 MiB transferred (9906.76 MiB/sec)
General Statistics	total time: 0.4112s total number of events: 4096
Threads fairness	events (avg/stddev): 4096.0000/0.00 execution time (avg/stddev): 0.4084/0.00

Disk I/O Benchmarking

Disk performance benchmarking in Sysbench can help assess read and write speeds on your storage. First, you’ll need to create a test file:

sysbench fileio --file-total-size=5G prepare

sysbench 1.0.20 (using system LuaJIT 2.1.0-beta3)

128 files, 40960Kb each, 5120Mb total
Creating files for the test...
Extra file open flags: (none)
Creating file test_file.0
Creating file test_file.1
Creating file test_file.2
Creating file test_file.3
Creating file test_file.4
Creating file test_file.5
Creating file test_file.6
Creating file test_file.7
Creating file test_file.8
Creating file test_file.9
Creating file test_file.10
Creating file test_file.11
Creating file test_file.12
Creating file test_file.13
Creating file test_file.14
...omitted for brevity...
Creating file test_file.127
5368709120 bytes written in 53.33 seconds (96.01 MiB/sec).

After creating the files, run the following command to benchmark disk I/O performance:

sysbench fileio --file-total-size=5G --file-test-mode=rndrw run

Photo by admingeek from Infotechys

• --file-total-size: Sets the size of the test file.
• --file-test-mode: Configures Sysbench to perform random read/write operations (rndrw).

Database Benchmarking

Sysbench supports benchmarking for MySQL and PostgreSQL databases. Ensure you have the appropriate database software installed and a test database created. Here’s a simple example with MySQL:

Prepare the database

sysbench oltp_read_write --db-driver=mysql --mysql-user=root --mysql-password=yourpassword --tables=10 --table-size=10000 prepare

Run the Database Benchmark

sysbench oltp_read_write --db-driver=mysql --mysql-user=root --mysql-password=yourpassword --tables=10 --table-size=10000 run

Sample Database Benchmark Result Table

Test	Value
Transactions/sec	500
Reads/sec	2500
Writes/sec	1500
Latency (avg)	4.1 ms

Sysbench Commands Cheat Sheet

To make things easier, here’s a table summarizing essential Sysbench commands for quick reference:

Benchmark Type	Command Example
CPU	sysbench cpu --cpu-max-prime=20000 run
Memory	sysbench memory --memory-block-size=1M --memory-total-size=10G run
Disk I/O	sysbench fileio --file-total-size=5G --file-test-mode=rndrw run
Database	sysbench oltp_read_write --db-driver=mysql --mysql-user=root --mysql-password=yourpassword run

Optimizing Performance Based on Results

Once you’ve gathered benchmark data, consider these steps to optimize performance:

1. CPU Optimization: If CPU performance is lower than expected, consider upgrading your processor or increasing CPU resources in virtualized environments.

2. Memory Tuning: Increase the available memory if your workload requires high memory throughput. Adjusting the swap size can also help in memory-bound applications.

3. Disk I/O: Upgrade to SSDs or NVMe storage if disk read/write speeds are low, or explore using RAID for improved redundancy and performance.

4. Database Optimization: Tune database settings such as cache size, query optimization, and indexing. For MySQL, for example, adjusting innodb_buffer_pool_size can lead to improved performance.